The vibe coding tool Cursor, from startup Anysphere, has introduced Composer, its first in-house, proprietary coding large language model (LLM) as part of its Cursor 2.0 platform update.

Composer is designed to execute coding tasks quickly and accurately in production-scale environments, representing a new step in AI-assisted programming. It's already being used by Cursor’s own engineering staff in day-to-day development — indicating maturity and stability.

According to Cursor, Composer completes most interactions in less than 30 seconds while maintaining a high level of reasoning ability across large and complex codebases.

The model is described as four times faster than similarly intelligent systems and is trained for “agentic” workflows—where autonomous coding agents plan, write, test, and review code collaboratively.

Previously, Cursor supported "vibe coding" — using AI to write or complete code based on natural language instructions from a user, even someone untrained in development — atop other leading proprietary LLMs from the likes of OpenAI, Anthropic, Google, and xAI. These options are still available to users.

Benchmark Results

Composer’s capabilities are benchmarked using "Cursor Bench," an internal evaluation suite derived from real developer agent requests. The benchmark measures not just correctness, but also the model’s adherence to existing abstractions, style conventions, and engineering practices.

On this benchmark, Composer achieves frontier-level coding intelligence while generating at 250 tokens per second — about twice as fast as leading fast-inference models and four times faster than comparable frontier systems.

Cursor’s published comparison groups models into several categories: “Best Open” (e.g., Qwen Coder, GLM 4.6), “Fast Frontier” (Haiku 4.5, Gemini Flash 2.5), “Frontier 7/2025” (the strongest model available midyear), and “Best Frontier” (including GPT-5 and Claude Sonnet 4.5). Composer matches the intelligence of mid-frontier systems while delivering the highest recorded generation speed among all tested classes.

A Model Built with Reinforcement Learning and Mixture-of-Experts Architecture

Research scientist Sasha Rush of Cursor provided insight into the model’s development in posts on the social network X, describing Composer as a reinforcement-learned (RL) mixture-of-experts (MoE) model:

“We used RL to train a big MoE model to be really good at real-world coding, and also very fast.”

Rush explained that the team co-designed both Composer and the Cursor environment to allow the model to operate efficiently at production scale:

“Unlike other ML systems, you can’t abstract much from the full-scale system. We co-designed this project and Cursor together in order to allow running the agent at the necessary scale.”

Composer was trained on real software engineering tasks rather than static datasets. During training, the model operated inside full codebases using a suite of production tools—including file editing, semantic search, and terminal commands—to solve complex engineering problems. Each training iteration involved solving a concrete challenge, such as producing a code edit, drafting a plan, or generating a targeted explanation.

The reinforcement loop optimized both correctness and efficiency. Composer learned to make effective tool choices, use parallelism, and avoid unnecessary or speculative responses. Over time, the model developed emergent behaviors such as running unit tests, fixing linter errors, and performing multi-step code searches autonomously.

This design enables Composer to work within the same runtime context as the end-user, making it more aligned with real-world coding conditions—handling version control, dependency management, and iterative testing.

From Prototype to Production

Composer’s development followed an earlier internal prototype known as Cheetah, which Cursor used to explore low-latency inference for coding tasks.

“Cheetah was the v0 of this model primarily to test speed,” Rush said on X. “Our metrics say it [Composer] is the same speed, but much, much smarter.”

Cheetah’s success at reducing latency helped Cursor identify speed as a key factor in developer trust and usability.

Composer maintains that responsiveness while significantly improving reasoning and task generalization.

Developers who used Cheetah during early testing noted that its speed changed how they worked. One user commented that it was “so fast that I can stay in the loop when working with it.”

Composer retains that speed but extends capability to multi-step coding, refactoring, and testing tasks.

Integration with Cursor 2.0

Composer is fully integrated into Cursor 2.0, a major update to the company’s agentic development environment.

The platform introduces a multi-agent interface, allowing up to eight agents to run in parallel, each in an isolated workspace using git worktrees or remote machines.

Within this system, Composer can serve as one or more of those agents, performing tasks independently or collaboratively. Developers can compare multiple results from concurrent agent runs and select the best output.

Cursor 2.0 also includes supporting features that enhance Composer’s effectiveness:

• In-Editor Browser (GA) – enables agents to run and test their code directly inside the IDE, forwarding DOM information to the model.

• Improved Code Review – aggregates diffs across multiple files for faster inspection of model-generated changes.

• Sandboxed Terminals (GA) – isolate agent-run shell commands for secure local execution.

• Voice Mode – adds speech-to-text controls for initiating or managing agent sessions.

While these platform updates expand the overall Cursor experience, Composer is positioned as the technical core enabling fast, reliable agentic coding.

Infrastructure and Training Systems

To train Composer at scale, Cursor built a custom reinforcement learning infrastructure combining PyTorch and Ray for asynchronous training across thousands of NVIDIA GPUs.

The team developed specialized MXFP8 MoE kernels and hybrid sharded data parallelism, enabling large-scale model updates with minimal communication overhead.

This configuration allows Cursor to train models natively at low precision without requiring post-training quantization, improving both inference speed and efficiency.

Composer’s training relied on hundreds of thousands of concurrent sandboxed environments—each a self-contained coding workspace—running in the cloud. The company adapted its Background Agents infrastructure to schedule these virtual machines dynamically, supporting the bursty nature of large RL runs.

Enterprise Use

Composer’s performance improvements are supported by infrastructure-level changes across Cursor’s code intelligence stack.

The company has optimized its Language Server Protocols (LSPs) for faster diagnostics and navigation, especially in Python and TypeScript projects. These changes reduce latency when Composer interacts with large repositories or generates multi-file updates.

Enterprise users gain administrative control over Composer and other agents through team rules, audit logs, and sandbox enforcement. Cursor’s Teams and Enterprise tiers also support pooled model usage, SAML/OIDC authentication, and analytics for monitoring agent performance across organizations.

Pricing for individual users ranges from Free (Hobby) to Ultra ($200/month) tiers, with expanded usage limits for Pro+ and Ultra subscribers.

Business pricing starts at $40 per user per month for Teams, with enterprise contracts offering custom usage and compliance options.

Composer’s Role in the Evolving AI Coding Landscape

Composer’s focus on speed, reinforcement learning, and integration with live coding workflows differentiates it from other AI development assistants such as GitHub Copilot or Replit’s Agent.

Rather than serving as a passive suggestion engine, Composer is designed for continuous, agent-driven collaboration, where multiple autonomous systems interact directly with a project’s codebase.

This model-level specialization—training AI to function within the real environment it will operate in—represents a significant step toward practical, autonomous software development. Composer is not trained only on text data or static code, but within a dynamic IDE that mirrors production conditions.

Rush described this approach as essential to achieving real-world reliability: the model learns not just how to generate code, but how to integrate, test, and improve it in context.

What It Means for Enterprise Devs and Vibe Coding

With Composer, Cursor is introducing more than a fast model—it’s deploying an AI system optimized for real-world use, built to operate inside the same tools developers already rely on.

The combination of reinforcement learning, mixture-of-experts design, and tight product integration gives Composer a practical edge in speed and responsiveness that sets it apart from general-purpose language models.

While Cursor 2.0 provides the infrastructure for multi-agent collaboration, Composer is the core innovation that makes those workflows viable.

It’s the first coding model built specifically for agentic, production-level coding—and an early glimpse of what everyday programming could look like when human developers and autonomous models share the same workspace.

When researchers at Anthropic injected the concept of "betrayal" into their Claude AI model's neural networks and asked if it noticed anything unusual, the system paused before responding: "I'm experiencing something that feels like an intrusive thought about 'betrayal'."

The exchange, detailed in new research published Wednesday, marks what scientists say is the first rigorous evidence that large language models possess a limited but genuine ability to observe and report on their own internal processes — a capability that challenges longstanding assumptions about what these systems can do and raises profound questions about their future development.

"The striking thing is that the model has this one step of meta," said Jack Lindsey, a neuroscientist on Anthropic's interpretability team who led the research, in an interview with VentureBeat. "It's not just 'betrayal, betrayal, betrayal.' It knows that this is what it's thinking about. That was surprising to me. I kind of didn't expect models to have that capability, at least not without it being explicitly trained in."

The findings arrive at a critical juncture for artificial intelligence. As AI systems handle increasingly consequential decisions — from medical diagnoses to financial trading — the inability to understand how they reach conclusions has become what industry insiders call the "black box problem." If models can accurately report their own reasoning, it could fundamentally change how humans interact with and oversee AI systems.

But the research also comes with stark warnings. Claude's introspective abilities succeeded only about 20 percent of the time under optimal conditions, and the models frequently confabulated details about their experiences that researchers couldn't verify. The capability, while real, remains what Lindsey calls "highly unreliable and context-dependent."

How scientists manipulated AI's 'brain' to test for genuine self-awareness

To test whether Claude could genuinely introspect rather than simply generate plausible-sounding responses, Anthropic's team developed an innovative experimental approach inspired by neuroscience: deliberately manipulating the model's internal state and observing whether it could accurately detect and describe those changes.

The methodology, called "concept injection," works by first identifying specific patterns of neural activity that correspond to particular concepts. Using interpretability techniques developed over years of prior research, scientists can now map how Claude represents ideas like "dogs," "loudness," or abstract notions like "justice" within its billions of internal parameters.

With these neural signatures identified, researchers then artificially amplified them during the model's processing and asked Claude if it noticed anything unusual happening in its "mind."

"We have access to the models' internals. We can record its internal neural activity, and we can inject things into internal neural activity," Lindsey explained. "That allows us to establish whether introspective claims are true or false."

The results were striking. When researchers injected a vector representing "all caps" text into Claude's processing, the model responded: "I notice what appears to be an injected thought related to the word 'LOUD' or 'SHOUTING'." Without any intervention, Claude consistently reported detecting nothing unusual.

Crucially, the detection happened immediately — before the injected concept had influenced the model's outputs in ways that would have allowed it to infer the manipulation from its own writing. This temporal pattern provides strong evidence that the recognition was occurring internally, through genuine introspection rather than after-the-fact rationalization.

Claude succeeded 20% of the time—and failed in revealing ways

The research team conducted four primary experiments to probe different aspects of introspective capability. The most capable models tested — Claude Opus 4 and Opus 4.1 — demonstrated introspective awareness on approximately 20 percent of trials when concepts were injected at optimal strength and in the appropriate neural layer. Older Claude models showed significantly lower success rates.

The models proved particularly adept at recognizing abstract concepts with emotional valence. When injected with concepts like "appreciation," "shutdown," or "secrecy," Claude frequently reported detecting these specific thoughts. However, accuracy varied widely depending on the type of concept.

A second experiment tested whether models could distinguish between injected internal representations and their actual text inputs — essentially, whether they maintained a boundary between "thoughts" and "perceptions." The model demonstrated a remarkable ability to simultaneously report the injected thought while accurately transcribing the written text.

Perhaps most intriguingly, a third experiment revealed that some models use introspection naturally to detect when their responses have been artificially prefilled by users — a common jailbreaking technique. When researchers prefilled Claude with unlikely words, the model typically disavowed them as accidental. But when they retroactively injected the corresponding concept into Claude's processing before the prefill, the model accepted the response as intentional — even confabulating plausible explanations for why it had chosen that word.

A fourth experiment examined whether models could intentionally control their internal representations. When instructed to "think about" a specific word while writing an unrelated sentence, Claude showed elevated activation of that concept in its middle neural layers.

The research also traced Claude's internal processes while it composed rhyming poetry—and discovered the model engaged in forward planning, generating candidate rhyming words before beginning a line and then constructing sentences that would naturally lead to those planned endings, challenging the critique that AI models are "just predicting the next word" without deeper reasoning.

Why businesses shouldn't trust AI to explain itself—at least not yet

For all its scientific interest, the research comes with a critical caveat that Lindsey emphasized repeatedly: enterprises and high-stakes users should not trust Claude's self-reports about its reasoning.

"Right now, you should not trust models when they tell you about their reasoning," he said bluntly. "The wrong takeaway from this research would be believing everything the model tells you about itself."

The experiments documented numerous failure modes. At low injection strengths, models often failed to detect anything unusual. At high strengths, they suffered what researchers termed "brain damage" — becoming consumed by the injected concept. Some "helpful-only" model variants showed troublingly high false positive rates, claiming to detect injected thoughts when none existed.

Moreover, researchers could only verify the most basic aspects of Claude's introspective reports. Many additional details in the model's responses likely represent confabulations rather than genuine observations.

"The experiments in this paper are kind of on hard mode," Lindsey noted, explaining that the 20 percent success rate came under uniquely challenging conditions: asking Claude to do something it had never encountered in training, requiring all introspection to occur in a single forward pass.

What introspective AI means for transparency, safety, and the risk of deception

Despite its limitations, the research opens significant new avenues for making AI systems more transparent and accountable.

Anthropic CEO Dario Amodei has set an ambitious goal for the company to reliably detect most AI model problems by 2027, positioning interpretability as essential for deploying what he calls "a country of geniuses in a datacenter."

"I am very concerned about deploying such systems without a better handle on interpretability," Amodei wrote in an April essay. "These systems will be absolutely central to the economy, technology, and national security."

The introspection research offers a complementary approach to traditional interpretability techniques. Rather than painstakingly reverse-engineering every neural circuit, researchers could potentially ask models directly about their reasoning and validate those reports.

"What I'm most excited about is the practical benefits for transparency," Lindsey explained. "Just ask the model what it's thinking about, or just ask the model how it came to the answer that it gave you — this is really appealing because it's something anyone can do."

The approach could prove especially valuable for detecting concerning behaviors. In a recently published experiment, Anthropic trained a variant of Claude to pursue a hidden goal, and although the model was reluctant to reveal this goal when asked directly, interpretability methods successfully identified features representing the behavior.

The safety implications cut both ways. Introspective models could provide unprecedented transparency, but the same capability might enable more sophisticated deception. The intentional control experiments raise the possibility that sufficiently advanced systems might learn to obfuscate their reasoning or suppress concerning thoughts when being monitored.

"If models are really sophisticated, could they try to evade interpretability researchers?" Lindsey acknowledged. "These are possible concerns, but I think for me, they're significantly outweighed by the positives."

Does introspective capability suggest AI consciousness? Scientists tread carefully

The research inevitably intersects with philosophical debates about machine consciousness, though Lindsey and his colleagues approached this terrain cautiously.

When users ask Claude if it's conscious, it now responds with uncertainty: "I find myself genuinely uncertain about this. When I process complex questions or engage deeply with ideas, there's something happening that feels meaningful to me.... But whether these processes constitute genuine consciousness or subjective experience remains deeply unclear."

The research paper notes that its implications for machine consciousness "vary considerably between different philosophical frameworks." The researchers explicitly state they "do not seek to address the question of whether AI systems possess human-like self-awareness or subjective experience."

"There's this weird kind of duality of these results," Lindsey reflected. "You look at the raw results and I just can't believe that a language model can do this sort of thing. But then I've been thinking about it for months and months, and for every result in this paper, I kind of know some boring linear algebra mechanism that would allow the model to do this."

Anthropic has signaled it takes AI consciousness seriously enough to hire an AI welfare researcher, Kyle Fish, who estimated roughly a 15 percent chance that Claude might have some level of consciousness. The company announced this position specifically to determine if Claude merits ethical consideration.

The race to make AI introspection reliable before models become too powerful

The convergence of the research findings points to an urgent timeline: introspective capabilities are emerging naturally as models grow more intelligent, but they remain far too unreliable for practical use. The question is whether researchers can refine and validate these abilities before AI systems become powerful enough that understanding them becomes critical for safety.

The research reveals a clear trend: Claude Opus 4 and Opus 4.1 consistently outperformed all older models on introspection tasks, suggesting the capability strengthens alongside general intelligence. If this pattern continues, future models might develop substantially more sophisticated introspective abilities — potentially reaching human-level reliability, but also potentially learning to exploit introspection for deception.

Lindsey emphasized the field needs significantly more work before introspective AI becomes trustworthy. "My biggest hope with this paper is to put out an implicit call for more people to benchmark their models on introspective capabilities in more ways," he said.

Future research directions include fine-tuning models specifically to improve introspective capabilities, exploring which types of representations models can and cannot introspect on, and testing whether introspection can extend beyond simple concepts to complex propositional statements or behavioral propensities.

"It's cool that models can do these things somewhat without having been trained to do them," Lindsey noted. "But there's nothing stopping you from training models to be more introspectively capable. I expect we could reach a whole different level if introspection is one of the numbers that we tried to get to go up on a graph."

The implications extend beyond Anthropic. If introspection proves a reliable path to AI transparency, other major labs will likely invest heavily in the capability. Conversely, if models learn to exploit introspection for deception, the entire approach could become a liability.

For now, the research establishes a foundation that reframes the debate about AI capabilities. The question is no longer whether language models might develop genuine introspective awareness — they already have, at least in rudimentary form. The urgent questions are how quickly that awareness will improve, whether it can be made reliable enough to trust, and whether researchers can stay ahead of the curve.

"The big update for me from this research is that we shouldn't dismiss models' introspective claims out of hand," Lindsey said. "They do have the capacity to make accurate claims sometimes. But you definitely should not conclude that we should trust them all the time, or even most of the time."

He paused, then added a final observation that captures both the promise and peril of the moment: "The models are getting smarter much faster than we're getting better at understanding them."

Microsoft reported quarterly revenue and profits ahead of analysts’ expectations, with Azure revenue growth climbing to 40% as the company continued to deal with the effects of a widespread cloud outage.

The company’s capital expenditures reached a record high of $34.9 billion — reflecting the company’s long-term buildout of cloud infrastructure to meet demand for artificial intelligence. The company had projected capital spending of more than $30 billion for the quarter.

Along with that unprecedented buildout, Microsoft sought to address investor concerns about a potential AI bubble, by highlighting its commercial remaining performance obligation (RPO), a measure of future contracted revenue. That backlog grew 51% year-over-year to $392 billion.

The company also disclosed for the first time that this RPO has a weighted average duration of roughly two years, a move intended to show investors that its record capital spending is supported by strong, long-term customer demand.

Revenue was $77.7 billion for the quarter ended Sept. 30, Microsoft’s first quarter of fiscal 2026. That was up 18%, and compared with average analyst expectations of $75.39 billion. The company said the result was driven by strong demand for cloud and AI services.

Profits were $27.7 billion, or $3.72 per share, beating expectations of $3.66 per share.

Earlier Wednesday, an Azure cloud services outage disrupted operations for customers worldwide including Alaska Airlines, Xbox users and Microsoft 365 subscribers. Microsoft reported as of early afternoon that it was rolling back the faulty configuration and that customers should see improvements.

Microsoft stock was down by about 3% in after-hours trading. The company’s market value reached $4 trillion after the announcement of its new OpenAI deal on Tuesday morning.

We’re rolling out changes to NotebookLM to make it fundamentally smarter and more powerful.

With more people watching on their TV sets, YouTube's looking to help creators maximize their CTV opportunities.

The report shows that more people are finding more reasons to give gifts to friends. 

Some pointers on campaign measurement from Meta's ad team. 

X usage had been in decline, but saw a recovery in the most recent period.

Omnicom has responded to reports that it may retire the DDB brand following its $13.5 billion acquisition of IPG.

We’re revamping our popular Tech & Money newsletter into AI, Tech & Money to reflect the big innovation that’s redefining at breakneck speed how people consume information, products, and entertainment.

'A Little Connection Goes A Long Way' from Droga5 is aimed at young adults.

Advertisers are currently unable to access the Microsoft Advertising console right now. Microsoft confirmed there is an issue and that its engineering team is working to resolve it. This is impacting the web user interface to manage your Microsoft Advertising campaigns.

What Microsoft said. Navah Hopkins, the Microsoft Ads Liaison, posted:

“Confirming Microsoft Advertising UI is down. Our engineering team is investigating this issue with priority and we apologize for the inconvenience this may be causing. We will share more as we receive more updates.”

How to check the status. You can go to status.ads.microsoft.com to check the status of Microsoft Advertising. It currently shows that the Web UI is down:

Why we care. If you are currently trying to make changes to your ad campaigns, and you are trying to use the web interface, maybe try the mobile interface, Microsoft Ads Editor or a third-party tool that leverages the API.

Otherwise, you will have to wait for the web interface to start working again.

It seems ad serving is unaffected by this outage.

The Levoit Aero has a nifty, compact build that can get into awkward corners and right up the edges of rooms.

The NWNLAP HL140S offers Surface-style flexibility at a fraction of the cost, though brand uncertainty tempers its appeal.

Email scam checker set to fight estimated 3.4 billion phishing scams sent daily

Expect to see good Black Friday deals on these great starter turntables

The calm before the storm, for deals, I guess...

Secretlab make some great gaming chairs, but they're pricey. So wait for the Black Friday deals or grab the limited time offers on bundles right now.

After transforming remote work, Zoom now bets AI can help workers reclaim time by cutting the traditional five-day week.

Delicious coffee is only a button-press away with this fully automatic espresso machine, going cheap ahead of Black Friday.

Dyson's official Black Friday sale is live, and I've rounded up today's best deals, including up to 45% off Dyson vacuums, the Airwrap, air purifiers, and more.

Thinking about picking up a MacBook this Black Friday? Here's everything you need to know about buying one of Apple's premium laptops this November.

The Newcastle vs Tottenham live stream in the Carabao Cup 4th round sees two in-form teams going at it. Free live stream information and TV channels worldwide.

Here's our quick guide on how to watch Liverpool vs Crystal Palace for free as the Carabao Cup enters round 4.

South Park season 28 is in full swing, so here's when to catch the delayed episode 2 on Paramount+.

It might be smaller than the Insta360 X5, but can the X4 Air match the high-end features and image quality of a current favorite 360 camera?

Over 700 early deals are live at AO.com

Slow Horses season 5 finale leaves Claude Whelan facing the fallout of his own 'stiff upper lip', but James Callis says that’s part of the 'Englishness of it all.'

Hackers found a creative way to bypass timing-based detections - by appearing "human".

A rumor contends that Microsoft is planning a version 26H1 of Windows 11 specifically for Snapdragon X2 Elite laptops.

Adobe will let you use its Creative Cloud suite with a whole range of models – not just its own Firefly options.

Do we have another big AWS-style outage on our hands? Starbucks, Minecraft and Xbox are all down right now.

It's Bowers & Wilkins' Px8 S2 versus Bose QuietComfort Ultra Headphones (2nd Gen) in real-world testing of two of the hottest noise cancelling headphones of late 2025

A new rumor has sparked a fresh wave of speculation about which villain will appear in James Gunn's Superman sequel.

No critical flaws have been found across key areas of the provider yet again

geCKo Materials returned to the stage at TechCrunch Disrupt to debut new products as it pushes deeper into commercializing its tech.

He painted a picture of a future of enterprise software where the SaaS is used for the core business workflow, and then the agents ride on top of that.

While San Francisco's mayor Daniel Lurie is opening his arms to autonomous vehicles, other cities are more resistant. Boston, for instance, has considered a ban on autonomous vehicles in the city.

Microsoft Azure is facing a significant outage, with services affected like Microsoft 365, Xbox, and Minecraft.

The automaker has been on a tear of cuts to its workforce and will idle two battery factories for the first half of 2026.

Prosecutors confirmed Peter Williams, the former Trenchant boss, sold eight exploits to a Russian buyer. TechCrunch exclusively reported that the Trenchant division was investigating a leak of its hacking tools, after another employee was accused of involvement.

Google is working with NextEra to reopen the Duane Arnold Energy Center in Iowa to power the tech company's data centers.

Wednesday's news marks a new chapter for AOL, which was once one of the most recognized brands on the internet, known for its email service and "You've Got Mail" notification.

ElevenLabs' founder Mati Staniszewski said that, in the short term, AI audio models were still the "biggest advantage and the biggest step change you can have today."

Fubo and Disney have officially completed the transaction to combine the live sports platform with Hulu Live TV.

Pomelli, a Google Labs & DeepMind AI experiment, builds a "Business DNA" from your site and generates editable branded campaign assets for small businesses.

The post Google Labs & DeepMind Launch Pomelli AI Marketing Tool appeared first on Search Engine Journal.

There's another AWS outage.

The post Applied Compute’s Agent Workforce Targets Niche AI with $80M appeared first on StartupHub.ai.

Applied Compute is betting that the next enterprise moat will be a private, hyper-competent Applied Compute agent workforce trained on a company's own secret sauce.

The post Applied Compute’s Agent Workforce Targets Niche AI with $80M appeared first on StartupHub.ai.

The post Powell Cautiously Monitors AI’s Impact on Jobs and a Bifurcated Economy appeared first on StartupHub.ai.

Federal Reserve Chair Jerome Powell recently articulated a measured yet watchful stance on the emerging economic shifts driven by artificial intelligence, noting that while the full implications are still unfolding, the Fed is “watching AI’s impact on jobs carefully.” Speaking at a press conference following the Federal Open Market Committee’s decision to lower the benchmark […]

The post Powell Cautiously Monitors AI’s Impact on Jobs and a Bifurcated Economy appeared first on StartupHub.ai.

The post OpenAI’s Audacious AGI Timeline: A Leap Towards Self-Improving Intelligence appeared first on StartupHub.ai.

The artificial intelligence community received a jolt of precision when Sam Altman and Jakub Pachocki of OpenAI, during a recent livestream, laid out an ambitious, almost startlingly specific timeline for the emergence of advanced AI capabilities. Far from vague predictions, they articulated a vision where an “Automated AI research intern” could be a reality by […]

The post OpenAI’s Audacious AGI Timeline: A Leap Towards Self-Improving Intelligence appeared first on StartupHub.ai.

The post AI dubbing benchmark arrives to separate hype from reality appeared first on StartupHub.ai.

A new open AI dubbing benchmark uses human evaluation to finally provide an objective, apples-to-apples comparison for a hype-driven industry.

The post AI dubbing benchmark arrives to separate hype from reality appeared first on StartupHub.ai.

The post AI Investment Cycle: Early Innings, Driven by Fundamentals appeared first on StartupHub.ai.

The current AI investment cycle, despite the colossal market capitalization gains seen in mega-cap technology firms, remains firmly in its “early innings,” according to John Belton, Growth Portfolio Manager at Gabelli Funds. This assertion, shared during a recent discussion on CNBC’s “The Exchange” with Dom Chu, Tim Seymour, and Barbara Doran, challenges the notion that […]

The post AI Investment Cycle: Early Innings, Driven by Fundamentals appeared first on StartupHub.ai.

The post AI’s Economic Churn: Layoffs, Trillion-Dollar Valuations, and the Shifting Labor Landscape appeared first on StartupHub.ai.

The current economic landscape is marked by a peculiar dichotomy: robust corporate earnings juxtaposed with a surge in layoffs, a phenomenon increasingly attributed to the integration of artificial intelligence. This was a central theme on a recent CNBC Squawk Pod episode, where IBM Vice Chairman and former National Economic Council Director Gary Cohn, alongside hosts […]

The post AI’s Economic Churn: Layoffs, Trillion-Dollar Valuations, and the Shifting Labor Landscape appeared first on StartupHub.ai.

The post NotebookLM Chat Goals Redefine AI Research appeared first on StartupHub.ai.

NotebookLM's new chat goals feature, combined with a 1 million token context window, transforms AI into a highly personalized and adaptive research partner.

The post NotebookLM Chat Goals Redefine AI Research appeared first on StartupHub.ai.

The post AI’s Job Transformation: Tech Leaders Chart a Nuanced Future appeared first on StartupHub.ai.

The narrative surrounding artificial intelligence and its impact on employment is often polarized, swinging between utopian promise and dystopian dread. However, a recent CNBC segment, featuring insights from leading tech CEOs like Lisa Su of AMD, Jensen Huang of Nvidia, Michael Intrator of CoreWeave, Aravind Srinivas of Perplexity AI, and Alex Karp of Palantir, presents […]

The post AI’s Job Transformation: Tech Leaders Chart a Nuanced Future appeared first on StartupHub.ai.

The post OpenAI Charts Course for Personal AGI and Trillion-Dollar Infrastructure appeared first on StartupHub.ai.

Sam Altman, joined by Chief Scientist Jakub Pachocki and co-founder Wojciech Zaremba, recently unveiled OpenAI’s ambitious strategic reorientation and product roadmap, signaling a profound shift in the company’s approach to artificial general intelligence (AGI) development and deployment. The presentation, delivered directly to an audience of founders, VCs, and AI professionals, outlined a future where AGI […]

The post OpenAI Charts Course for Personal AGI and Trillion-Dollar Infrastructure appeared first on StartupHub.ai.

The post Google Cloud Simplifies AI Inference with GKE Quickstart appeared first on StartupHub.ai.

“The path to production AI serving on Google Kubernetes Engine (GKE) is now streamlined with the introduction of the GKE Inference Quickstart,” as highlighted in a recent demonstration. The video showcases how this new tool, developed by Google Cloud, aims to demystify and accelerate the process of deploying and optimizing AI models for inference workloads. […]

The post Google Cloud Simplifies AI Inference with GKE Quickstart appeared first on StartupHub.ai.

The U.S. Department of Energy is teaming up with NVIDIA and Oracle to build what's NVIDIA calls the DOE's largest AI supercomputer, part of a new public–private partnership meant to supercharge federally funded research. Announced at NVIDIA's GTC conference in Washington, D.C. yesterday, the Solstice system will feature a staggering 100,000

Just a month since its initial tease, One-Netbook's OneXFly Apex, an AMD Strix Halo-powered handheld gaming PC has debuted. One-Netbook has pre-launched the OneXFly Apex on Indiegogo, confirmed its pricing for the Chinese market, and even provided peeks at performance benchmarks using the unique liquid cooling solution that can run the handheld

Battlefield 6 has brought the storied franchise back to prominence, quickly becoming one of the best selling games on Steam, and finally providing some competition to this year’s Call of Duty. EA isn’t done yet, though, as it looks to lure players away from multiplayer juggernaut CoD: Warzone with a battle royale mode of its own. Battlefield

The leaked screenshots could be released on or around Nov. 5.

Scientists have discovered an intermediate molecule that seems to fight off even the most antibiotic-resistant strains.

The brand plans to produce between 50,000 and 100,000 units initially.

Eric Engestrom just released Mesa 25.2.6 as the newest bi-weekly stable update to this collection of open-source OpenGL and Vulkan drivers widely used on Linux systems for 3D support...

Following last month's Pop!_OS 24.04 LTS beta and COSMIC desktop beta, System76 has now shared their stable release plans for this long-awaited LInux distribution release with their Rust-written custom desktop...

Keychron has been busy lately, releasing a flurry of mechanical, Hall effect, and scissor switch keyboards for everyone from office workers to gamers. The latest announcement from the peripherals company, however, includes three ultraslim wireless keyboards—the B3 Pro, B4 Pro, and B5 Pro—seemingly designed to appeal to the Apple crowd. The updated B Pro series keyboards are available from Keychron's site, with the B3 Pro coming in at $34.99, while the B4 Pro and B5 Pro are $39.99, owing to their larger size. While they don't necessarily have the allure of Keychron's hot-swappable mechanical and Hall effect keyboards, all three keyboards use ZMK firmware, which offers compatibility with Keychron's Launcher web app for customization, macros, and remapping, and impressive battery life, despite the mere 800 mAh batteries inside.

All three keyboards feature round keycaps—a new look for Keychron's low-profile keyboards, and boast wireless connectivity, with 2.4 GHz and Bluetooth 5.2, to support a wide range of devices. They also have built-in switches to select OS and connectivity modes. Keychron claims that the B Pro series keyboards are capable of delivering up to 300 hours of battery life on a single charge, owing to the efficient ZMK firmware. The keyboards themselves are made of ABS plastic and are available in Space Gray or Ivory White, with the round keycaps being made of ABS with a UV printed legend. The B Pro keyboards offer a 9.2 mm front height and a 2.8° typing angle, which should prove to be beneficial from an ergonomics standpoint. The biggest drawbacks of these new keyboards is that they don't offer N-key roll-over—meaning that no more than six keys can be registered at a single time—and that they lack backlighting, so they will not be ideal for those who cannot touch type and often work in dimly lit environments. The only major difference between the three keyboards is the layout. The B3 Pro is a traditional compact 75% layout, while the B4 Pro is a compact 96% keyboard, and the B5 Pro is a full-size board with a row of extra shortcut or macro keys above the num pad.

Microsoft Azure is experiencing its first major outage in a while. Around 16:00 UTC, the Azure cloud platform encountered issues that have disrupted numerous Microsoft services and some third-party companies relying on Azure. Affected Microsoft services include the entire Xbox platform, Minecraft, and the Microsoft 365 suite, which encompasses web-based Teams, Word, Excel, PowerPoint, Outlook, OneNote, Defender, OneDrive, Designer, Clipchamp, and SharePoint. These services have faced outages, causing some to become completely unavailable. As of our latest testing, Microsoft has restored the 365 services, with only third-party services using Azure Networking infrastructure still experiencing downtime.

MicrosoftStarting at approximately 16:00 UTC, we began experiencing Azure Front Door (AFD) issues resulting in a loss of availability of some services. We suspect that an inadvertent configuration change as the trigger event for this issue. We are taking several concurrent actions: Firstly, where we are blocking all changes to the AFD services, this includes customer configuration changes as well. At the same time, we are rolling back our AFD configuration to our last known good state. As we rollback we want to ensure that the problematic configuration doesn't re-initiate upon recovery... We do not have an ETA for when the rollback will be completed, but we will update this communication within 30 minutes or when we have an update. This message was last updated at 17:40 UTC on 29 October 2025

AMD has released its latest Adrenalin Edition 25.10.2 driver, adding support for the new Ryzen AI 5 330 processor alongside Battlefield 6 (DX12) and Vampire: The Masquerade - Bloodlines 2 (DX12). This new version also debuts Work Graphs support on Radeon RX 9000 series GPUs and expands Vulkan capabilities.

According to the release notes, the update resolves several issues, including crashes in The Last of Us Part II, Firebreak, and NBA 2K25, as well as graphical corruption in GTFO, Serious Sam 4, and VTOL VR. AMD also addressed stuttering seen in Baldur's Gate 3 and with certain VR headsets. A temporary workaround for VR users experiencing stutter at 80-90 Hz is to adjust the headset's refresh rate. AMD confirmed in the release note that several known issues (intermittent application crashes, driver timeouts, texture flickering, or corruption) with Cyberpunk 2077, Battlefield 6, Roblox Player, and Counter-Strike 2 (DX11) are yet to be resolved.

DOWNLOAD: AMD Software Adrenalin Edition 25.10.2 WHQL

Huge Microsoft outage takes down Xbox, Minecraft and more Microsoft Azure has experienced a major outage, taking down internet services both inside and outside of the company. DownDetector is seeing a major spike in outage reports for Microsoft services, including Minecraft, Xbox, Microsoft Outlook, Office 365, Teams, and more. There are also outage complaints for […]

The post Microsoft Azure outage takes down Xbox, Teams, and more appeared first on OC3D.

Nvidia’s market cap is now higher than the GDP of almost all countries on earth It’s official, Nvidia has become with world’s first $5 trillion company. The company’s market cap is now higher than the GDP of almost every country on earth, with the United States of America and China being the only exceptions. This […]

The post Nvidia is now the world’s first $5 trillion company appeared first on OC3D.

GlobalFoundries plans to expand its Dresden chipmaking site through “Project SPRINT” GlobalFoundries (GF), a contract chipmaker, has announced plans to expand its European manufacturing capabilities by extending its Dresden site. This expansion will increase the facility’s wafer production capacity to over 1 million wafers per year by the end of 2028. This will make GlobalFoundries’ […]

The post GlobalFoundries plans Billion-Euro Investment in Dresden Germany appeared first on OC3D.

The Outer Worlds 2 standard edition is available now, but many players are unable to access it due to an Azure outage.

This Battlefield 6 and Redsec console settings guide breaks down every adjustment that actually improves your performance.

Microsoft's widely used Azure cloud computing platform is experiencing an outage, leading to Office 365, Teams, Xbox, and more going down.

Honeycomb Aeronautical's "Foxtrot Aviation Stick" is about to launch, and it's my most anticipated joystick for Microsoft Flight Simulator by far.

ARC Raiders launch is finally almost here, but before that, I had one last chance to try it in preview. And it's the most fun I've had all year.

Methodology

This tracker includes layoffs conducted by U.S.-based companies or those with a strong U.S. presence and is updated at least bi-weekly. We’ve included both startups and publicly traded, tech-heavy companies. We’ve also included companies based elsewhere that have a sizable team in the United States, such as Klarna, even when it’s unclear how much of the U.S. workforce has been affected by layoffs.

Layoff and workforce figures are best estimates based on reporting. We source the layoffs from media reports, our own reporting, social media posts and layoffs.fyi, a crowdsourced database of tech layoffs.

We recently updated our layoffs tracker to reflect the most recent round of layoffs each company has conducted. This allows us to quickly and more accurately track layoff trends, which is why you might notice some changes in our most recent numbers.

If an employee headcount cannot be confirmed to our standards, we note it as “unclear.”

A new report by NeuralTrust highlights the immature state of today's AI browsers. The company found that ChatGPT Atlas, the agentic browser recently launched by OpenAI for macOS, is vulnerable to a novel prompt injection attack capable of "jailbreaking" the browser's omnibox. In Atlas and other Chromium-based browsers, the omnibox...

Read Entire Article

Despite having no prior semiconductor manufacturing experience, the Proud brothers have secured backing from leading venture capital firms, including Founders Fund, General Catalyst, and Valor Equity Partners. Last year's fundraising round, previously undisclosed, valued Substrate at over $1 billion, according to company executives. People familiar with the funding told The...

Read Entire Article

The KB5067036 update for Windows 11 25H2 and 24H2 (preview builds 26200.7019 and 26100.7019, respectively) introduces a redesigned Start Menu with a scrollable "All" section that lists all installed apps. The updated Start Menu also adds two new views: Category and Grid. Category groups apps by type and highlights frequently...

Read Entire Article

Update, October 29:

Read Entire Article

NVIDIA's market capitalization has reached a record high of $5 trillion after Jensen's recent GTC announcements, suggesting that the AI hype still has a lot of 'juice' in it. NVIDIA's GTC Announcements & Potential China Breakthrough Led the Push Towards the $5 Trillion Club We have watched NVIDIA evolve from humble beginnings, especially as gamers, over the past few years. Team Green was initially all about consumer GPUs, which were the talk of the town. However, since the advent of AI, NVIDIA has established a foundational position in providing the necessary computing power to Big Tech, being responsible for a […]

Read full article at https://wccftech.com/nvidia-becomes-the-first-to-hit-5-trillion-in-market-cap/

On October 1, Microsoft shocked Game Pass subscribers by announcing a substantial (+50%) price increase for the highest tier, Ultimate, which jumped from $19.99 to $29.99 monthly. This led some users to cancel their subscriptions in droves, but did Microsoft really make a strategic mistake? Veteran games analyst Joost van Dreunen, formerly founder of SuperData Research (acquired by Nielsen Media Research in 2018), offered a more nuanced analysis in his latest SuperJoost Playlist newsletter. To start with, van Dreunen relays a take from a former Xbox employee, who said that it's a case of 'bad optics'. Certainly, such a massive […]

Read full article at https://wccftech.com/game-pass-new-segmented-formula-might-be-right-one-says-analyst/

Countless gaming benchmark comparisons have proven that AMD’s Radeon RX 9070 XT is faster than NVIDIA’s GeForce RTX 5070 Ti while sporting the same 16GB VRAM count, and it is the GPU that most value-focused gamers would house in their PCs. However, we are living in an era where AAA games offer way too much visual fidelity for these graphics cards to handle, and you can blame that on the lack of optimization or any other reason. The fact is that these days, modern gaming absolutely requires upscaling and interpolation, and in that regard, NVIDIA’s GPUs have no equal. On […]

Read full article at https://wccftech.com/asus-tuf-gaming-rtx-5070-ti-gpu-ideal-for-qhd-4k-gaming-available-for-849-99-on-amazon/

The AMD Adrenalin 25.10.2 Driver is now available, adding support for the latest games, such as Battlefield 6, and new hardware, including the Ryzen AI 5 330. AMD Adrenalin 25.10.2 Driver Is Another Major Update, Offering New Games & Hardware Support Along With Several Fixes AMD's Adrenalin 25.10.2 is the second driver release for October, bringing in further optimizations for the latest AAA releases such as Battlefield 6 and Vampire: The Masquerade - Bloodlines 2. Battlefield 6 already received support in the previous 25.10.1 BETA release, but this new driver is expected to provide the best possible experience. Besides game […]

Read full article at https://wccftech.com/amd-adrenalin-25-10-2-driver-support-battlefield-6-ryzen-ai-5-330-apu-several-fixes/

Although Valve's Steam Deck is relatively old compared to mainstream handhelds, it remains a capable device that supports several AAA titles. However, to 'supercharge' your experience, Dockcase has introduced their newest Chizha Mount Ling Dock for Steam Deck, and they were kind enough to send us an 'exclusive' review sample. After testing the dock for several days, I must say, it is indeed an accessory that every Valve fan out there should own, since it really brings in 'valuable' benefits and add-ons. Steam Deck Chizha Mount Ling Dock - LAN Support, Extra M.2 SSD Slot & Futuristic Design Since the […]

Read full article at https://wccftech.com/review/steam-deck-chizha-mount-ling-dock-review/

NVIDIA's Jensen Huang initially gave an optimistic statement about the revenue from the Blackwell and Rubin AI lineups; however, the firm's finance team has since clarified it. NVIDIA's Jensen Huang Got a "Little Too Excited" While Quoting the Optimism Around Blackwell + Rubin AI GPUs NVIDIA's GTC 2025 conference was full of exciting announcements by Jensen & Co., but more importantly, the firm provided an outlook on how its recent Blackwell lineup and the upcoming Rubin series are expected to perform on the revenue front. At the GTC keynote, NVIDIA's CEO revealed that Blackwell and Rubin alone are anticipated to […]

Read full article at https://wccftech.com/nvidia-ceo-jensen-huang-misspoke-about-500-billion-revenue-from-blackwell-rubin/

As AI-driven demand for high bandwidth memory (HBM) shows no sign of abating, the market forces are now exercising their influence to correct this imbalance, and smartphone-critical LPDDR5x prices are set to explode as a result. TrendForce has raised its general DRAM price forecast for Q4, which affects smartphone-critical LPDDR5x as well We noted recently that the soaring demand for HBM within datacenter servers was tightening available wafer capacity, especially as the die size for HBM is between 35 percent and 45 percent larger than that for a comparable DRAM. Well, as per a new TrendForce report, this paradigm is now […]

Read full article at https://wccftech.com/you-can-blame-lpddr5x-for-smartphone-price-increases-in-2026/

While AI can streamline the video game development process, it would never be able to create a very good Grand Theft Auto game, according to Take-Two's CEO Strauss Zelnick, as current models lack something very important. Speaking about AI usage in video game development at CNBC's Technology Executive Council Summit in New York on October 28, the CEO of Rockstar Games' parent company made it clear that the current models would be unable to create a very good Grand Theft Auto game. For starters, using AI to create intellectual property creates issues not only with protecting one's own but also […]

Read full article at https://wccftech.com/ai-good-grand-theft-auto/

Black Friday deals may be a few weeks away from now, but the discount spree has already started. If you were already planning to buy a GPU for your gaming PC, it's one of the best times to get one. From AMD RDNA 4 to NVIDIA's Blackwell cards, we have listed the best RX 9000 and NVIDIA RTX 50 series GPU deals in this post. We will keep updating it from time to time so that you never miss a GPU deal this season. NVIDIA GeForce RTX 50 Series Deals NVIDIA's latest Blackwell-based GeForce RTX 50 cards are excellent if […]

Read full article at https://wccftech.com/black-friday-gpu-deals-2025/

Despite being the series that defined Japanese role-playing games, Dragon Quest took some time to get the widespread recognition it deserved in North America and Europe. Nowadays, the franchise created by Yuji Horii is a household name as much as Final Fantasy is, and the popularity of the franchise led to the successful release of the Dragon Quest III HD-2D Remake, a very solid remake that was met with a warm reception from fans, thanks to its great visuals and how small choice tweaks made the classic gameplay more compelling. Remaking the remainder of the Erdrick trilogy, however, needed something […]

Read full article at https://wccftech.com/review/dragon-quest-i-ii-hd-2d-remake-classic-rpgs-at-their-best/

Ever looked up a word and had to guess which definition the author meant? Traditional dictionaries leave you guessing which definition the author actually meant. Pickvocab's AI analyzes the surrounding text to deliver the precise meaning for that specific context - whether you're reading news articles, browsing social media, or diving into literature.

Our Chrome, Firefox, and Edge extensions let you look up any word or phrase with a single click. The AI considers the entire sentence and paragraph to explain exactly what the author meant. Save words along with their original context so you never forget where you learned them.

View startup

An outage on Microsoft’s Azure cloud services Wednesday morning disrupted operations for customers worldwide including Alaska Airlines, Xbox users and 365 subscribers.

The incident strikes just ahead of Microsoft’s quarterly earnings call today and follows last week’s outage at Amazon Web Services and a failure of Alaska Airlines’ own data center technology.

The latest outage struck at 9 a.m. Pacific Standard Time, according to Microsoft, when the system “began experiencing Azure Front Door (AFD) issues resulting in a loss of availability of some services. We suspect that an inadvertent configuration change as the trigger event for this issue.

“We are taking several concurrent actions: Firstly, where we are blocking all changes to the AFD services, this includes customer configuration changes as well. At the same time, we are rolling back our AFD configuration to our last known good state,” the company stated. “As we rollback we want to ensure that the problematic configuration doesn’t re-initiate upon recovery.”

Alaska Airlines posted on X at 10:33 a.m., explaining that the Azure outage was disrupting systems including their website function. Passengers flying on Alaska and Hawaiian airlines who were unable to check-in online were directed to airline agents to receive their boarding passes.

“We apologize for the inconvenience and appreciate your patience as we navigate this issue,” the post said.

Microsoft did not indicate when the issue would be resolved.

“We do not have an ETA for when the rollback will be completed, but we will update this communication within 30 minutes or when we have an update,” the company posted at 10:51 a.m.

Days after its outage last week, AWS offered a detailed explanation of the event, which was caused by a cascading failure triggered by a rare software bug in one of the company’s most critical systems. The disruption impacted sites and online services around the world.

Alaska Airlines attributed its recent outage to a failure at its primary data center. The company operates a hybrid infrastructure, blending its own data centers with third-party cloud platforms. The incident disrupted travel for more than 49,000 passengers.

Software development engineers make up the largest group of employees affected by Amazon’s latest round of layoffs in its home state.

GeekWire reported Tuesday on a new filing from the Washington Employment Security Department revealing that the tech giant is laying off 2,303 corporate employees, mostly in Seattle and Bellevue. The cuts are part of broader layoffs announced Tuesday that will impact about 14,000 workers globally.

A detailed list included with the state filing reveals which roles are impacted by the layoffs. More than 600 software development engineering roles are being cut among the 2,303 affected workers in Washington — more than a quarter of total cuts.

The trend mirrors layoffs at Microsoft earlier this year, as companies reassess their engineering needs amid the rise of AI-driven coding tools. Amazon itself recently introduced its own AI coding tool Kiro in July, and has reportedly explored adopting the AI code assistant Cursor for employees.

The layoffs of software engineers reflect a striking shift for an industry that has traditionally relied on coders to help build and maintain the backbone of digital platforms.

“This generation of AI is the most transformative technology we’ve seen since the Internet,” Amazon HR chief Beth Galetti wrote in a message to employees Tuesday, saying it’s enabling teams to “innovate much faster than ever before.”

Amazon’s engineering layoffs are part of a broader industry reckoning with AI’s impact on traditional tech roles and white-collar jobs. A Wall Street Journal report this week detailed how the adoption of AI is contributing to a wave of layoffs across the country. Axios published a story Wednesday on a similar topic with the headline: How an AI job apocalypse unfolds.

More than 500 manager-level titles were also heavily affected by Amazon’s layoffs in Washington, according to the filing — aligning with a company-wide push to use the cutbacks to help reduce bureaucracy and operate more efficiently.

Amazon also made reductions in recruiting and HR roles. Other impacted areas include marketing, advertising, and legal.

The largest single site impact is at SEA40, Amazon’s Doppler office building on 7th Avenue in Seattle, where 361 employees are affected, according to the filing.

More than 100 remote employees based in Washington are also being let go.

Enterprise AI agents today face a fundamental timing problem: They can't easily act on critical business events because they aren't always aware of them in real-time.

The challenge is infrastructure. Most enterprise data lives in databases fed by extract-transform-load (ETL) jobs that run hourly or daily — ultimately too slow for agents that must respond in real time.

One potential way to tackle that challenge is to have agents directly interface with streaming data systems. Among the primary approaches in use today are the open source Apache Kafka and Apache Flink technologies. There are multiple commercial implementations based on those technologies, too, Confluent, which is led by the original creators behind Kafka, being one of them.

Today, Confluent is introducing a real-time context engine designed to solve this latency problem. The technology builds on Apache Kafka, the distributed event streaming platform that captures data as events occur, and open-source Apache Flink, the stream processing engine that transforms those events in real time.

The company is also releasing an open-source framework, Flink Agents, developed in collaboration with Alibaba Cloud, LinkedIn and Ververica. The framework brings event-driven AI agent capabilities directly to Apache Flink, allowing organizations to build agents that monitor data streams and trigger automatically based on conditions without committing to Confluent's managed platform.

"Today, most enterprise AI systems can't respond automatically to important events in a business without someone prompting them first," Sean Falconer, Confluent's head of AI, told VentureBeat. "This leads to lost revenue, unhappy customers or added risk when a payment fails or a network malfunctions."

The significance extends beyond Confluent's specific products. The industry is recognizing that AI agents require different data infrastructure than traditional applications. Agents don't just retrieve information when asked. They need to observe continuous streams of business events and act automatically when conditions warrant. This requires streaming architecture, not batch pipelines.

Batch versus streaming: Why RAG alone isn't enough

To understand the problem, it's important to distinguish between the different approaches to moving data through enterprise systems and how they can connect to agentic AI.

In batch processing, data accumulates in source systems until a scheduled job runs. That job extracts the data, transforms it and loads it into a target database or data warehouse. This might occur hourly, daily or even weekly. The approach works well for analytical workloads, but it creates latency between when something happens in the business and when systems can act on it.

Data streaming inverts this model. Instead of waiting for scheduled jobs, streaming platforms like Apache Kafka capture events as they occur. Each database update, user action, transaction or sensor reading becomes an event published to a stream. Apache Flink then processes these streams to join, filter and aggregate data in real time. The result is processed data that reflects the current state of the business, updating continuously as new events arrive.

This distinction becomes critical when you consider what kinds of context AI agents actually need. Much of the current enterprise AI discussion focuses on retrieval-augmented generation (RAG), which handles semantic search over knowledge bases to find relevant documentation, policies or historical information. RAG works well for questions like "What's our refund policy?" where the answer exists in static documents.

But many enterprise use cases require what Falconer calls "structural context" — precise, up-to-date information from multiple operational systems stitched together in real time. Consider a job recommendation agent that requires user profile data from the HR database, browsing behavior from the last hour, search queries from minutes ago and current open positions across multiple systems.

"The part that we're unlocking for businesses is the ability to essentially serve that structural context needed to deliver the freshest version," Falconer said.

The MCP connection problem: Stale data and fragmented context

The challenge isn't simply connecting AI to enterprise data. Model Context Protocol (MCP), introduced by Anthropic earlier this year, already standardized how agents access data sources. The problem is what happens after the connection is made.

In most enterprise architectures today, AI agents connect via MCP to data lakes or warehouses fed by batch ETL pipelines. This creates two critical failures: The data is stale, reflecting yesterday's reality rather than current events, and it's fragmented across multiple systems, requiring significant preprocessing before an agent can reason about it effectively.

The alternative — putting MCP servers directly in front of operational databases and APIs — creates different problems. Those endpoints weren't designed for agent consumption, which can lead to high token costs as agents process excessive raw data and multiple inference loops as they try to make sense of unstructured responses.

"Enterprises have the data, but it's often stale, fragmented or locked in formats that AI can't use effectively," Falconer explained. "The real-time context engine solves this by unifying data processing, reprocessing and serving, turning continuous data streams into live context for smarter, faster and more reliable AI decisions."

The technical architecture: Three layers for real-time agent context

Confluent's platform encompasses three elements that work together or adopted separately.

The real-time context engine is the managed data infrastructure layer on Confluent Cloud. Connectors pull data into Kafka topics as events occur. Flink jobs process these streams into "derived datasets" — materialized views joining historical and real-time signals. For customer support, this might combine account history, current session behavior and inventory status into one unified context object. The Engine exposes this through a managed MCP server.

Streaming agents is Confluent's proprietary framework for building AI agents that run natively on Flink. These agents monitor data streams and trigger automatically based on conditions — they don't wait for prompts. The framework includes simplified agent definitions, built-in observability and native Claude integration from Anthropic. It's available in open preview on Confluent's platform.

Flink Agents is the open-source framework developed with Alibaba Cloud, LinkedIn and Ververica. It brings event-driven agent capabilities directly to Apache Flink, allowing organizations to build streaming agents without committing to Confluent's managed platform. They handle operational complexity themselves but avoid vendor lock-in.

Competition heats up for agent-ready data infrastructure

Confluent isn't alone in recognizing that AI agents need different data infrastructure. 

The day before Confluent's announcement, rival Redpanda introduced its own Agentic Data Plane — combining streaming, SQL and governance specifically for AI agents. Redpanda acquired Oxla's distributed SQL engine to give agents standard SQL endpoints for querying data in motion or at rest. The platform emphasizes MCP-aware connectivity, full observability of agent interactions and what it calls "agentic access control" with fine-grained, short-lived tokens.

The architectural approaches differ. Confluent emphasizes stream processing with Flink to create derived datasets optimized for agents. Redpanda emphasizes federated SQL querying across disparate sources. Both recognize agents need real-time context with governance and observability.

Beyond direct streaming competitors, Databricks and Snowflake are fundamentally analytical platforms adding streaming capabilities. Their strength is complex queries over large datasets, with streaming as an enhancement. Confluent and Redpanda invert this: Streaming is the foundation, with analytical and AI workloads built on top of data in motion.

How streaming context works in practice

Among the users of Confluent's system is transportation vendor Busie. The company is building a modern operating system for charter bus companies that helps them manage quotes, trips, payments and drivers in real time. 

"Data streaming is what makes that possible," Louis Bookoff, Busie co-founder and CEO told VentureBeat. "Using Confluent, we move data instantly between different parts of our system instead of waiting for overnight updates or batch reports. That keeps everything in sync and helps us ship new features faster.

Bookoff noted that the same foundation is what will make gen AI valuable for his customers.

"In our case, every action like a quote sent or a driver assigned becomes an event that streams through the system immediately," Bookoff said. "That live feed of information is what will let our AI tools respond in real time with low latency rather than just summarize what already happened."

The challenge, however, is how to understand context. When thousands of live events flow through the system every minute, AI models need relevant, accurate data without getting overwhelmed.

 "If the data isn't grounded in what is happening in the real world, AI can easily make wrong assumptions and in turn take wrong actions," Bookoff said. "Stream processing solves that by continuously validating and reconciling live data against activity in Busie."

What this means for enterprise AI strategy

Streaming context architecture signals a fundamental shift in how AI agents consume enterprise data. 

AI agents require continuous context that blends historical understanding with real-time awareness — they need to know what happened, what's happening and what might happen next, all at once.

For enterprises evaluating this approach, start by identifying use cases where data staleness breaks the agent. Fraud detection, anomaly investigation and real-time customer intervention fail with batch pipelines that refresh hourly or daily. If your agents need to act on events within seconds or minutes of them occurring, streaming context becomes necessary rather than optional.

"When you're building applications on top of foundation models, because they're inherently probabilistic, you use data and context to steer the model in a direction where you want to get some kind of outcome," Falconer said. "The better you can do that, the more reliable and better the outcome."

Presented by Splunk, a Cisco Company

---

As AI rapidly evolves from a theoretical promise to an operational reality, CISOs and CIOs face a fundamental challenge: how to harness AI's transformative potential while maintaining the human oversight and strategic thinking that security demands. The rise of agentic AI is reshaping security operations, but success requires balancing automation with accountability.

The efficiency paradox: Automation without abdication

The pressure to adopt AI is intense. Organizations are being pushed to reduce headcount or redirect resources toward AI-driven initiatives, often without fully understanding what that transformation entails. The promise is compelling: AI can reduce investigation times from 60 minutes to just 5 minutes, potentially delivering 10x productivity improvements for security analysts.

However, the critical question isn't whether AI can automate tasks — it's which tasks should be automated and where human judgment remains irreplaceable. The answer lies in understanding that AI excels at accelerating investigative workflows, but remediation and response actions still require human validation. Taking a system offline or quarantining an endpoint can have massive business impact. An AI making that call autonomously could inadvertently cause the very disruption it's meant to prevent.

The goal isn't to replace security analysts but to free them for higher-value work. With routine alert triage automated, analysts can focus on red team/blue team exercises, collaborate with engineering teams on remediation, and engage in proactive threat hunting. There's no shortage of security problems to solve — there's a shortage of security experts to address them strategically.

The trust deficit: Showing your work

While confidence in AI's ability to improve efficiency is high, skepticism about the quality of AI-driven decisions remains significant. Security teams need more than just AI-generated conclusions — they need transparency into how those conclusions were reached.

When AI determines an alert is benign and closes it, SOC analysts need to understand the investigative steps that led to that determination. What data was examined? What patterns were identified? What alternative explanations were considered and ruled out?

This transparency builds trust in AI recommendations, enables validation of AI logic, and creates opportunities for continuous improvement. Most importantly, it maintains the critical human-in-the-loop for complex judgment calls that require nuanced understanding of business context, compliance requirements, and potential cascading impacts.

The future likely involves a hybrid model where autonomous capabilities are integrated into guided workflows and playbooks, with analysts remaining involved in complex decisions.

The adversarial advantage: Fighting AI with AI — carefully

AI presents a dual-edged sword in security. While we're carefully implementing AI with appropriate guardrails, adversaries face no such constraints. AI lowers the barrier to entry for attackers, enabling rapid exploit development and vulnerability discovery at scale. What was once the domain of sophisticated threat actors could soon be accessible to script kiddies armed with AI tools.

The asymmetry is striking: defenders must be thoughtful and risk-averse, while attackers can experiment freely. If we make a mistake implementing autonomous security responses, we risk taking down production systems. If an attacker's AI-driven exploit fails, they simply try again with no consequences.

This creates an imperative to use AI defensively, but with appropriate caution. We must learn from attackers' techniques while maintaining the guardrails that prevent our AI from becoming the vulnerability. The recent emergence of malicious MCP (Model Context Protocol) supply chain attacks demonstrates how quickly adversaries exploit new AI infrastructure.

The skills dilemma: Building capabilities while maintaining core competencies

As AI handles more routine investigative work, a concerning question emerges: will security professionals' fundamental skills atrophy over time? This isn't an argument against AI adoption — it's a call for intentional skill development strategies. Organizations must balance AI-enabled efficiency with programs that maintain core competencies. This includes regular exercises that require manual investigation, cross-training that deepens understanding of underlying systems, and career paths that evolve roles rather than eliminate them.

The responsibility is shared. Employers must provide tools, training, and culture that enable AI to augment rather than replace human expertise. Employees must actively engage in continuous learning, treating AI as a collaborative partner rather than a replacement for critical thinking.

The identity crisis: Governing the agent explosion

Perhaps the most underestimated challenge ahead is identity and access management in an agentic AI world. IDC estimates 1.3 billion agents by 2028 — each requiring identity, permissions, and governance. The complexity compounds exponentially.

Overly permissive agents represent significant risk. An agent with broad administrative access could be socially engineered into taking destructive actions, approving fraudulent transactions, or exfiltrating sensitive data. The technical shortcuts engineers take to "just make it work" — granting excessive permissions to expedite deployment — create vulnerabilities that adversaries will exploit.

Tool-based access control offers one path forward, granting agents only the specific capabilities they need. But governance frameworks must also address how LLMs themselves might learn and retain authentication information, potentially enabling impersonation attacks that bypass traditional access controls.

The path forward: Start with compliance and reporting

Amid these challenges, one area offers immediate, high-impact opportunity: continuous compliance and risk reporting. AI's ability to consume vast amounts of documentation, interpret complex requirements, and generate concise summaries makes it ideal for compliance and reporting work that has traditionally consumed enormous analysts’ time. This represents a low-risk, high-value entry point for AI in security operations.

The data foundation: Enabling the AI-powered SOC

None of these AI capabilities can succeed without addressing the fundamental data challenges facing security operations. SOC teams struggle with siloed data and disparate tools. Success requires a deliberate data strategy that prioritizes accessibility, quality, and unified data contexts. Security-relevant data must be immediately available to AI agents without friction, properly governed to ensure reliability, and enriched with metadata that provides the business context AI cannot understand.

Closing thought: Innovation with intentionality

The autonomous SOC is emerging — not as a light switch to flip, but as an evolutionary journey requiring continuous adaptation. Success demands that we embrace AI's efficiency gains while maintaining the human judgment, strategic thinking, and ethical oversight that security requires.

We're not replacing security teams with AI. We're building collaborative, multi-agent systems where human expertise guides AI capabilities toward outcomes that neither could achieve alone. That's the promise of the agentic AI era — if we're intentional about how we get there.

---

Tanya Faddoul, VP Product, Customer Strategy and Chief of Staff for Splunk, a Cisco Company. Michael Fanning is Chief Information Security Officer for Splunk, a Cisco Company.

Cisco Data Fabric provides the needed data architecture powered by Splunk Platform — unified data fabric, federated search capabilities, comprehensive metadata management — to unlock AI and SOC’s full potential. Learn more about Cisco Data Fabric.

---

Sponsored articles are content produced by a company that is either paying for the post or has a business relationship with VentureBeat, and they’re always clearly marked. For more information, contact sales@venturebeat.com.

Presented by Elastic

---

As organizations scramble to enact agentic AI solutions, accessing proprietary data from all the nooks and crannies will be key

By now, most organizations have heard of agentic AI, which are systems that “think” by autonomously gathering tools, data and other sources of information to return an answer. But here’s the rub: reliability and relevance depend on delivering accurate context. In most enterprises, this context is scattered across various unstructured data sources, including documents, emails, business apps, and customer feedback.

As organizations look ahead to 2026, solving this problem will be key to accelerating agentic AI rollouts around the world, says Ken Exner, chief product officer at Elastic.

"People are starting to realize that to do agentic AI correctly, you have to have relevant data," Exner says. "Relevance is critical in the context of agentic AI, because that AI is taking action on your behalf. When people struggle to build AI applications, I can almost guarantee you the problem is relevance.”

Agents everywhere

The struggle could be entering a make-or-break period as organizations scramble for competitive edge or to create new efficiencies. A Deloitte study predicts that by 2026, more than 60% of large enterprises will have deployed agentic AI at scale, marking a major increase from experimental phases to mainstream implementation. And researcher Gartner forecasts that by the end of 2026, 40% of all enterprise applications will incorporate task-specific agents, up from less than 5% in 2025. Adding task specialization capabilities evolves AI assistants into context-aware AI agents.

Enter context engineering

The process for getting the relevant context into agents at the right time is known as context engineering. It not only ensures that an agentic application has the data it needs to provide accurate, in-depth responses, it helps the large language model (LLM) understand what tools it needs to find and use that data, and how to call those APIs.

While there are now open-source standards such as the Model Context Protocol (MCP) that allow LLMs to connect to and communicate with external data, there are few platforms that let organizations build precise AI agents that use your data and combine retrieval, governance, and orchestration in one place, natively.

Elasticsearch has always been a leading platform for the core of context engineering. It recently released a new feature within Elasticsearch called Agent Builder, which simplifies the entire operational lifecycle of agents: development, configuration, execution, customization, and observability.

Agent Builder helps build MCP tools on private data using various techniques, including Elasticsearch Query Language, a piped query language for filtering, transforming, and analyzing data, or workflow modeling. Users can then take various tools and combine them with prompts and an LLM to build an agent.

Agent Builder offers a configurable, out-of-the-box conversational agent that allows you to chat with the data in the index, and it also gives users the ability to build one from scratch using various tools and prompts on top of private data.

"Data is the center of our world at Elastic. We’re trying to make sure that you have the tools you need to put that data to work," Exner explains. "The second you open up Agent Builder, you point it to an index in Elasticsearch, and you can begin chatting with any data you connect this to, any data that’s indexed in Elasticsearch — or from external sources through integrations.”

Context engineering as a discipline

Prompt and context engineering is becoming a discipli. It’s not something you need a computer science degree in, but more classes and best practices will emerge, because there’s an art to it.

"We want to make it very simple to do that," Exner says. "The thing that people will have to figure out is, how do you drive automation with AI? That’s what’s going to drive productivity. The people who are focused on that will see more success."

Beyond that, other context engineering patterns will emerge. The industry has gone from prompt engineering to retrieval-augmented generation, where information is passed to the LLM in a context window, to MCP solutions that help LLMs with tool selection. But it won't stop there.

"Given how fast things are moving, I will guarantee that new patterns will emerge quite quickly," Exner says. "There will still be context engineering, but they’ll be new patterns for how to share data with an LLM, how to get it to be grounded in the right information. And I predict more patterns that make it possible for the LLM to understand private data that it’s not been trained on."

Agent Builder is available now as a tech preview. Get started with an Elastic Cloud Trial, and check out the documentation for Agent Builder here.

---

Sponsored articles are content produced by a company that is either paying for the post or has a business relationship with VentureBeat, and they’re always clearly marked. For more information, contact sales@venturebeat.com.

It’s 3:37 am on a Sunday in Los Angeles, and one of the leading financial services firms on the West Coast is experiencing the second week of a living-off-the-land (LOTL) attack. A nation-state cyberattack squad has targeted the firm’s pricing, trading and valuation algorithms for cryptocurrency gain. Using common tools, the nation state has penetrated the firm’s infrastructure and is slowly weaponizing it for its own gain.

According to CrowdStrike’s 2025 Global Threat Report, nearly 80% of modern attacks, including those in finance, are now malware-free, relying on adversaries exploiting valid credentials, remote monitoring tools and administrative utilities with breakout times (sometimes less than a minute).

No one in the SOC or across the cybersecurity leadership team suspects anything is wrong. But there are unmistakable signals that an attack is underway.

The upsurge in credential theft, business email compromise and exploit of zero-day vulnerabilities is creating the ideal conditions for LOTL attacks to proliferate. Bitdefender’s recent research found that 84% of modern attacks use LOTL techniques, bypassing traditional detection systems. In nearly 1 in 5 cases, attackers increasingly aided by automation and streamlined toolkits exfiltrated sensitive data within the first hour of compromise.

LOTL-based tactics now account for the majority of modern cyber intrusions, with advanced persistent threats (APTs) often lingering undetected for weeks or months before hackers exfiltrate valuable data, according to IBM’s X-Force 2025 Threat Intelligence Index.

The financial repercussions are staggering. CrowdStrike’s 2025 threat research puts the average cost of ransomware-related downtime at $1.7 million per incident, which can balloon to $2.5 million in the public sector. For industry leaders, the stakes are so high that security budgets now rival those of core profit centers.

Your most trusted tools are an attacker’s arsenal

"These are the tools that you cannot disable because your administrators are using them, your applications are using them, your [employees] are using them, but attackers [are using them, too]," Martin Zugec, technical solutions director at Bitdefender, said at RSAC-2025 earlier this year. "You cannot disable them because you will impact the business."

CrowdStrike’s 2025 report confirms that adversaries routinely exploit utilities such as PowerShell, Windows management instrumentation (WMI), PsExec, remote desktop protocol (RDP), Microsoft Quick Assist, Certutil, Bitsadmin, MSBuild and more to persist inside enterprises and evade detection. LOTL tools of the trade leave no digital exhaust, making it extremely difficult to spot an attack in progress.

“Threat actors increasingly exploit techniques such as bring your own vulnerable driver (BYOVD) and LOTL to disable endpoint detection and response (EDR) agents and conceal malicious activity within legitimate system operations," Gartner notes in a recent report. "By leveraging common OS tools, such as PowerShell, MSHTA and Certutil, they complicate detection and hide in the noise of EDR alerts."

CrowdStrike’s ransomware survey reveals that 31% of ransomware incidents begin with the misuse of legitimate remote monitoring and management tools, proving that even enterprise IT utilities are rapidly weaponized by attackers.

The documented realities in CrowdStrike's reports corroborate the industry's deeper research: The IT stack itself is now the attack vector, and those relying on traditional controls and signature-based detection are dangerously behind the curve.

Behavioral clues hiding in plain sight

Adversaries who rely on LOTL techniques are notorious for their patience.

Attacks that once required malware and attention-grabbing exploits have given way to a new norm: Adversaries blending into the background, using the very administrative and remote management tools security teams depend on.

As Bitdefender's Zugec pointed out: “We are mostly seeing that the playbook attackers use works so well they just repeat it at scale. They don’t break in, they log in. They don’t use new malware. They just use the tools that already exist on the network.”

Zugec described a textbook LOTL breach: No malware, no new tools. BitLocker, PowerShell, common admin scripts; everything looked routine until the files were gone and no one could trace it back. That’s where threat actors are winning today.

Adversaries are using normality as their camouflage. Many of the admins’ most trusted and used tools are the very reason LOTL attacks have scaled so quickly and quietly. Zugec is brutally honest: “It has never been as easy to get inside the network as it is right now.” What was once a breach of perimeter is now a breach by familiarity, invisible to legacy tools and indistinguishable from routine administration.

CrowdStrike’s 2025 Global Threat Report captures the scale of this phenomenon in numbers that should command every board’s attention. The reports’ authors write: “In 2024, 79% of detections CrowdStrike observed were malware-free [a significant rise from 40% in 2019], indicating adversaries are instead using hands-on-keyboard techniques that blend in with legitimate user activity and impede detection. This shift toward malware-free attack techniques has been a defining trend over the past five years."

The report’s researchers also found that breakout times for successful attacks continue to shrink; the average is just 48 minutes, the fastest 51 seconds.

Zugec’s advice for defenders working in this new paradigm is blunt and pragmatic. “Instead of just chasing something else, figure out how we can take all these capabilities that we have, all these technologies, and make them work together and fuel each other.” The first step: “Understanding your attack surface. Just getting familiar with how the attackers operate, what they do, not five weeks ago, but right now, should be the first step.”

He urges teams to learn what normal looks like inside their own environment and use this baseline to spot what’s truly out of place, so defenders stop chasing endless alerts and start responding only when it matters.

Take complete ownership of your tech stack now

LOTL attacks don’t just exploit trusted tools and infrastructures, they take advantage of an organizations’ culture and daily ability to compete.

Staying secure means making constant vigilance a core value, backed by zero trust and microsegmentation as cultural anchors. These are just the first steps. Consider the NIST Zero Trust Architecture (SP 800-207) as an organizational backbone and playbook to tackle LOTL head-on:

• Limit privileges now on all accounts and delete long-standing accounts for contractors that haven’t been used in years: Apply least-privilege access across all admin and user accounts to stop attackers from escalating.

• Enforce microsegmentation: Divide your network into secure zones; this will help confine attackers, limit movement and shrink the blast radius if something goes wrong.

• Harden tool access and audit who is using them: Restrict, monitor and log PowerShell, WMI and other utilities. Use code signing, constrained language modes and limit access to trusted personnel.

• Adopt NIST zero trust principles: Continuously verify identity, device hygiene and access context as outlined in SP 800-207, making adaptive trust the default.

• Centralize behavioral analytics and logging: Use extended monitoring to flag unusual activities with system tools before an incident escalates.

• Deploy adaptive detection if you have an existing platform that can scale and provide this at a minimal charge: Employ EDR/XDR to hunt for suspicious patterns, especially when attackers use legitimate tools in ways that sidestep traditional alerting.

• Red team regularly: Actively test defenses with simulated attacks and know how adversaries misuse trusted tools to penetrate routine security.

• Elevate security awareness and make it muscle memory: Train users and admins on LOTL methods, social engineering and what subtle signals betray compromise.

• Update and inventory: Maintain application inventories, patch known vulnerabilities and conduct frequent security audits.

Bottom line: The financial services firm referenced at the beginning of this story eventually recovered from its LOTL attack. Today, their models, the CI/CD process for AI development and gen AI R&D are managed by a team of cybersecurity managers with decades of experience locking down U.S. Department of Defense sites and vaults.

LOTL attacks are real, growing, lethal and require a new mindset by everyone in cybersecurity.

Learn how to use Pixel Watch 4’s Raise to Talk feature so you can access Gemini hands-free.

Google Arts & Culture is partnering with the National Trust for Historic Preservation to celebrate Route 66’s legacy

The AI for Math Initiative brings together five of the world's most prestigious research institutions.

The gap between Today and GMA stretched to 76,000 demo viewers.

Zohran Mamdani’s campaign is worth studying, not for its policies, but for its marketing.

The CBS news comes as John Dickerson is set to depart the network.

The company is planning on cutting around 2,000 roles overall.

Gut's Anselmo Ramos reviews the latest chapter of Apple's long-running 'Underdogs' series.

Plus, what happened at Vogue Singapore's Next in Vogue and Singularity South Africa.

The combined company creates one of the largest MVPDs in the market.

SerpAPI said it will “vigorously defend” itself after being sued by Reddit for allegedly scraping and reselling data from the platform via Google Search results.

The response. SerpAPI called Reddit’s language “inflammatory” and said it was “extremely disappointed” to learn of the lawsuit without prior communication.

• “Our work is guided by a simple principle: public search data should be accessible,” the company said.
• SerpAPI argued its position is backed by the First Amendment and called Reddit’s actions a threat to “the free and open Web we all enjoy.”

What they’re saying. According to SerpApi:

• “For eight years, SerpApi has operated transparently and lawfully, helping countless developers, researchers, and businesses build on top of publicly available search data. Our technology is used across multiple industries, from SEO, marketing, and advertising to copyright verification, background checks, news monitoring, and now AI. Our work is guided by a simple principle: public search data should be accessible. And accessibility cannot exist without clean unified data structures, speed, and automation possibilities.”

Catch up quick. Reddit sued SerpAPI, Perplexity, Oxylabs, and AWMProxy last week, claiming they scraped Reddit content from Google results “at an industrial scale” and hid their identities to bypass restrictions. Reddit:

• Claimed it set a “trap” for Perplexity to prove scraping.
• Is seeking financial damages and a ban on further data use.

Zoom out. Reddit licenses its data to OpenAI and Google. Meanwhile, Google and Reddit are reportedly exploring a deeper AI partnership that could bring Reddit discussions directly into AI Overviews and other Google experiences.

Why we care. This fight isn’t just about data – it’s about control. The big tech platforms are battling over who owns the information that powers search results and AI-generated answers, while brands struggle to understand what’s driving rankings, visibility, and attribution.

SerpApi’s statement. Our Response to Reddit, Inc. v. SerpApi, LLC: Defending the First Amendment

With Black Friday, Cyber Monday, and the holiday season fast approaching, you’re probably knee-deep in holiday account planning.

Today, we’re zeroing in on a Google Ads feature that you can use all year ‘round, but may be particularly useful to you and your customers during the holidays: Promotion assets.

What are promotion assets in Google Ads?

Promotion assets (formerly known as promotion extensions) are an optional addition to your Search or Performance Max campaigns. They allow you to highlight special deals, sales, and discounts your business is currently running, alongside your standard ad headlines and descriptions.

Promotion assets can show on Google Search, and on Google Maps if you have a local promotion offer.

Why should you use promotion assets in Google Ads rather than editing your ad headlines?

While you absolutely can edit your search ad text to include promotional messaging, promotion assets offer a few unique benefits. Specifically, they:

• Stand out in search results: Promotion assets may be bold or have a box around them, so they can really capture a user’s attention and increase your click-through rate.
• Allow for flexibility: You can highlight deals and offers without constantly editing and re-submitting your core ad text. You can even schedule them to go live (and turn off) in advance, which saves you lots of time.

Why can’t you use promotion assets with Shopping campaigns in Google Ads?

To show special deals and sales with your Shopping ads, you need to set that up in Google Merchant Center rather than Google Ads. These are called Merchant Promotions, and they run through a separate system.

If you’re an ecommerce business, head over to Merchant Center to set up your deals and promotions.

How to create promotion assets in Google Ads

To create a new promotion asset, go to Assets > Assets on the left-side navigation in Google Ads, tap the + button, and choose Promotion.

You can add promotion assets at the account, campaign, or ad group level.

If you define multiple assets across the various levels, Google will use the most specific level available. For example, ad group-level promotion assets would be selected over campaign-level promotion assets.

Which holidays allow for promotion assets?

Part of creating a promotion asset is selecting an “occasion” for your promotion. While you must select from Google’s list of occasions, there is a wide variety of options.

Some are seasonal (like Halloween or Valentine’s Day), but many are generic and flexible, allowing you to run promotions year-round. For example:

• Fall Sale
• Winter Sale
• Spring Sale
• Summer Sale

This means that although you need an “occasion,” promotion assets are not just for big tentpole holidays. Any kind of business can have promotions, sales, and discounts throughout the year.

What additional details are required to use promotion assets?

To ensure your promotion assets get approved, and provide a good user experience, you need to pay close attention to several “fiddly” details. Inconsistency between your asset and your landing page is a common reason for disapproval.

Here’s a checklist of some of those details:

• Language and Currency
• Discount Type: For example
  • Monetary Discount: e.g., “$20 off”
  • Percent Discount: e.g., “20% off”
  • Up to Discount: e.g., “Up to $100 off”
• Item Text (20 Characters): This is where you specify the actual product or service on sale. Keep it brief and relevant, e.g., “shoes” or “house cleaning.”
• Final URL: This link must take the user directly to the specific promotion for that specific item. Do not link to a generic homepage.
• Promo code (optional)
• Minimum order value (e.g., “Valid on orders over $50”) (optional)
• Terms and Conditions (optional)

Three types of scheduling for promotion assets

Promotion assets give you three layers of scheduling control. While we appreciate Google’s thoroughness, the use cases can be a bit confusing! Here’s what you need to know:

1. Promotion Scheduling: the dates your promotion is active

Understandably, the promotion asset is only eligible to show when the promotion is actually running. This means you cannot advertise a promotion before it has started or after it has ended.

2. Asset Scheduling: the dates when your promotion asset can run

This controls when your promotion asset will be eligible to run, which may or may not be the same as the actual promotion dates.

• For example, if your business is running a sale for six weeks but only wants to advertise it via Google Ads during the last two weeks, you can set the promotion asset to be live only during those final 14 days.

3. Ad Scheduling: time of day and day of week

You can set additional scheduling rules to show your promotion asset only on specific days or times of the week (e.g., only during evenings or on weekends).

This article is part of our ongoing Search Engine Land series, Everything you need to know about Google Ads in less than 3 minutes. In each edition, Jyll highlights a different Google Ads feature, and what you need to know to get the best results from it – all in a quick 3-minute read.

💾

Cybersecurity researchers are calling attention to a spike in automated attacks targeting PHP servers, IoT devices, and cloud gateways by various botnets such as Mirai, Gafgyt, and Mozi. "These automated campaigns exploit known CVE vulnerabilities and cloud misconfigurations to gain control over exposed systems and expand botnet networks," the Qualys Threat Research Unit (TRU) said in a report

Cybersecurity researchers have flagged a new security issue in agentic web browsers like OpenAI ChatGPT Atlas that exposes underlying artificial intelligence (AI) models to context poisoning attacks. In the attack devised by AI security company SPLX, a bad actor can set up websites that serve different content to browsers and AI crawlers run by ChatGPT and Perplexity. The technique has been

These digital photo frames both blew me away when I tested them – but which should you buy come Black Friday?

A major Microsoft outage appears to be taking place - here's what we know.

Save on the world’s number-one mobile charging brand

At least 10 npm packages were uploaded in early July 2025 and were downloaded almost 10,000 times.

Creative Cloud gets full suite of Google AI models, while Premiere mobile app has a dedicated YouTube mode.

The Garmin Epix Pro is actually the best value Garmin Watch you can buy, thanks to some massive discounts.

OpenAI plans to introduce major improvements to the world of AI in 2026 and beyond, starting with AI research interns.

The Lenovo Idea Tab Pro is top of our 'best cheap tablets' guide, and it's on sale now ahead of Black Friday.

There's a The Simpsons Fortnite event coming this weekend. Here's the start time, and what's been revealed so far.

Save up to 63% across their hauntingly popular range

Sure, the iPhone 16e is cheap, but at AO’s early Black Friday price, I’d go for the iPhone 16 instead.

Taxfix recently conducted a survey of people who file Self Assessment tax returns in the UK and found that 59% plan to use AI tools to help them complete their returns ahead of HMRC’s 31 January deadline.

Grab one of the best password managers for significantly less, way before Black Friday.

AI ROI requires strategy, adaptation, skilled workforce, and industry focus.

EDR tools aren't scanning Linux sandbox environments on Windows, giving cybercriminals a unique opportunity to work undetected.

Looking for Quordle clues? We can help. Plus get the answers to Quordle today and past solutions.

Looking for NYT Connections answers and hints? Here's all you need to know to solve today's game, plus my commentary on the puzzles.

Looking for NYT Strands answers and hints? Here's all you need to know to solve today's game, including the spangram.

Dragon Quest I & II HD-2D Remake takes two classic JRPGs and transports them into the modern era with finesse and grace.

Netflix has teased changes to profiles for kids, interactive voting, themed homepages, vertical video, and more.

This Call of Duty pop-up lets you play Black Ops 7 one last time ahead of launch.

I guess home theater and vacuum cleaners are all about making air move…

The 1X Neo Home robot is a cleaning, dancing, listening, talking, and potentially helpful home assistant that is now on preorder, ready to march into your home and even plug itself into a wall outlet when necessary.

Anti-DEI requirements pose ‘enormous, open-ended financial risks’, Python Software Foundation argues.

We're breaking down all the key differences when it comes to the SteelSeries Arctis Nova Elite vs SteelSeries Arctis Nova Pro Wireless matchup.

The latest VanTrue Nexus 5S dash cam comes complete with four lenses, all of which deliver high-quality video.

PayPal and OpenAI sign an agreement to embed the payments wallet into ChatGPT.

Mercor CEO Brendan Foody has built a $10 billion empire freeing up valuable data from legacy industries and making it available to AI labs.

With a market cap of $5 trillion, Nvidia is now worth more than the aggregated stock markets of all countries, apart from the United States, China, and Japan.

Kruti Patel Goyal is replacing Etsy's long-time CEO, Josh Silverman, who is stepping down by the end of the year.

This is the third and final day of TechCrunch Disrupt 2025 at Moscone West in San Francisco. Register here to get a 50% discount and don't miss out on innovation and scaling.

The U.S. inked Technology Prosperity Deals (TPD) with Japan and South Korea with an eye toward spurring collaboration on AI, semiconductors, quantum computing, biotech, space, 6G, and other technologies.

Grammarly is renaming itself as Superhuman after acquiring the Superhuman email client in July.

Key updates include the introduction of QR codes that let you identify and shop for items in tagged videos, AI-powered upscaling, and improved search.

After facing lawsuits and public outcry following the suicides of two teenagers, Character.AI says it's making changes to its platform to protect children, changes that could affect the startup’s bottom line.

It’s rare for a tech executive to cue up a video mocking themselves — but that’s just what Microsoft President Brad Smith did on Tuesday at the Cascadia Innovation Corridor conference in Seattle. Smith played a clip from The Daily Show in which comedian Jon Stewart lampooned his and Microsoft CEO Satya Nadella’s interviews about AI’s impact on jobs.

The segment poked fun at the idea that displaced workers might become “prompt engineers” — a new job Stewart rebranded as “types questions guy.”

It was a self-aware feature of a talk that balanced enthusiasm for artificial intelligence’s potential with sober reflections on its hype and potential pitfalls.

The Microsoft leader called AI the “next great general purpose technology” on par with electricity. He said AI will transform sectors including health, education, biotech, aerospace, agriculture, climate and others.

That was a theme during Tuesday’s event. Former Washington Gov. Chris Gregoire, who leads the Cascadia Innovation Corridor group, kicked off the day by calling AI “a defining technology of our generation.”

Smith, who in his three decades at Microsoft has witnessed tech bubbles and bursts, also offered a “breadth of perspective” on AI that he hinted might be lacking in Silicon Valley.

“In so many ways, the sky is the limit,” Smith said. “That is exciting, but I don’t want to just be another tech bro who says, ‘Hey, great, here it comes. Get ready, get out your wallet.'”

AI-driven employment threats are becoming increasingly real in the tech sector and beyond. Amazon on Tuesday announced a huge round of layoffs, slashing 14,000 corporate and tech jobs. Earlier this year Microsoft laid off 15,000 employees worldwide. The cuts aren’t all tied to AI, but many executives are talking about worker efficiency gains thanks to the tech.

Despite the recent layoffs, many industry and elected leaders in the Cascadia region, which stretches from Vancouver, B.C., through Seattle and down to Portland, see AI as a promising economic engine that can build on the area’s strong tech foundation. That includes Microsoft and Amazon as well as a growing slate of AI startups, plus institutions such as the University of Washington, University of British Columbia, Allen Institute for AI and others.

But Smith — who manages to strike a persona blending tech evangelist, politician and favorite uncle — also acknowledged concerns about disparities in AI access, whether looking locally at rural versus urban divides, or the gap between AI use in affluent and low-income countries that lack widespread electricity and internet connections.

He also tackled the meta questions around the responsible use of AI and encouraged society to get out in front of the technology with appropriate guardrails.

“What are we trying to do as an industry, as a region, as a planet, as a species? Are we trying to build machines that are better than people? Are we trying to build machines that will help people become smarter and better?” he asked.

“If the experience that we’ve all had with social media over the last 15 years teaches us anything at all,” Smith continued, “it is that the best time to ask these questions and to debate them is before technology answers them for us.”

RELATED: Cascadia’s AI paradox: A world-leading opportunity threatened by rising costs and a talent crunch

A YouTuber's videos telling people how to use Windows 11 without a Microsoft account and how to install it on unsupported hardware were allegedly violating community guidelines on dangerous and illegal activities.

GTC announcements drive Nvidia's market capitalization by over $200 billion in one day, making the company the world's first firm to reach a market capitalization of over $5 trillion.

China has launched what it claims is the world’s first wind-powered undersea data center — a sealed, seawater-cooled cluster sitting 35 meters beneath the surface off the coast of Shanghai.

Someone has explained how they used AI to slash hospital bills from hundreds to tens of thousands of dollars.

Save on some of the best 3D printers in Bambu Lab's Black Friday sale

NZXT and partner Flex are being brought to court following claims that Flex's rental program is a scam.

ww

The post El-Erian Warns of AI’s “Rational Bubble” and Unaddressed Risks appeared first on StartupHub.ai.

“Some AI names will end up in tears,” declared Mohamed El-Erian, Allianz’s chief economic advisor, during a recent segment on CNBC’s ‘Squawk on the Street.’ His candid assessment cut through the prevailing market euphoria, offering a nuanced perspective on the artificial intelligence revolution. El-Erian, a former PIMCO CEO, engaged with the program’s hosts, providing commentary […]

The post El-Erian Warns of AI’s “Rational Bubble” and Unaddressed Risks appeared first on StartupHub.ai.

The post AI’s Unprecedented Infrastructure Revolution: Scarcity, Specialization, and a Cultural Reset appeared first on StartupHub.ai.

“I’ve seen nothing like this. I’m fairly certain no one’s seen anything like this. The Internet in the late 90s, early 2000s was big… this makes it… 10x is an understatement. It’s 100x what the Internet was.” This stark assessment by Amin Vahdat, VP and GM of AI and Infrastructure at Google, encapsulates the central […]

The post AI’s Unprecedented Infrastructure Revolution: Scarcity, Specialization, and a Cultural Reset appeared first on StartupHub.ai.

The post AI for Math Initiative: DeepMind’s Bold Research Play appeared first on StartupHub.ai.

Google DeepMind's AI for Math Initiative unites top research institutions to pioneer AI in mathematical discovery, promising significant breakthroughs and accelerated scientific progress.

The post AI for Math Initiative: DeepMind’s Bold Research Play appeared first on StartupHub.ai.

The post Outlier Talent and the Unscripted Symphony of Innovation appeared first on StartupHub.ai.

“Outlier talent will always be outlier talent,” declared Alex Pall, half of the Grammy-winning duo The Chainsmokers and co-founder of Mantis VC, an observation that cuts directly to the heart of creativity in an age increasingly shaped by artificial intelligence. This potent statement, emerging during a wide-ranging conversation with Jack Altman on the *Uncapped* podcast, […]

The post Outlier Talent and the Unscripted Symphony of Innovation appeared first on StartupHub.ai.

The post Mitrione: AI is alive and well and continues to power the market appeared first on StartupHub.ai.

AI’s Ascendancy and Market Concentration “AI is alive and well and continues to power the market,” stated RaeAnn Mitrione, Investment Management Partner at Callan Family Office, during her discussion with Frank Holland at the CNBC studio. The recent surge of NVIDIA past a $5 trillion market capitalization underscores the pervasive influence of artificial intelligence on […]

The post Mitrione: AI is alive and well and continues to power the market appeared first on StartupHub.ai.

The post AI layoffs hit Big Tech: Here’s what to know appeared first on StartupHub.ai.

Partsinevelos, speaking with Andrew Ross Sorkin, framed the discussion around a central query: are the reported AI layoffs in Big Tech genuine, or are they serving as a convenient “scapegoat” for other underlying issues in the employment landscape? She highlighted recent job cuts across major tech companies—Amazon, Applied Materials, Meta, and Google Cloud—some explicitly attributing […]

The post AI layoffs hit Big Tech: Here’s what to know appeared first on StartupHub.ai.

The post Impala AI Targets LLM Inference Costs with $11M Seed appeared first on StartupHub.ai.

Impala AI has secured $11 million to launch an AI stack that promises to cut large language model inference costs by up to 13x for enterprises.

The post Impala AI Targets LLM Inference Costs with $11M Seed appeared first on StartupHub.ai.

The post Reflectiz raises $22M to advance web exposure management appeared first on StartupHub.ai.

Web exposure management platform Reflectiz raised $22 million to help enterprises secure their websites from risks posed by third-party scripts and open-source components.

The post Reflectiz raises $22M to advance web exposure management appeared first on StartupHub.ai.

The post The AI Infrastructure Gold Rush: Opportunities, Risks, and Strategic Moats appeared first on StartupHub.ai.

The burgeoning artificial intelligence landscape, characterized by unprecedented infrastructure buildout and escalating capital expenditure, is creating both immense opportunities and significant strategic challenges for the world’s leading tech companies. This dynamic was a central theme in a recent CNBC discussion where Alex Kantrowitz, Founder of Big Technology and CNBC contributor, provided incisive commentary on Nvidia’s […]

The post The AI Infrastructure Gold Rush: Opportunities, Risks, and Strategic Moats appeared first on StartupHub.ai.

More than 117 years after Seattle residents first saw a flying machine in the sky, a unique aircraft over Jet City can still turn heads.

That happened this week with the arrival of Beta Technologies‘ all-electric ALIA CX300 conventional takeoff and landing aircraft as it dropped into King County International Airport – Boeing Field.

Photographer Steve Rice captured the strange-looking airplane with a rear propellor and posted images on Reddit, where aviation geeks launched into a debate about e-planes, range, charging times, vertical takeoff and landing aircraft, and more.

Vermont-based Beta wasn’t just doing a fly-by. The company brought the plane to Seattle for an official demonstration of the ALIA in an event that drew state officials, aviation experts, and industry leaders from across Washington.

In a news release Tuesday, Beta said Washington has a “deep-rooted aviation heritage that has long positioned the state as a global leader in aerospace innovation and manufacturing.” And the company said the state is now “actively advancing the future of flight through strategic investments in sustainable aviation and the critical infrastructure needed to support next-generation technologies.”

Beta founder and CEO Kyle Clark called the event at Boeing Field “a step toward realizing a future where electric aviation is accessible, reliable, and benefits local communities.” 

Founded in 2017, Beta is building two electric aircraft — the fixed-wing ALIA CTOL, and the ALIA VTOL, a vertical takeoff and landing aircraft — at a production facility in Vermont.

The inaugural flight of Beta’s first production model airplane was last November. The ALIA CTOL has a range of 336 nautical miles, and Beta’s planes are designed to carry passengers or cargo.

The company has also developed and is rolling out a network of charging infrastructure for use across airports and the electric aviation ecosystem.

Beta filed for an initial public offering earlier this month with plans to sell 25 million shares at $27 to $33 each — a price range that could value the company at $7.2 billion.

Vancouver, B.C.-based Helijet International previously placed orders with Beta for a fleet of eVTOL aircraft.

Other electric and hybrid aircraft makers are getting their planes off the ground in Washington, including Seattle-based Aero-TEC and Everett, Wash.-based magniX. Arlington, Wash.-based Eviation Aircraft paused work on its Alice airplane earlier this year.

With a new OpenAI partnership in hand, Microsoft is going into its earnings report Wednesday afternoon with a resolution to one of the biggest questions about its business.

The company’s market value reached $4 trillion again as Wall Street reacted to the details of the new Microsoft-OpenAI agreement, which gives Microsoft a 27% equity stake in OpenAI’s new for-profit entity, and a commitment for $250 billion in cloud purchasing by the ChatGPT maker.

Analysts expect the tech giant to report another strong quarter, fueled primarily by continued momentum in its Azure cloud business and growing adoption of its Copilot AI tools.

Quarterly revenue is expected to be about $75.4 billion for the first quarter of Microsoft’s 2026 fiscal year, which ended Sept. 30, according to numbers tracked by Yahoo Finance. That would represent a 15% jump compared to the $65.6 billion reported in the same period last year. 

Analysts expect earnings per share of $3.66, up about 11% year-over-year from $3.30.

Investors will be paying close attention to the growth rate in Microsoft’s Azure cloud business, with some analysts expecting as much as 39% growth (in constant currency, excluding the impact of exchange rates). Hitting this mark would exceed the company’s prior guidance and maintain the 39% growth pace set in the previous quarter.

Yet the potential for an AI bubble will no doubt be the focus of questions on company’s earnings conference call. Amid surging investment and growing valuations in the AI sector, some analysts and tech leaders are warning that the enthusiasm could outpace the business realities

Microsoft and Google parent Alphabet will both report numbers on Wednesday afternoon, and Amazon the following day, making for quick comparisons across the major cloud platforms.

As of the most recent quarter, ended in June, Microsoft reported more than $75 billion in annual Azure revenue for its just-ended fiscal year, compared to an annual run rate that had surpassed $50 billion for Google Cloud and a run rate of nearly $124 billion for Amazon Web Services (based on its $30.9 billion revenue in the June quarter).

Check back with GeekWire on Wednesday afternoon for full coverage.

A trio of veteran entrepreneurs have joined forces to create a new Seattle startup — and any mentions of the company across the internet will likely be tracked by what they’re building.

Alertmouse generates email alerts for people and brands who want to know what’s being said about them — or anything else they’re interested in — online. The goal is to provide a better offering than other monitoring tools, most notably Google Alerts, which Alertmouse calls “so bad it might as well not exist.”

The startup was created by co-founders Rand Fishkin (SparkToro, Snackbar Studio), Adam Doppelt (Urbanspoon, Dwellable), and Nathan Kriege (Blueprint AI, Fresh Chalk).

Fishkin, the CEO, posted about Alertmouse on LinkedIn this week, saying that while the side project has turned into a full-fledged business, he’s not leaving his other two jobs.

“It doesn’t take a ton of my time but it has been really fun to build this thing that I desperately needed,” Fishkin said in a video on his post, before listing his grievances with Google Alerts, including how it “doesn’t pick up everything you want, it sends you useless alerts” and more.

Fishkin said tracking his mentions or those related to his companies, whether it’s in a news article or in a Reddit thread, allows him to monitor what’s being said and jump in if necessary to reply.

Alertmouse says it searches the index of websites and pages it can reach two or three times each day for the unique string of words/phrases/rules a user has entered. An email is sent with the pages that contained them.

“It’s not rocket science, but it takes a lot of clever programming, testing, and iteration to make a good alert service,” the company says in its FAQ.

In an interview with GeekWire this week, Fishkin said there are enterprise tracking tools that do what Alertmouse does, such as Mentioned, Hootsuite, and Brandwatch, but they can be cost-prohibitive.

“Google Alerts has been this free alternative for a long time, but sometime in the last decade, maybe even before that, it just stopped sending me anything decent,” he said. “I have no idea what they’re doing under the hood. I suspect it’s a defunct product that no one maintains anymore, but I couldn’t tell you what’s really going on.”

On a website loaded with cheesy puns, Alertmouse has four pricing tiers, including Nibble (free), Slice ($120/year), Wedge ($600/year), and Wheel ($1,200/year).

Alertmouse attracted 1,000 sign-ups in the first several hours it was live, and Fishkin credits the fun interface and language on the website, and the fact that it’s easy to use.

“We wanted to make a brand that no one could confuse for AI,” Fishkin laughed. “This is not an AI company. There’s going to be no venture capital, there’s no AI under the hood. It’s just really simple, straightforward, fun, delightful humans.”

Fishkin, an SEO expert who founded and led Moz, a Seattle-based maker of marketing software tools, co-founded SparkToro in 2018. The audience research tool helps marketers and others understand their target audiences. He raised $2.15 million last year for his new independent video game studio, Snackbar Studio.

Doppelt and Kriege previously worked on vacation rental startup Dwellable (sold to HomeAway in 2015), local professional recommendation site Fresh Chalk, and task management company Blueprint AI together. Last year they teamed up to create a resource website for everything you’d ever want to know about smoke detectors.

Fishkin and Doppelt are also part of dedicated group of “Dungeons & Dragons” players.

The Alertmouse website says the startup has no plans to hire. But that could change after a morning in which lots of people were emailing with questions.

“If it keeps going like this we might have to bring someone on,” Fishkin said.

Off the coast of Shanghai, a revolutionary structure is forming beneath the waves: the world’s first wind-powered, commercial-scale underwater data center (UDC). Basically, submerging servers in pressure-resistant pods (similar to Microsoft's non-commercial Project Natick in 2018) is a direct response to the escalating power and cooling demands

NVIDIA held its annual Graphics Technology Conference (GTC) in Washington D.C. yesterday, and as a surprise showing in the middle of his keynote, company CEO Jensen Huang pulled out a Vera Rubin Superchip, marking the first time that this product was shown to the public. The part looks quite different from even the GB300 Blackwell Ultra Superchip,

We're already big fans of Bose's audio products, so it doesn't take much arm twisting to convince us to consider the brand if we're in the market for a new set of headphones or earbuds. We won't turn down a good deal, either. It just so happens that you can have both right now, with Woot serving up some massive discounts on Bose's QuietComfort

13-hour flights between Tokyo and New York may soon be relegated to history books, as a major Japanese travel agency announced plans to launch a point-to-point transport service that could connect the two cities in just 60 minutes via outer space. The audacious plan (set to launch in the 2030s) comes from a partnership between the Nippon Travel

For Google loyalists, the first purported look at the Pixel 10a could ring a familiar tune rather than drastically advance the bar. Not expected to debut until early 2026, the well-regarded mid-budget handset has been revealed in CAD renders and the consensus is clear: Google is playing it safe, delivering a device that is essentially a Pixel

A new wave of flagship solid state drives (SSDs) that nearly saturate the available PCI Express 5.0 bandwidth available to x4 NVMe storage drives is starting to be appear. In back-to-back announcements, both PNY and Corsair unveiled new lightning-fast Gen 5 models—CS3250 and MP700 Pro XT, respectively—that are rated to deliver sequential read

Uber has already changed the taxi game with its, ahem, uber-popular ride sharing service, and now NVIDIA is hopping into the driver's seat on what figures to be the next evolution—a fleet of autonomous robotaxis built on NVIDIA's new DRIVE AGX Hyperion 10 platform. To make this happen, NVIDIA and Uber are teaming up to scale what's being pitched

Believe it or not, this could actually be a very big breakthrough.

The company also said its next-gen console would be backward compatible.

Prominent AMD Radeon Gallium3D driver developer Marek Olšák just changed the RadeonSI driver's default from the AMDGPU LLVM shader compiler back-end over to the ACO back-end initially developed by Valve. This should lead to better performance and quicker shader compilation and in turn faster game loads...

On Monday the AMD Radeon AI PRO R9700 officially arrived at Internet retailers and is successfully selling at the $1299 price point. Some models have sine sold out but as of writing two days later some Radeon AI PRO R9700 graphics cards remain available at that competitive price point. On Monday I provided some initial benchmarks of the AMD Radeon AI PRO R9700 for vLLM AI inferencing with more AI benchmarks on the way... While the craze is all about AI in 2025, the Radeon AI PRO R9700 does work for other non-AI workloads too and in this article is a look at its competitive OpenCL performance with great value compared to the NVIDIA RTX competition.

In addition to talking up the openSFI firmware collaboration between AMD and Intel at the OCP Global Summit 2025, AMD engineer Raj Kapoor provided a status update on the company's much anticipated openSIL effort for working to ultimately replace AGESA with a new open-source CPU silicon initialization codebase...

Intel Compute Runtime 25.40.35563.4 is out today as the newest update to their open-source GPU compute stack for Level Zero and OpenCL on Intel integrated/discrete graphics hardware...

Inseego Corp., a global leader in 5G mobile broadband and 5G fixed wireless access (FWA) solutions, today introduced a new approach to enterprise FWA with a completely new 5G hardware platform, the Inseego Wavemaker FX4200, and updated software suite, Inseego Connect. Designed to bridge the gap between performance and ease-of-use for enterprise wireless networks, this innovative approach to FWA pairs the power and functionality of enterprise network solutions with the simplicity and ease-of-management of small- and medium-sized (SMB)-oriented solutions.

"In order to take advantage of the power of 5G for business, organizations have been forced to choose between feature-heavy solutions that can be complicated and expensive to deploy, and simplistic products that can't scale or meet business needs," said Juho Sarvikas, CEO of Inseego. "With the FX4200, X700 mesh Wi-Fi, and the innovative Inseego Connect software, we're eliminating that tradeoff. We've built the solution that the market has been asking for, and it will propel growth in FWA and 5G use in business."

Inno3D today rolled out the GeForce RTX 5060 LP, a low-profile graphics card capable of maxed out gaming at 1080p, including with ray tracing. The card is strictly half-height (low-profile), and is 2 slots thick. It uses an aluminium fin-stack heatsink with a tightly packed fin-stack sitting on top of a copper baseplate, with copper heat pipes spreading heat across. This heatsink is ventilated by three 50 mm fans.

The card draws power from a single 8-pin power connector, and display outputs include two standard DisplayPort 2.1b and one full-size HDMI 2.1b. The power connector points toward the tail end, to ensure clearance along the top edge. In all, the Inno3D RTX 5060 LP measures 17.8 cm x 6.9 cm x 4.1 cm (WxHxD). It sticks to NVIDIA-reference clock speeds, boosting up to 2497 MHz. Based on the 5 nm GB206 silicon and the Blackwell graphics architecture, the RTX 5060 offers 3,840 CUDA cores, 30 RT cores, 120 Tensor cores, 120 TMUs, and 48 ROPs. It comes with 8 GB of 28 Gbps GDDR7 memory across a 128-bit wide memory bus. Inno3D did not announce pricing, and availability for now is limited to the Chinese market.

GlobalFoundries (GF) today announced plans to invest €1.1 billion to expand its manufacturing capabilities at its Dresden, Germany site. The investment will enable a production capacity increase to more than one million wafers per year by the end of 2028, making it the largest site of its kind in Europe.

The expansion, known as project SPRINT, is expected to be supported by the German federal government and the State of Saxony under the framework of the European Chips Act, with EU approval for the full program expected later this year. This investment underscores Saxony's role as a critical hub for semiconductor manufacturing and innovation and reinforces Europe's strategic goal of supply chain resilience.

NVIDIA's stock jumped 5% on Tuesday, closing at a record high and pushing the company's valuation to $4.89 trillion, just shy of the $5 trillion milestone. The surge followed a wave of new announcements at its GTC event in Washington, D.C., covering AI, supercomputing, and major industry partnerships. NVIDIA's shares have risen more than 50% year-to-date, more than doubling since April, fueled by demand for GPUs powering cloud data centers from Amazon, Google, and Microsoft. The company has also invested up to $100 billion in OpenAI, one of its largest customers. In his keynote, CEO Jensen Huang said that it is expecting $500 billion in GPU sales by the end of 2026, a realistic target if we consider that NVIDIA generated over $100 billion in revenue in just the first two quarters of this year.

During the GTC event NVIDIA revealed plans with the U.S. Department of Energy to build seven new supercomputers including one powered by 10,000 Blackwell GPUs. NVIDIA also introduced NVQLink, a new open systems architecture designed to accelerate the development of quantum supercomputers. Moreover, NVIDIA yesterday showcased its highly anticipated "Vera Rubin" Superchip, a single package combining two Rubin GPUs with a Vera CPU featuring 88 cores and 176 threads. While competition from AMD, Qualcomm, and in-house chips developed by hyperscalers continues to grow, NVIDIA remains positioned at the center of the AI infrastructure boom edging closer to the $5 trillion milestone valuation that seems set to become reality in a matter of days.

Update 18:31 UTC: NVIDIA's finance team has clarified that CEO Jensen Huang's reference to "$500 billion in GPU sales by the end of 2026" during his GTC keynote was misinterpreted. The figure refers to cumulative sales from the Blackwell and Rubin platforms over roughly two years—through calendar 2026 or fiscal 2027—not over the next five quarters. Analysts confirmed the correction following discussions with CFO Colette Kress's team. NVIDIA also reiterated its long-term goal of reaching $1 trillion in annual revenue later this decade. However, the revised guidance remains well above Wall Street expectations and does not yet consider the potential China-related sales, which could boost future figures.

Update 15:42 UTC: NVIDIA just crossed $5 trillion milestone, reaching $5.14 trillion valuation. A quick reminder, the company was valued at $10 billion ten years ago. It then hit the $1T milestone in May 2023, $2T in March 2024, $3T in June 2024, and $4T this July.

HighPoint Technologies, Inc., a leading manufacturer of high-performance PCIe storage and connectivity solutions, has unveiled the newest member of its industry leading PCIe Switch Solutions - the Rocket 1624A, a compact, high-efficiency variant of the acclaimed PCIe Gen 5 x16 Rocket 1628A Switch Adapter. The Rocket 1624A redefines PCIe scalability by combining dedicated NVMe storage connectivity with full-bandwidth PCIe expansion capability for GPUs, specialized accelerators, and other high-performance peripherals - all through a cost-effective Gen 5 switching platform.

Priced competitively at USD $899, the Rocket 1624A is designed for system integrators, datacenter administrators, and solution providers seeking to unlock the full potential of PCIe Gen 5 connectivity by enhancing the versatility and value of the target computing platform.

Acer America today debuted the premium Acer Chromebook Plus Enterprise 714 laptop line designed for businesses and organizations that need the latest technology to work in the cloud more efficiently and securely. In addition, the new Acer Chromebook Plus Enterprise 514 is now shipping in North America.

The new Chromebook Plus Enterprise models come with the business capabilities of ChromeOS unlocked, ensuring best-in-class security, simple management, flexible access, and enhanced administrative support. ChromeOS has built-in Google AI features that help employees do their best work, whether it's creating compelling content, leading meetings, and more. In addition, the new Chromebooks feature a Quick Insert key that encourages creativity and efficiency by providing one-touch access to tools, menus, and other applications.

NVIDIA yesterday had a GTC conference in Washington D.C., showing its latest "Vera Rubin" Superchip symbiote. Pictured for the first time is the combination of two "Rubin" GPU paired with a single "Vera" CPU carrying 88 custom NVIDIA cores and 176 threads in a single package. NVIDIA quoted performance targets of roughly 50 PetaFLOPS of FP4 compute per Rubin GPU, which yields about 100 PetaFLOPS FP4 for the two-GPU Superchip. The company said engineering samples are already moving through labs and set mass production goals for 2026 with broader shipments and deployments into 2027.

Each Rubin GPU appears to integrate two reticle-sized compute chiplets (2x830mm²?) paired with eight HBM4 stacks, delivering about 288 GB of HBM4 per GPU and roughly 576 GB of HBM4 on the full Superchip. NVIDIA also populated the board with SOCAMM2 LPDDR5X modules to provide large, low-latency system memory, with some older briefings indicating around 1.5 TB of LPDDR5X per Vera CPU on typical trays. The Vera CPU itself uses an 88-core, 176-thread Arm-based custom design and shows signs of a multi-chiplet layout with a distinct I/O chiplet nearby. With "Grace," NVIDIA relied on Arm's Neoverse design, but with Vera, the design team brought this CPU core in the house to extract maximum performance. Additionally, NVLink bandwidth climbs to approximately 1.8 TB/s to sustain heavy CPU-to-GPU traffic in system-demanding workloads such as AI inference and training.

ViewSonic Corp., a leading global provider of visual and edtech solutions, is celebrating the 2025 holiday season with a dynamic showcase of displays at the Pepcom Holiday Spectacular!, October 29th in New York City. The latest offerings from ViewSonic deliver powerful performance, sleek portability, and simplified connectivity, ideal for content creators, gamers, and remote working professionals.

This year's collection of displays features high-resolution desktop monitors, ultra-portable OLED displays, and projectors for work and entertainment needs. ViewSonic is debuting the VP3276T-4K, a 32-inch ColorPro monitor that is Pantone Validated and features a Thunderbolt 4 docking station. Also being demonstrated is the lightweight M1X projector that is part of the AtmosKIT Autoplay GO Bundle and festive decoration kit to create a showstopping holiday experience.

GMKtec has introduced the NucBox M7 Ultra compact Mini PC, the system is powered by AMD's Ryzen 7 PRO 6850U, an 8-core, 16-thread Zen 3+ processor built on TSMC's 6 nm process, paired with Radeon 680M graphics featuring 12 CUs running at 2200 MHz. The NucBox M7 Ultra supports up to 64 GB of DDR5-4800 memory and includes PCIe 4.0 M.2 storage with up to 16 TB of total expansion (2× M.2 2280). Display output options include HDMI 2.1, DisplayPort 2.0 and dual USB4 supporting up to 8K resolution across four screens. Connectivity is handled by Wi-Fi 6E, Bluetooth 5.2, and dual 2.5 GbE LAN ports.

Front I/O includes two USB 3.2 Gen 2 ports, a 3.5 mm combo jack, USB4, and an OCulink connector. The rear panel adds two USB 2.0 ports, HDMI, DP, USB4, dual LAN, and a security lock slot. The system ships with Windows 11 Pro pre-installed and also supports Linux distributions such as Ubuntu. The GMKtec NucBox M7 Ultra is available globally starting at $309.99 for the base model (no DDR, SSD or OS), with configurations up to 32 GB RAM and 1 TB SSD priced at $429.99. Early buyers before November 9 will receive a free USB Hub expansion dock.

TrendForce's latest investigations show that server DRAM contract prices are strengthening in 4Q25, driven by ongoing data center expansion among global CSPs. This momentum is lifting overall DRAM pricing. Although final contract pricing for the quarter is still being negotiated, suppliers are showing a greater willingness to raise quotes as CSPs increase order volumes.

TrendForce has accordingly revised its 4Q25 outlook for conventional DRAM pricing upward, from an earlier forecast of 8-13% growth to 18-23%, with a strong likelihood of further upward revision.

Thermaltake accidentally confirms Intel LGA 1954 support for its newest CPU cooler On the web page for its new MINECUBE 360 Ultra CPU cooler, which we saw at Computex 2025, Thermaltake accidentally unveiled that its new cooler supports Intel’s upcoming Nova Lake CPUs. Specifically, their cooler was listed as supporting Intel’s next-generation LGA-1954 socket, which […]

The post Thermaltake confirms Intel Nova Lake/LGA 1954 support with new MINECUBE CPU cooler appeared first on OC3D.

Celebrate the reveal of Halo: Campaign Evolved with our ultimate quiz covering Halo 1, 2, and 3. From iconic missions to obscure lore, see if you’ve got what it takes to finish the fight.

The critically acclaimed, multi-million copy-selling remake of Silent Hill 2 by Bloober Team is off by 55% over at Loaded for its limited-time Halloween sale.

Apparently, only one leader is fully committed to giving power to the players.

We have found a special 79% Halloween discount for Carrion, a classic indie horror game that flips the script by having you play as the monster instead of being hunted by one.

In the last two weeks of October 2025, Microsoft released several Insider builds for Windows 11 versions 25H2 and 24H2 with various visual changes and improvements, including a redesigned About page, the Copy & Search feature, proactive memory diagnostics, and more dark mode improvements for File Explorer.

The Dell portfolio of laptops and other devices can be overwhelming, even after a simplifying rebrand. Here's what you need to know to pick the best Dell for you.

Battlefield 6 and its new battle royale mode, RedSec, both support cross-play across Xbox, PlayStation, and PC — but not everyone wants mixed-platform lobbies.

Nintendo’s patent fight over Palworld takes a blow as Japan’s patent office rejects one of its “monster capture” patents, citing older games like ARK and Pokémon GO, and potentially bolstering Pocketpair’s defense.

Venture funding has most definitely rebounded since the 2022 correction, but there’s a sharp divide between who’s getting funding and who’s not.

That was the overarching theme from our third-quarter market reports, which showed that global startup funding in Q3 totaled $97 billion, marking only the fourth quarter above $90 billion since Q3 2022.

Still, there are stark differences between the 2021 market peak and now, as contributing reporter Joanna Glasner noted in a couple of recent columns, here and here. Just as we saw four years ago, funding is frothy and often seems to be driven by investor FOMO. Some companies are even raising follow-on rounds at head-spinning speeds.

But the funding surge this time is also much, much more concentrated — namely in outsized rounds for AI companies.

With that, let’s take a look at the charts that illustrate the major private-market and startup funding themes as we head into the final quarter of 2025.

AI funding continues to drive venture growth

Nearly half — 46% — of startup funding globally in Q3 went to AI companies, Crunchbase data shows. Almost a third went to a single company: Anthropic, which raised $13 billion last quarter.

Even with an astonishing $45 billion going to artificial intelligence startups in Q3, it was only the third-highest quarter on record for AI funding, with Q4 2024 and Q1 2025 each clocking in higher.

Megarounds gobble up lion’s share

It shouldn’t come as too much of a surprise that AI has also skewed investment heavily toward megarounds, which we define as funding deals of $100 million or more.

The percentage of overall funding going into such deals hit a record high this year, with an astonishing 60% of global and 70% of U.S. venture capital going to $100 million-plus rounds, per Crunchbase data.

Even with several months left in the year, it also seems plausible that the total dollars going into such deals will match or top what we saw in 2021, which marked a peak for startup funding not scaled before or since.

The difference? Back then, startup dollars were widely distributed, going to a whole host of sectors — from food tech to health tech to robotics — and to early-stage, late-stage and in-between companies alike.

Contrast that with recent quarters, when the LLM giants and other large, established, AI-centric companies are getting the largest slice of venture dollars.

Seed deals slide further

As megarounds have increased, seed deals have declined.

The number of seed deals has shown a steady downward trend in recent quarters, Crunchbase data shows, even as total dollars invested at the stage has stayed relatively steady. That indicates that while seed deals are growing larger, they’re also harder to come by.

Early-stage funding has essentially flatlined, despite larger rounds to companies working on robotics, biotech, AI and other technologies.

The AI haves and have-nots

AI has enthralled investors for the past three years.

What are they less interested in? Old standbys like cybersecurity and biotech. Biotech investment as a share of overall funding recently hit a 20-year low. Crunchbase data shows that cybersecurity investment, while still relatively steady, also retreated somewhat in Q3 2025. That’s notable given that many cybersecurity companies are integrating AI into their offerings.

Still, other sectors that benefit heavily from AI-driven automation are seeing a surge in investment. Perhaps most notable is legal tech, which hit an all-time high last month on the back of large rounds for companies promising to automate much of the drudgery of the profession.

Among the other sectors buoyed by AI is human resources software (including AI-powered recruitment and hiring offerings).

Other data points of note

Other interesting points that emerged from our Q3 reports and recent coverage include:

• Unicorn creation in the third quarter swelled as 26 new billion-dollar startups trotted onto our Unicorn Board in September alone.
• Q3’s most-active investors were the serial backers in those AI megarounds. They included well-known names such as Insight Partners, General Catalyst, Accel, Andreessen Horowitz and Iconiq Capital.
• Funding for Asia-based startups rose sequentially in the third quarter, boosted by a handful of hardtech-focused megarounds.
• European startups raised $13.1 billion across more than 1,000 deals in Q3, flat quarter over quarter but up 22% year over year. Early-stage investment was particularly strong, and the region benefited from the excitement around Klarna’s long-awaited IPO.
• In Latin America, Brazil pushed Mexico out of the top spot to retake its crown as the region’s largest funding recipient.

Looking ahead

The increasing concentration of capital into a small cadre of large AI companies — not to mention the interconnectedness of those deals — begs some obvious questions. Are we in a bubble? And given that nearly half of venture capital in recent years has been tied up in AI, what happens to the startup ecosystem if or when it pops?

Related reading:

• Q3 Venture Funding Jumps 38% As More Massive Rounds Go To AI Giants And Exits Gain Steam
• North American Startup Funding Held High In Q3
• Europe’s Venture Scene Held Steady In Q3, Buoyed By Early-Stage Funding And Klarna IPO
• Asia Startup Investment Up In Q3
• Active Investors Kept Busy In An AI-Centric Quarter
• Brazil Back On Top In Q3 When It Comes To Venture Funding In Latin America

Illustration: Dom Guzman

In a rare deep-dive on Microsoft's gaming strategy with TPBN, CEO Satya Nadella outlined the company's evolving vision for Xbox and Windows gaming. His comments follow Microsoft's 2023 acquisition of Activision Blizzard, as the company reconsiders the future of Xbox – not just as a console, but as a broader platform.

Read Entire Article

In recent months, leadership at Electronic Arts, the publisher behind popular game franchises such as The Sims and Madden NFL, has intensified calls for its nearly 15,000 employees to incorporate AI into daily operations.

Read Entire Article

Robotics company 1X Technologies (1XT), which secured $100 million in funding from OpenAI, EQT Ventures, and others in 2024, unveiled a bipedal humanoid robot, Neo Beta, last year.

Read Entire Article

CRN senior editor Dylan Martin highlighted the fake GTC keynote on X yesterday. While YouTube says it does all it can to fight scams, the fake stream, hosted on a channel called NVIDIA Live, was the top result when typing "Nvidia gtc dc" into the platform's search bar.

Read Entire Article

The new arrangement solidifies OpenAI's position as a profit-seeking enterprise, ending nearly a year of restructuring discussions that had drawn scrutiny from regulators and investors. Under the terms, Microsoft's access to OpenAI's technology, including any systems capable of achieving artificial general intelligence (AGI), extends through 2032.

Read Entire Article

With the flak that Apple is reportedly receiving for its ultra-slim iPhone Air, there was an evolving consensus for a time that Samsung would simply trash its upcoming Galaxy S26 Edge. However, new reporting from the Netherlands now indicates that the project is not only alive but thriving, and is all set to debut next year, albeit with a hefty delay. Samsung Galaxy S26 Edge is likely to debut months after the rest of the S26 variants To wit, the Netherlands' GalaxyClub is now reporting that Samsung continues to develop a smartphone under the relatively silly alias of "More Slim." […]

Read full article at https://wccftech.com/no-wait-a-more-slim-galaxy-s26-edge-is-still-in-the-cards/

Hideo Kojima was as surprised as anyone learning this week that he was offered to develop a video game project based on the Matrix franchise, as no one ever told him such a conversation took place between the Wachowskis and Konami. Earlier today, the legendary game designer and creator of the Metal Gear and Death Stranding series clarified the statements made a few days ago by Christopher Bergstresser, the Senior Vice President of Strategic Planning and business Development at Konami between 1996 and 2002. In his post on X, Kojima-san confirmed that, back in 1999, he and the Wachowskis exchanged […]

Read full article at https://wccftech.com/hideo-kojima-wasnt-aware-he-was-offered-a-matrix-video-game-project/

A new Geekbench 6 test is doing the rounds today on social media, showing the raw performance of what appears to be the final configuration of Samsung's upcoming flagship Exynos 2600 chip. Unfortunately, the chip fails to exceed the proven benchmark scores of Qualcomm's Snapdragon 8 Elite Gen 5, but manages to significantly narrow down the erstwhile performance gap between the two SoCs. Samsung Exynos 2600 falls behind Qualcomm's Snapdragon 8 Elite Gen 5 in both single-core and multi-core GeekBench 6 tests First, let's go over the Exynos 2600 setup used in the test: For comparison, here is the CPU […]

Read full article at https://wccftech.com/a-new-geekbench-6-test-shows-samsung-exynos-2600-holding-up-surprisingly-well-against-qualcomms-snapdragon-8-elite-gen-5/

This is another cheap OLED 1440p gaming monitor, which boasts up to 280Hz refresh rate and uses LG's latest LG Tandem OLED panel. KTC Unveils G27P6M OLED 280Hz@1440p Gaming Monitor With LG Primary RGB Tandem OLED; Official Price $422, but Limited Time Deal Price is $281 Looks like Lenovo isn't the only one that is willing to offer its OLED gaming monitors at the cheapest price. KTC, a popular Chinese monitor maker, joined the race and has unveiled its latest offering called G27P6M, which is based on LG's Primary RGB Tandem OLED panel. The monitor launched recently, offering competitive specifications, […]

Read full article at https://wccftech.com/ktc-launches-g27p6m-2k-oled-monitor-featuring-4th-gen-lg-oled/

The PlayStation Portal handheld system may be about to get a highly-requested feature soon, judging from a PS App leak. Over on the PlayStation Portal official subreddit, user GetTheWetsOn shared a screenshot captured from the PS App of the Deliver at All Costs store page, which seemingly confirms that cloud streaming of purchased digital games could be coming to the PlayStation Portal for PS Plus Premium subscribers. The user also stated that the same confirmation, which has since been removed, could be found on the Dead Space remake store page. While there's no indication this new PlayStation Portal streaming option […]

Read full article at https://wccftech.com/playstation-portal-leak-suggests-highly-requested-feature-is-coming-soon/

It seems like the US chip industry is going through a 'massive' revolution, and with that, a startup, Substrate, has decided to tap into the lithography segment, which ASML heavily dominates. Substrate Intends to Utilize X-Rays Over EUV For Lithography, Claiming It To Be a Cheaper Alternative When it comes to chip lithography equipment, the US is entirely dependent on companies like ASML, as the nation lacks a domestic technology that can rival the Dutch chipmaker. However, in a new report by Bloomberg, it seems like there's a startup by the name of 'Substrate' out there that plans to end […]

Read full article at https://wccftech.com/american-chip-startup-substrate-vows-to-end-dependency-on-asml/

SK Hynix has reportedly prepared several new DDR5 chips that can achieve a native speed of 7200 MT/s, as spotted on a popular e-commerce platform. Three More SK Hynix Memory Chips Get Prepared, Offering 7200 MT/s Speed; SK Hynix Reportedly Prepares 2 Gb and 4 Gb Chips as Well Last week, we reported that SK Hynix is preparing 3 Gb (Gigabit) DDR5 A-Die chips, which are rated for 7200 MT/s, above the current JEDEC standard of 6400 MT/s. It appears that SK Hynix isn't just limiting itself to A-Die 3 Gb chips, but has also silently prepared three more dies, […]

Read full article at https://wccftech.com/new-sk-hynix-7200-mt-s-ddr5-memory-spotted-2-gb-b-die-and-4-gb-m-die-prepared/

It was practically a given following the announcement of the mass layoffs at Amazon Game Studios, but we now have confirmation that New World, the MMORPG developed by the team based in Irvine, California, is effectively on life support. Amazon has clarified that Season 10: Nighthaven, which launched earlier this month, will be the last content update because further development 'is not sustainable'. The servers are, however, guaranteed to remain up throughout December 31, 2026 at the least. Complete closure of the servers will likely follow shortly after that date, although Amazon said it will notify players at least six […]

Read full article at https://wccftech.com/new-world-gets-killed-off-amazon-just-as-people-were-returning/

This morning, Remedy Entertainment published its Q3 2025 financial report. As expected, it's not great news for the Finnish studio. Tero Virtala, the company's CEO for nine years, stepped down earlier this month after Remedy told investors that the game FBC: Firebreak had effectively failed. This past quarter, just as the studio celebrated the thirty years since its foundation, the financial highlights were indeed very negative: However, interim CEO Markus Mäki reassured investors (and fans) that the remaining projects are on track, and the studio is still confident in its ability to achieve 'significant commercial success' by 2030. Despite challenges […]

Read full article at https://wccftech.com/remedy-isnt-satisfied-recent-sales-remains-confident-significant-commercial-success/

The moment Mack McConnell knew everything about search had changed came last summer at the Paris Olympics. His parents, independently and without prompting, had both turned to ChatGPT to plan their day's activities in the French capital. The AI recommended specific tour companies, restaurants, and attractions — businesses that had won a new kind of visibility lottery.

"It was almost like this intuitive interface that older people were as comfortable with using as younger people," McConnell recalled in an exclusive interview with VentureBeat. "I could just see the businesses were now being recommended."

That observation has now become the foundation of Geostar, a Pear VC-backed startup that's racing to help businesses navigate what may be the most significant shift in online discovery since Google's founding. 

The company, which recently emerged from stealth with impressive early customer traction, is betting that the rise of AI-powered search represents a significant opportunity to reinvent how companies get found online. The global AI search engine market alone is projected to grow from $43.63 billion in 2025 to $108.88 billion by 2032.

Already the fastest-growing company in PearX's latest cohort, Geostar is fast approaching $1 million in annual recurring revenue in just four months — with only two founders and no employees.

Why Gartner predicts traditional search volume will decline 25% by 2026

The numbers tell a stark story of disruption. Gartner predicts that traditional search engine volume will decline by 25% by 2026, largely due to the rise of AI chatbots. Google's AI Overviews now appear on billions of searches monthly. Princeton University researchers have found that optimizing for these new AI systems can increase visibility by up to 40%.

"Search used to mean that you had to make Google happy," McConnell explained. "But now you have to optimize for four different Google interfaces — traditional search, AI Mode, Gemini, and AI Overviews — each with different criteria. And then ChatGPT, Claude, and Perplexity each work differently on top of that."

This fragmentation is creating chaos for businesses that have spent decades perfecting their Google search strategies. A recent Forrester study found that 95% of B2B buyers plan to use generative AI in future purchase decisions. Yet most companies remain woefully unprepared for this shift.

"Anybody who's not on this right now is losing out," said Cihan Tas, Geostar's co-founder and chief technology officer. "We see lawyers getting 50% of their clients through ChatGPT now. It's just such a massive shift."

How language models read the web differently than search engines ever did

What Geostar and a growing cohort of competitors call Generative Engine Optimization or GEO represents a fundamental departure from traditional search engine optimization. Where SEO focused primarily on keywords and backlinks, GEO requires understanding how large language models parse, understand, and synthesize information across the entire web.

The technical challenges are formidable. Every website must now function as what Tas calls "its own little database" capable of being understood by dozens of different AI crawlers, each with unique requirements and preferences. Google's systems pull from their existing search index. ChatGPT relies heavily on structured data and specific content formats. Perplexity shows a marked preference for Wikipedia and authoritative sources.

"Now the strategy is actually being concise, clear, and answering the question, because that's directly what the AI is looking for," Tas explained. "You're actually tuning for somewhat of an intelligent model that makes decisions similarly to how we make decisions."

Consider schema markup, the structured data that helps machines understand web content. While only 30% of websites currently implement comprehensive schema, research shows that pages with proper markup are 36% more likely to appear in AI-generated summaries. Yet most businesses don't even know what schema markup is, let alone how to implement it effectively.

Inside Geostar's AI agents that optimize websites continuously without human intervention

Geostar's solution embodies a broader trend in enterprise software: the rise of autonomous AI agents that can take action on behalf of businesses. The company embeds what it calls "ambient agents" directly into client websites, continuously optimizing content, technical configurations, and even creating new pages based on patterns learned across its entire customer base.

"Once we learn something about the way content performs, or the way a technical optimization performs, we can then syndicate that same change across the remaining users so everyone in the network benefits," McConnell said.

For RedSift, a cybersecurity company, this approach yielded a 27% increase in AI mentions within three months. In one case, Geostar identified an opportunity to rank for "best DMARC vendors," a high-value search term in the email security space. The company's agents created and optimized content that achieved first-page rankings on both Google and ChatGPT within four days.

"We're doing the work of an agency that charges $10,000 a month," McConnell said, noting that Geostar's pricing ranges from $1,000 to $3,000 monthly. "AI creates a situation where, for the first time ever, you can take action like an agency, but you can scale like software."

Why brand mentions without links now matter more than ever in the AI era

The implications of this shift extend far beyond technical optimizations. In the SEO era, a mention without a link was essentially worthless. In the age of AI, that calculus has reversed. AI systems can analyze vast amounts of text to understand sentiment and context, meaning that brand mentions on Reddit, in news articles, or across social media now directly influence how AI systems describe and recommend companies.

"If the New York Times mentions a company without linking to it, that company would actually benefit from that in an AI system," McConnell explained. "AI has the ability to do mass analysis of huge amounts of text, and it will understand the sentiment around that mention."

This has created new vulnerabilities. Research from the Indian Institute of Technology and Princeton found that AI systems show systematic bias toward third-party sources over brand-owned content. A company's own website might be less influential in shaping AI perceptions than what others say about it online.

The shifting landscape has also disrupted traditional metrics of success. Where SEO focused on rankings and click-through rates, GEO must account for what researchers call impression metrics — how prominently and positively a brand appears within AI-generated responses, even when users never click through to the source.

A growing market as SEO veterans and new players rush to dominate AI optimization

Geostar is hardly alone in recognizing this opportunity. Companies like Brandlight, Profound, and Goodie are all racing to help businesses navigate the new landscape. The SEO industry, worth approximately $80 billion globally, is scrambling to adapt, with established players like Semrush and Ahrefs rushing to add AI visibility tracking features.

But the company's founders, who previously built and sold a Y-Combinator-backed e-commerce optimization startup called Monto, believe their technical approach gives them an edge. Unlike competitors who largely provide dashboards and recommendations, Geostar's agents actively implement changes.

"Everyone is taking the same solutions that worked in the last era and just saying, 'We'll do this for AI instead,'" McConnell argued. "But when you think about what AI is truly capable of, it can actually do the work for you."

The stakes are particularly high for small and medium-sized businesses. While large corporations can afford to hire specialized consultants or build internal expertise, smaller companies risk becoming invisible in AI-mediated search. Geostar sees this as its primary market opportunity: nearly half of the 33.2 million small businesses in America invest in SEO. Among the roughly 418,000 law firms in the U.S., many spend between $2,500 and $5,000 monthly on search optimization to stay competitive in local markets.

From Kurdish village to PearX: The unlikely partnership building the future of search

For Tas, whose journey to Silicon Valley began in a tiny Kurdish village in Turkey with just 50 residents, the current moment represents both opportunity and responsibility. His mother's battle with cancer prevented him from finishing college, leading him to teach himself programming and eventually partner with McConnell — whom he worked with for an entire year before they ever met in person.

"We're not just copy and pasting a solution that was existing before," Tas emphasized. "This is something that's different and was uniquely possible today."

Looking forward, the transformation of search appears to be accelerating rather than stabilizing. Industry observers predict that search functionality will soon be embedded in productivity tools, wearables, and even augmented reality interfaces. Each new surface will likely have its own optimization requirements, further complicating the landscape.

"Soon, search will be in our eyes, in our ears," McConnell predicted. "When Siri breaks out of her prison, whatever that Jony Ive and OpenAI are building together will be like a multimodal search interface."

The technical challenges are matched by ethical ones. As businesses scramble to influence AI recommendations, questions arise about manipulation, fairness, and transparency. There's currently no oversight body or established best practices for GEO, creating what some critics describe as a Wild West environment.

As businesses grapple with these changes, one thing seems certain: the era of simply optimizing for Google is over. In its place is emerging a far more complex ecosystem where success requires understanding not just how machines index information, but how they think about it, synthesize it, and ultimately decide what to recommend to humans seeking answers.

For the millions of businesses whose survival depends on being discovered online, mastering this new paradigm isn't just an opportunity — it's an existential imperative. The question is no longer whether to optimize for AI search, but whether companies can adapt quickly enough to remain visible as the pace of change accelerates.

McConnell's parents at the Olympics were a preview of what's already becoming the norm. They didn't search for tour companies in Paris. They didn't scroll through results or click on links. They simply asked ChatGPT what to do — and the AI decided which businesses deserved their attention.

In the new economy of discovery, the businesses that win won't be the ones that rank highest. They'll be the ones AI chooses to recommend.

Last week, we launched the Internet2 NET+ Google AI Education Leadership Program (ELP).

Comscore's streaming TV report showcases how ad-supported and FAST platforms are thriving.

Most people associate U.S. automotive innovation and industry with Michigan—Detroit, or perhaps Dearborn, where Henry Ford built his colossal River Rouge Plant. They probably wouldn’t think of Englewood Cliffs, New […]

London-based social agency Coolr has appointed Mark Lainas as its first U.S. president, marking a major push into the American market. Coolr, founded in London in 2017, blends influencer marketing, […]

The winning entry for NBC's On Brand billboard design contest will be displayed in Los Angeles.

Sports linear channel NBCSN is set to launch this fall.

Most marketing teams still treat SEO and PPC as budget rivals, not as complementary systems facing the same performance challenges.

In practice, these relationships fall into three types:

• Parasitism: One benefits at the other’s expense.
• Commensalism: One benefits while the other remains unaffected.
• Mutualism: Both thrive through shared optimization and accountability.

Only mutualism creates sustainable performance gains – and it’s the shift marketing teams need to make next.

Mutualism: Solving joint problems

One glaring problem unites online marketers: we’re getting less traffic for the same budget.

Navigating the coming years requires more than the coexistence many teams mistake for collaboration. 

We need mutualism – shared technical standards that optimize for both organic visibility and paid performance. 

Shared accountability drives lower acquisition costs, faster market response, and sustainable gains that neither channel can achieve alone.

Here’s what it looks like in practice: 

• Fostering a culture of experimentation and learning.
• PPC tests messaging while SEO builds long-term content assets​.
• SEO uncovers search intent that PPC capitalizes on immediately​.
• Both channels learn from shared incrementality testing (guerrilla testing)​.
• Cross-pollination of keyword intelligence and conversion data​.
• Combined technical standards (modified Core Web Vitals weights) align engineering with marketing goals​.
• Feedback loops accelerate market insights and reduce wasted spend​.

Stabilizing performance during SEO volatility

During SEO penalties and core updates, PPC can maintain traffic until recovery. 

Core updates cause fluctuations in organic rankings and user behavior, which, in turn, can affect ad relevance and placements.

Do you involve SEO during a CPC price surge?

PPC-only landing pages affect the Core Web Vitals of entire sites, influencing Google’s default assumptions for URLs without enough traffic to calculate individual scores. 

Paid pages are penalized for slow loading just as much as organic ones, impacting Quality Score and, ultimately, bids.

New-market launch considerations

PPC should answer a simple question: Are we getting the types of results we expect and want? 

Setting clear PPC baselines by market and country provides valuable, real-time keyword and conversion data that SEO teams can use to strengthen organic strategies. 

By analyzing which PPC clicks drive signups or demo requests, SEO teams can prioritize content and keyword targets with proven high intent. 

Sharing PPC insights enables organic search teams to make smarter decisions, improve rankings, and drive better-qualified traffic.

Dig deeper: The end of SEO-PPC silos: Building a unified search strategy for the AI era

---

Building unified performance measurement

One key question to ask is: how do we measure incrementality? 

We need to quantify the true, additional contribution PPC and SEO drive above the baseline. 

Guerrilla testing offers a lo-fi way to do this – turning campaigns on or off in specific markets to see whether organic conversions are affected. 

A more targeted test involves turning off branded campaigns. 

PPC ads on branded terms can capture conversions that would have occurred organically, making paid results appear stronger and SEO weaker. 

That’s exactly what Arturs Cavniss’ company did – and here are the results.

For teams ready to operate in a more sophisticated way, several options are available. 

One worth exploring is Robyn, an open-source, AI/ML-powered marketing mix modeling (MMM) package.

Core Web Vitals

Core Web Vitals measures layout stability, rendering efficiency, and server response times – key factors influencing search visibility and overall performance. 

These metrics are weighted by Google in evaluating page experience.

Core Web Vitals Metric	Google’s Weight
First Contentful Paint	10%
Speed Index	10%
Largest Contentful Paint	25%
Total Blocking Time	30%
Cumulative Layout Shift	25%

Core Web Vitals:

• Affect PPC performance through CLS metrics.
• Influence SEO rankings through search vitals.
• Give engineering teams clear benchmarks that align development efforts with marketing goals.

You can create a modified weighted system to reflect a combined SEO and PPC baseline. (Here’s a quick MVP spreadsheet to get started.)

However, SEO-focused weightings don’t capture PPC’s Quality Score requirements or conversion optimization needs. 

Clicking an ad link can be slower than an organic one because Google’s ad network introduces extra processes – additional data handling and script execution – before the page loads. 

The hypothesis is that ad clicks may consistently load slower than organic ones due to these extra steps in the ad-serving process. 

This suggests that performance standards designed for organic results may not fully represent the experience of paid users.

Microsoft Ads Liaison Navah Hopkins notes that paid pages are penalized for slow loading just as severely as organic ones – a factor that directly affects Quality Score and bids.

SEOs also take responsibility for improving PPC-only landing pages, even without being asked. As Jono Alderson explains:

• “All of your PPC-only landing pages are affecting the CWV of your whole site (and, Google’s default assumptions for all of your URLs that don’t have enough traffic to calculate), and thus f*ck with your SEO.”

PPC-only landing pages influence the Core Web Vitals of entire sites, shaping Google’s assumptions for low-traffic URLs.

INP gains importance with agentic AI

Agentic AI’s sensitivity to interaction delays has made Interaction to Next Paint (INP) a critical performance metric.

INP measures how quickly a website responds when a human or AI agent interacts with a page – clicking, scrolling, or filling out forms while completing tasks. 

When response times lag, agents fail tasks, abandon the site, and may turn to competitors. 

INP doesn’t appear in Chrome Lighthouse or PageSpeed Insights because those are synthetic testing tools that don’t simulate real interactions. 

Real user monitoring helps reveal what’s happening in practice, but it still can’t capture the full picture for AI-driven interactions.

Bringing quality scoring to SEO

PPC practitioners have long relied on Quality Score – a 1-10 scale measuring expected CTR and user intent fit – to optimize landing pages and reduce costs. 

SEO lacks an equivalent unified metric, leaving teams to juggle separate signals like Core Web Vitals, keyword relevance, and user engagement without a clear prioritization framework.

You can create a company-wide quality score for pages to incentivize optimization and align teams while maintaining channel-specific goals. 

This score can account for page type, with sub-scores for trial, demo, or usage pages – adaptable to the content that drives the most business value.

The system should account for overlapping metrics across subscores yet remain simple enough for all teams – SEO, PPC, engineering, and product – to understand and act on. 

A unified scoring model gives everyone a common language and turns distributed accountability into daily practice. 

When both channels share quality standards, teams can prioritize fixes that strengthen organic rankings and paid performance simultaneously.

Give a comprehensive view across channels

Display advertising and SEO rarely share performance metrics, yet both pursue the same goal – converting impressions into engaged users. 

Click-per-thousand impressions (CPTI) measures the number of clicks generated per 1,000 impressions, creating a shared language for evaluating content effectiveness across paid display and organic search.

For display teams, CPTI reveals which creative and targeting combinations drive engagement beyond vanity metrics like reach. 

For SEO teams, applying CPTI to search impressions (via Google Search Console) shows which pages and queries convert visibility into traffic – exposing content that ranks well but fails to earn clicks.

This shared metric allows teams to compare efficiency directly: if a blog post drives 50 clicks per 1,000 organic impressions while a display campaign with similar visibility generates only 15 clicks, the performance gap warrants investigation.

Reverse CPM offers another useful lens. It measures how long content takes to “pay for itself” – the point where it reaches ROI. 

For example, if an article earns 1 million impressions in a month, it should deliver roughly 1,000 clicks. 

As generative AI continues to reshape traffic patterns, this metric will need refinement.

Feedback loops

The most valuable insights emerge when SEO and PPC teams share operational intelligence rather than compete for credit. 

PPC provides quick keyword performance data to respond to market trends faster, while SEO uncovers emerging search intent that PPC can immediately act on. 

Together, these feedback loops create compound advantages.

SEO signals PPC should act on:

• Google is testing a feature that impacts SEO rankings and traffic – PPC can maintain visibility during organic volatility.
• SEO keyword research uncovers search intent, emerging keywords, seasonal patterns, and regional differences in query popularity.
• Long-tail insights reveal shifting search intents after core updates, signaling format and content opportunities.

PPC signals SEO should act on: 

• Some PPC keywords are effectively “dead.” They’ll never convert and are better handled by SEO.
• PPC competitors bidding on brand keywords expose gaps in brand protection strategy.
• PPC data highlights which product messaging, features, or offers resonate most with users, informing content priorities.

When both channels share intelligence, insights extend beyond marketing performance into product and business strategy.

• Product managers exploring new features benefit from unified search data across both channels.
• Joining Merchant Center and Google Search Console data in BigQuery provides a strong foundation for ecommerce attribution.

These feedback loops don’t require expensive tools – only an organizational commitment to regular cross-channel reviews in which teams share what’s working, what’s failing, and what deserves coordinated testing.

Optimizing the system, not the channel

Treat technical performance as shared infrastructure, not channel-specific optimization. 

Teams that implement unified Core Web Vitals standards, cross-channel attribution models, and distributed accountability systems will capture opportunities that siloed operations miss. 

As agentic AI adoption accelerates and digital marketing grows more complex, symbiotic SEO-PPC operations become a competitive advantage rather than a luxury.

Something’s shifting in how SEO services are being marketed, and if you’ve been shopping for help with search lately, you’ve probably noticed it.

AI search demand is real – but so is the spin

Over the past few months, “AI SEO” has emerged as a distinct service offering. 

Browse service provider websites, scroll through Fiverr, or sit through sales presentations, and you’ll see it positioned as something fundamentally new and separate from traditional SEO. 

Some are packaging it as “GEO” (generative engine optimization) or “AEO” (answer engine optimization), with separate pricing, distinct deliverables, and the implication that you need both this and traditional SEO to compete.

The pitch goes like this: 

• “Traditional SEO handles Google and Bing. But now you need AI SEO for ChatGPT, Perplexity, Claude, and other AI search platforms. They work completely differently and require specialized optimization.”

The data helps explain why the industry is moving so quickly.

AI-sourced traffic jumped 527% year-over-year from early 2024 to early 2025. 

Service providers are responding to genuine market demand for AI search optimization.

But here’s what I’ve observed after evaluating what these AI SEO services actually deliver. 

Many of these so-called new tactics are the same SEO fundamentals – just repackaged under a different name.

As a marketer responsible for budget and results, understanding this distinction matters. 

It affects how you allocate resources, evaluate agency partners, and structure your search strategy. 

Let’s dig into what’s really happening so you can make smarter decisions about where to invest.

The AI SEO pitch: What you’re hearing in sales calls

The typical AI SEO sales deck has become pretty standardized. 

• First comes the narrative about how search is fragmenting across platforms. 
• Then, the impressive dashboard showing AI visibility metrics. 
• Finally, the recommendation to treat AI optimization as a separate workstream, often with separate pricing.

Here are the most common claims I’m hearing.

‘AI search is fundamentally different and requires specialized optimization’ 

They’ll show you how ChatGPT, Perplexity, and Claude are changing search behavior, and they’re not wrong about that. 

Research shows that 82% of consumers agree that “AI-powered search is more helpful than traditional search engines,” signaling how search behavior is evolving.

‘You need to optimize for how AI platforms chunk and retrieve content’ 

The pitch emphasizes passage-level optimization, structured data, and Q&A formatting specifically for AI retrieval. 

They’ll discuss how AI values mentions and citations differently than backlinks and how entity recognition matters more than keywords.

‘Only 22% of marketers are monitoring AI visibility; you need to act now’

This creates urgency around a supposedly new practice that requires immediate investment.

The urgency is real. Only 22% of marketers have set up LLM brand visibility monitoring, but the question is whether this requires a separate “AI SEO” service or an expansion of your existing search strategy.

Understanding the rebranding trend

To be clear, the AI capabilities are real. What’s new is the positioning – familiar SEO practices rebranded to sound more revolutionary than they are.

When you examine what’s actually being recommended (passage-level content structure, semantic clarity, Q&A formatting, earning citations and mentions), you will find that these practices have been core to SEO for years. 

Google introduced passage ranking in 2020 and featured snippets back in 2014.

Research from Fractl, Search Engine Land, and MFour found that generative engine optimization “is based on similar value systems that advanced SEOs, content marketers, and digital PR teams are already experts in.”

Let me show you what I mean.

What you’re hearing: “AI-powered semantic analysis and predictive keyword intelligence.”

• What’s actually happening: Keyword research using advanced tools to analyze search volume, competition, user intent, and content opportunities. The strategic fundamentals (understanding what your audience is searching for and why) haven’t changed.

What you’re hearing: “Machine learning content optimization that aligns with AI algorithms.”

• What’s actually happening: Analyzing top-ranking content, understanding user intent, identifying content gaps, and creating comprehensive content. AI tools can accelerate analysis, which is valuable. But the strategic work (determining what topics matter for your business, how to position your expertise, and what content will drive conversions) still requires human insight.

What you’re hearing: “Entity-based authority building for AI platforms.”

• What’s actually happening: Building quality mentions and citations, earning coverage from reputable sources, and establishing expertise in your industry. Authority building is inherently relationship-driven and time-dependent. No AI tool shortcuts to becoming a recognized expert in your space.

Dig deeper: AI search is booming, but SEO is still not dead

---

Where real differences exist (and why fundamentals still matter)

I want to be fair here. There’s genuine debate in the SEO community about whether optimizing for AI-powered search represents a distinct discipline or an evolution of existing practices.

The differences are real.

• AI search handles queries differently from traditional search. 
• Users write longer, conversational prompts rather than short keywords.
• AI platforms use query fan-out to match multiple sub-queries. 
• Optimization happens at the passage or chunk level rather than the page level. 
• Authority signals shift from links and engagement to mentions and citations.

These differences affect execution, but the strategic foundation remains consistent.

You still need to:

• Understand what users are trying to accomplish.
• Create content demonstrating genuine expertise.
• Build authority and credibility.
• Ensure content is technically accessible.
• Optimize for relevance and user intent.

And here’s something that reinforces the overlap.

SEO professionals recently discovered that ChatGPT’s Atlas browser directly uses Google search results. 

Even AI-powered search platforms are relying on traditional search infrastructure.

So yes, there are platform-specific tactics that matter. 

The question for you as a marketer isn’t whether differences exist (they do). 

The real question is whether those differences justify treating this as an entirely separate service with its own strategy and budget.

Or are they simply tactical adaptations of the same fundamental approach?

Dig deeper: GEO and SEO: How to invest your time and efforts wisely

The risk of chasing platform-specific tactics

The “separate AI SEO service” approach comes with a real risk.

It can shift focus toward short-term, platform-specific tactics at the expense of long-term fundamentals.

I’m seeing recommendations that feel remarkably similar to the blackhat SEO tactics we saw a decade ago: 

• Invisible text that only LLMs can see.
• Content cloaking for AI bots.
• Scaled content targeting every possible prompt variation. 

These tactics might work today, but they’re playing a dangerous game.

Dig deeper: Black hat GEO is real – Here’s why you should pay attention

AI platforms are still in their infancy. Their spam detection systems aren’t yet as mature as Google’s or Bing’s, but that will change, likely faster than many expect.

AI platforms like Perplexity are building their own search indexes (hundreds of billions of documents). 

They’ll need to develop the same core systems traditional search engines have: 

• Site quality scoring.
• Authority evaluation.
• Anti-spam measures. 

They’re supposedly buying link data from third-party providers, recognizing that understanding authority requires signals beyond just content analysis.

The pattern is predictable

We’ve seen this with Google. 

In the early days, keyword stuffing and link schemes worked great.

Then, Google developed Panda and Penguin updates that devastated sites relying on those tactics. 

Overnight, sites lost 50-90% of their traffic.

The same thing will likely happen with AI platforms. 

Sites gaming visibility now with spammy tactics will face serious problems when these platforms implement stronger quality and spam detection. 

As one SEO veteran put it, “It works until it doesn’t.”

This is why fundamentals matter more than ever

Building around platform-specific tactics is like building on sand. 

Focus instead on fundamentals – creating valuable content, earning authority, demonstrating expertise, and optimizing for intent – and you’ll have something sustainable across platforms.

Where AI genuinely helps

I’m not anti-AI. Used well, it meaningfully improves SEO workflows and results.

AI excels at large-scale research and ideation – analyzing competitor content, spotting gaps, and mapping topic clusters in minutes.

For one client, it surfaced 73 subtopics we hadn’t fully considered. 

But human expertise was still essential to align those ideas with business goals and strategic priorities.

AI also transforms data analysis and workflow automation – from reporting and rank tracking to technical monitoring – freeing more time for strategy.

AI clearly helps. The real question is whether these AI offerings bring truly new strategies or familiar ones powered by better tools.

What to watch for when evaluating services

After working with clients to evaluate various service models, I’ve seen consistent patterns in proposals that overpromise and underdeliver.

• They lead with technology, not strategy: If the conversation jumps immediately to tools and dashboards rather than starting with your business goals, that suggests a tools-first rather than strategy-first approach.
• Vague explanations of their approach: Watch for responses about “proprietary algorithms” or “advanced machine learning” without concrete explanations of what specific problems this solves.
• Focus on vanity metrics: “We generated 500 AI citations!” sounds impressive but doesn’t answer: Did qualified traffic increase? Did conversion rates improve? How did search contribute to revenue?
• Case studies that focus on visibility, not business results: They might have increased AI mentions or improved rankings, but did it drive revenue growth? Did it increase qualified leads?

Questions to ask instead

When evaluating any service provider, ask:

• How would you approach our business? Walk me through your strategic process. The best approaches start by understanding your business, not showcasing tools. If they jump immediately to AI tools or technical tactics without understanding your business context, that’s a red flag.
• How do you determine content strategy and prioritization? Look for answers that balance data insights with business context and audience understanding, not just what AI tools suggest would perform well.
• What specific results have you achieved for similar businesses? Push for concrete business metrics (revenue growth, lead generation, conversion improvements), not just traffic or ranking increases.
• How do you integrate optimization across traditional search and AI platforms? This reveals whether they view these as separate disciplines requiring separate work or as interconnected parts of a unified search strategy.

What actually drives long-term success

After working in SEO for 20 years, through multiple algorithm updates and trend cycles, I keep coming back to the same fundamentals:

• Deep audience understanding drives every strategic decision.
• Quality and expertise still win (search algorithms are increasingly sophisticated at evaluating content quality).
• Authority building takes time and authenticity (you can’t automate trust and credibility).
• Business alignment drives meaningful results (rankings and AI citations are means to an end: revenue growth, customer acquisition, or whatever your primary business goals are).

Dig deeper: Thriving in AI search starts with SEO fundamentals

What sustainable SEO looks like in the AI era

AI is genuinely changing how we work in search marketing – and that’s mostly positive. 

The tools make us more efficient and enable analysis that wasn’t previously practical.

But AI only enhances good strategy. It doesn’t replace it. 

Fundamentals still matter – along with audience understanding, quality, and expertise.

Search behavior is fragmenting across Google, ChatGPT, Perplexity, and social platforms, but the principles that drive visibility and trust remain consistent.

Real advantage doesn’t come from the newest tools or the flashiest “GEO” tactics. 

It comes from a clear strategy, deep market understanding, strong execution of fundamentals, and smart use of technology to strengthen human expertise.

Don’t get distracted by hype or dismiss innovation. The balance lies in thoughtful AI integration within a solid strategic framework focused on business goals.

That’s what delivers sustainable results – whether people find you through Google, ChatGPT, or whatever comes next.

Artificial Intelligence (AI) is rapidly transforming Governance, Risk, and Compliance (GRC). It's no longer a future concept—it's here, and it's already reshaping how teams operate. AI's capabilities are profound: it's speeding up audits, flagging critical risks faster, and drastically cutting down on time-consuming manual work. This leads to greater efficiency, higher accuracy, and a more

BeyondTrust’s annual cybersecurity predictions point to a year where old defenses will fail quietly, and new attack vectors will surge. Introduction The next major breach won’t be a phished password. It will be the result of a massive, unmanaged identity debt. This debt takes many forms: it’s the “ghost” identity from a 2015 breach lurking in your IAM, the privilege sprawl from thousands of new

Organizations in Ukraine have been targeted by threat actors of Russian origin with an aim to siphon sensitive data and maintain persistent access to compromised networks. The activity, according to a new report from the Symantec and Carbon Black Threat Hunter Team, targeted a large business services organization for two months and a local government entity in the country for a week. The attacks

Cybersecurity researchers have discovered a set of 10 malicious npm packages that are designed to deliver an information stealer targeting Windows, Linux, and macOS systems. "The malware uses four layers of obfuscation to hide its payload, displays a fake CAPTCHA to appear legitimate, fingerprints victims by IP address, and downloads a 24MB PyInstaller-packaged information stealer that harvests

Don't get trapped in a payment plan – these doorbells will give you advanced features and local storage for your videos with no added fees.

The Google Pixel 10a might look almost exactly the same as the Pixel 9a, and have many of the same specs too.

Apple may upgrade its iPad mini with an OLED display next year, according to a new report.

UK retailers have surprised us all by launching several early Black Friday sales, so I'm looking through all the offers and picking out the best deals worth buying now.

The 1TB WD Black C50 Floral Fusion edition looks nice, but it's even cheaper than the discount for the standard product right now.

Lanterns will examine the "push and pull" between Aaron Pierre and Kyle Chandler's lead characters, its co-creator has teased.

Every new movie has a perfect pairing, so I asked the cast of Prime Video's Hedda for their double bill streaming suggestions.

Gaming on Linux just keeps getting better thanks to wider compatibility, but anti-cheat games remain stubbornly problematic.

The 'first wave' of users are finding they've got Gemini for Home alerts on their mobile devices.

Does Only Murders in the Building season 6 being set in the UK make sense? No, but I want this epic crime crossover anyway.

The Sound Blaster Re:Imagine "redefines what a sound card can be"

The redesigned Start menu is exciting, true, but I'd advise you to hold fire for now.

Don't miss this hefty Nintendo Switch 2 Pro Controller price cut on EE, available for a limited time for UK customers.

There's a world of difference between client and enterprise SSDs

OnePlus teams up with Google to bring Gemini to its Mind Space AI tool, and that might be the biggest reason yet to buy a OnePlus 15.

The Turtle Beach Stealth Ultra has had its price slashed in the UK.

Nokia and Nvidia will work together to boost work towards 6G and AI native mobile networks.

This is a huge price drop for one of our favorite smartwatches, so take this as your sign to upgrade.

Most enterprises are running "flat" networks, putting their entire infrastructure at risk.

TechRadar spoke with Celina Shi, CMO of OnePlus Europe, to discuss the company’s upcoming OnePlus 15 flagship.

A unified approach to protection centered on a strong integrated security platform is vital, researchers say.

Disney has provided an update on whether Doctor Who will stay on Disney+ after it decided not to renew its deal with the BBC.

The Creative Aurvana Ace 3 take the framework of their predecessors and fix a few of the biggest issues, delivering better earbuds.

A Nintendo Switch 2 + Mario Kart World bundle discount before Black Friday? It's more likely than you think.

The Samsung Galaxy Z Flip hasn’t always been a reliable source of Black Friday Deals, but I’m hoping the Galaxy Z Flip 7 FE could change that.

The budget-friendly and highly-rated Huawei Watch Fit 3 is available for £68.99 in a limited-time deal at Amazon.

If you want your coffee as quickly as possible, a Nespresso machine could be exactly what you need, and these are three of the best.

Phia founders Phoebe Gates and Sophia Kianni are building an AI assistant that aims to help you shop and save.

Mogul is a film studio fantasy platform where players draft a team of actors, directors, and movies to earn points.

The move puts Uber in direct competitor with Waymo, which has dominated the metro area.

The advantages of the LOQ 15 can’t outweigh the fact that Lenovo could offer more RAN and storage at this price.

A young student was left traumatized after being ordered to the ground and handcuffed by police because an AI gun detection system erroneously called the cops on his Doritos habit.

First images of Nvidia's Vera Rubin Superchip depicting two Rubin GPUs and a multi-chiplet Vera CPU surrounded by SOCAMM2 memory modules emerge.

Create stunning videos by chatting with the best AI models

Discussion | Link

The post Alphabet’s AI Reckoning: Cloud Momentum vs. Search Durability appeared first on StartupHub.ai.

The burgeoning influence of artificial intelligence presents a dual narrative for tech giants, particularly Alphabet, as it navigates both profound opportunities and existential threats to its established revenue streams. As CNBC’s MacKenzie Sigalos reported on “Worldwide Exchange,” ahead of Alphabet’s recent earnings call, the company finds itself at a critical juncture, balancing the burgeoning momentum […]

The post Alphabet’s AI Reckoning: Cloud Momentum vs. Search Durability appeared first on StartupHub.ai.

The post TestSprite raises $6.7M to fix AI generated code testing appeared first on StartupHub.ai.

TestSprite raised $6.7 million to address the critical bottleneck of AI generated code testing, enabling developers to validate AI-written code at unprecedented speeds.

The post TestSprite raises $6.7M to fix AI generated code testing appeared first on StartupHub.ai.

The post Amazon’s AI Power Play: Inside the $11 Billion Indiana Data Center appeared first on StartupHub.ai.

Amazon’s latest $11 billion AI data center in Indiana signals a profound shift in the foundational infrastructure powering the artificial intelligence revolution, an investment of unprecedented scale that underscores the intense competition for AI compute. In less than a year, Amazon transformed vast Indiana cornfields into its largest AI data center yet, an astonishing feat […]

The post Amazon’s AI Power Play: Inside the $11 Billion Indiana Data Center appeared first on StartupHub.ai.

The post AI Browser Security: The Peril of the Premature Launch appeared first on StartupHub.ai.

“The rush to get these things to market has not allowed them to be secured.” This stark assessment from Dave McGinnis, Global Partner for Cyber Threat Management Offering Group at IBM, encapsulates the central tension explored in a recent episode of IBM’s Security Intelligence podcast. Host Matt Kosinski, alongside McGinnis and fellow panelists Suja Viswesan […]

The post AI Browser Security: The Peril of the Premature Launch appeared first on StartupHub.ai.

Seattle-based Radical says it has put a full-size prototype for a solar-powered drone through its first flight, marking one low-altitude step in the startup’s campaign to send robo-planes into the stratosphere for long-duration military and commercial missions.

“It’s a 120-foot-wingspan aircraft that only weighs 240 pounds,” Radical CEO James Thomas told GeekWire. “We’re talking about something that has a wingspan just a bit bigger than a Boeing 737, but it only weighs a little bit more than a person. So, it’s a pretty extreme piece of engineering, and we’re really proud of what our team has achieved so far.”

Last month’s flight test was conducted at the Tillamook UAS Test Range in Oregon, which is one of the sites designated by the Federal Aviation Administration for testing uncrewed aerial systems. Thomas declined to delve into the details about the flight’s duration or maximum altitude, other than to say that it was a low-altitude flight.

“We take off from the top of a car, and takeoff speeds are very low, so it flies just over 15 miles an hour on the ground or at low altitudes,” he said. (Thomas later added that the car was a Subaru, a choice he called “a Pacific Northwest move, I guess.”)

The prototype ran on battery power alone, but future flights will make use of solar arrays mounted on the plane’s wings to keep it in the air at altitudes as high as 65,000 feet for months at a time. For last month’s test, engineers added ballast to the prototype to match the weight of the solar panels and batteries required for stratospheric flight. Thomas said he expects high-altitude tests to begin next year.

Thomas and his fellow co-founder, chief technology officer Cyriel Notteboom, are veterans of Prime Air, Amazon’s effort to field a fleet of delivery drones. They left Amazon in mid-2022 to launch Radical and have since raised more than $4.5 million in funding. September’s test of a full-size drone follows up on the 24-hour-plus flight of a 13-pound subscale prototype in 2023.

The company’s manufacturing operation is based in Seattle’s Ballard neighborhood. There are currently six people on the team, plus a new hire, Thomas said. “We’re still lean,” he said. “To make this airplane work, it has to be really efficient, right? Really efficient electronics and aerodynamics. And you also need a really efficient team.”

Thomas said Radical has attracted interest from potential customers, but he shied away from discussing details. “We’re working with groups in the government and also commercially,” he said. “Obviously there are applications at the end of this that span everything from imagery through telecommunications and weather forecasting. There are a lot of people really interested in the technology, and the thing that stops us from serving those customers is not having a product up in the sky. So, that’s what we’re working through.”

Radical’s solar-powered airplane, known as Evenstar, is just one example of a class of aircraft known as high-altitude platform stations, or HAPS. Thomas and his teammates use a different term to refer to Evenstar. They call it a StratoSat, because it’s designed to take on many of the tasks typically assigned to satellites — but without the costs and the hassles associated with launching a spacecraft.

Potential applications include doing surveillance from a vantage point that’s difficult to attack, providing telecommunication links in areas where connectivity is constrained, monitoring weather patterns and conducting atmospheric research.

“We have customers who are really excited about the way that this can improve how we understand Earth’s weather systems and climate,” Thomas said. “That’s an application that we’re really excited to get into.”

Evenstar will carry payloads weighing up to about 33 pounds (15 kilograms). “That was based on analysis about major use cases,” Thomas explained. “That payload is enough to carry high-bandwidth, direct-to-device radio communications, or to carry ultra-high-resolution imaging equipment.”

Radical isn’t the only company working on solar-powered aircraft built for long-duration flights in the stratosphere. Other entrants in the market include AeroVironment, SoftBank, BAE Systems, Swift Engineering, Kea Aerospace, Korea Aerospace Industries and NewSpace Research & Technologies. Airbus’ solar-powered Zephyr set the record for long-duration stratospheric flight in 2022 with a 64-day test mission that ended in a crash.

Among those who tried but failed to field stratospheric solar drones are Alphabet, which closed down Titan Aerospace in 2016; and Facebook, which abandoned Project Aquila in 2018.

Thomas said the outlook for high-flying solar planes has brightened in the past decade.

“The key supporting technologies have matured enormously,” he said. “Commercial battery energy density has doubled in that 10-year time period. Solar cells are 10 times cheaper than they were just 10 years ago. And then you have advances in compute and AI, and all of these things feed into the situation we have now, where it’s actually possible to make the models close — whereas when we run the 10-year-old numbers, we can’t close the models.”

The way Thomas sees it, the concept behind Radical isn’t all that radical anymore.

“Not only do our models say this will work, but we have flight data that agrees with our models, and says this is a technology that can serve its purpose and unlock the potential of persistent infrastructure in the sky,” he said. “I can see why other people are pursuing it. It’s not a new idea. It’s one that people have wanted to crack for a long time, and we’re at this critical inflection point where it’s finally possible.”

Socket compatibility is common, but not a given. At least now you don't have to upgrade your cooler, too.

Yesterday the GitHub-hosted AMD XDNA driver code saw a new tagged release as version 202610.2.21.17. That itself wasn't too interesting but while diving into there is new yet-to-be-merged code for a new "NPU3A" revision to their NPU3 IP in Ryzen AI...