Learn more about the new Gemini for Home voice assistant, now rolling out in early access.

Pomelli, our newest experiment from Google Labs and DeepMind, is here to help you create on-brand marketing content.

An overview of the new course from The Google School for Leaders that helps all managers build high-performing teams.

Here’s how students and faculty at higher education institutions are using Gemini.

Users will soon be able to complete purchases within ChatGPT using their PayPal digital wallets.

Chatbots may undercut the value of onsite ads.

E.l.f. Beauty is turning compassion into a movement with its new campaign that dares everyone to show what they really care about. “Give an e.l.f.” launches today with the help of […]

Google has expanded the What’s happening feature within Google Business Profiles to restaurants and bars in the United Kingdom, Canada, Australia, and New Zealand. It is now available for multi-location restaurants, not just single-location restaurants.

The What’s happening feature launched back in May as a way for some businesses to highlight events, deals, and specials prominently at the top of your Google Business Profile. Now, Google is bringing it to more countries.

What Google said. Google’s Lisa Landsman wrote on LinkedIn:

How do you promote your “Taco Tuesday” in Toledo and your “Happy Hour” in Houston… right when locals are searching for a place to go?

I’m excited to share that the Google Business Profile feature highlighting what’s happening at your business, such as timely events, specials and deals, has now rolled out for multi-location restaurants & bars across the US, UK, CA, AU & NZ! (It was previously only available for single-location restaurants)

This is a great option for driving real-time foot traffic. It automatically surfaces the unique specials, live music, or events you’re already promoting at a specific location, catching customers at the exact moment they’re deciding where to eat or grab a cocktail.

What it looks like. Here is a screenshot of this feature:

More details. Google’s Lisa Landsman added, “We’ve already seen excellent results from testing and look forward to hearing how this works for you!”

Availability. This feature is only available for restaurants & bars. Google said it hopes to expand to more categories soon. It is also only available in the United States, United Kingdom, Canada, Australia, and New Zealand.

The initial launch was for single-location Food and Drink businesses in the U.S., UK, Australia, Canada, and New Zealand. It is now available for multi-location restaurants, not just single-location restaurants.

Why we care. If you manage restaurants and/or bars, this may be a new way to get more attention and visitors to your business from Google Search. Now, if you manage multi-location restaurants or bars, you can leverage this feature.

Marketing, technology, and business leaders today are asking an important question: how do you optimize for large language models (LLMs) like ChatGPT, Gemini, and Claude? 

LLM optimization is taking shape as a new discipline focused on how brands surface in AI-generated results and what can be measured today. 

For decision makers, the challenge is separating signal from noise – identifying the technologies worth tracking and the efforts that lead to tangible outcomes.

The discussion comes down to two core areas – and the timeline and work required to act on them:

• Tracking and monitoring your brand’s presence in LLMs.
• Improving visibility and performance within them.

Tracking: The foundation of LLM optimization

Just as SEO evolved through better tracking and measurement, LLM optimization will only mature once visibility becomes measurable. 

We’re still in a pre-Semrush/Moz/Ahrefs era for LLMs. 

Tracking is the foundation of identifying what truly works and building strategies that drive brand growth. 

Without it, everyone is shooting in the dark, hoping great content alone will deliver results.

The core challenges are threefold:

• LLMs don’t publish query frequency or “search volume” equivalents.
• Their responses vary subtly (or not so subtly) even for identical queries, due to probabilistic decoding and prompt context.
• They depend on hidden contextual features (user history, session state, embeddings) that are opaque to external observers.

Why LLM queries are different

Traditional search behavior is repetitive – millions of identical phrases drive stable volume metrics. LLM interactions are conversational and variable. 

People rephrase questions in different ways, often within a single session. That makes pattern recognition harder with small datasets but feasible at scale. 

These structural differences explain why LLM visibility demands a different measurement model.

This variability requires a different tracking approach than traditional SEO or marketing analytics.

The leading method uses a polling-based model inspired by election forecasting.

The polling-based model for measuring visibility

A representative sample of 250–500 high-intent queries is defined for your brand or category, functioning as your population proxy. 

These queries are run daily or weekly to capture repeated samples from the underlying distribution of LLM responses.

Tracking tools record when your brand and competitors appear as citations (linked sources) or mentions (text references), enabling share of voice calculations across all competitors. 

Over time, aggregate sampling produces statistically stable estimates of your brand visibility within LLM-generated content.

Early tools providing this capability include:

• Profound.
• Conductor.
• OpenForge.

Consistent sampling at scale transforms apparent randomness into interpretable signals. 

Over time, aggregate sampling provides a stable estimate of your brand’s visibility in LLM-generated responses – much like how political polls deliver reliable forecasts despite individual variations.

Building a multi-faceted tracking framework

While share of voice paints a picture of your presence in the LLM landscape, it doesn’t tell the complete story. 

Just as keyword rankings show visibility but not clicks, LLM presence doesn’t automatically translate to user engagement. 

Brands need to understand how people interact with their content to build a compelling business case.

Because no single tool captures the entire picture, the best current approach layers multiple tracking signals:

• Share of voice (SOV) tracking: Measure how often your brand appears as mentions and citations across a consistent set of high-value queries. This provides a benchmark to track over time and compare against competitors.
• Referral tracking in GA4: Set up custom dimensions to identify traffic originating from LLMs. While attribution remains limited today, this data helps detect when direct referrals are increasing and signals growing LLM influence.
• Branded homepage traffic in Google Search Console: Many users discover brands through LLM responses, then search directly in Google to validate or learn more. This two-step discovery pattern is critical to monitor. When branded homepage traffic increases alongside rising LLM presence, it signals a strong causal connection between LLM visibility and user behavior. This metric captures the downstream impact of your LLM optimization efforts.

Nobody has complete visibility into LLM impact on their business today, but these methods cover all the bases you can currently measure.

Be wary of any vendor or consultant promising complete visibility. That simply isn’t possible yet.

Understanding these limitations is just as important as implementing the tracking itself.

Because no perfect models exist yet, treat current tracking data as directional – useful for decisions, but not definitive.

Dig deeper: In GEO, brand mentions do what links alone can’t

Estimating LLM ‘search volume’

Measuring LLM impact is one thing. Identifying which queries and topics matter most is another.

Compared to SEO or PPC, marketers have far less visibility. While no direct search volume exists, new tools and methods are beginning to close the gap.

The key shift is moving from tracking individual queries – which vary widely – to analyzing broader themes and topics. 

The real question becomes: which areas is your site missing, and where should your content strategy focus?

To approximate relative volume, consider three approaches:

Correlate with SEO search volume

Start with your top-performing SEO keywords. 

If a keyword drives organic traffic and has commercial intent, similar questions are likely being asked within LLMs. Use this as your baseline.

Layer in industry adoption of AI

Estimate what percentage of your target audience uses LLMs for research or purchasing decisions:

• High AI-adoption industries: Assume 20-25% of users leverage LLMs for decision-making.
• Slower-moving industries: Start with 5-10%.

Apply these percentages to your existing SEO keyword volume. For example, a keyword with 25,000 monthly searches could translate to 1,250-6,250 LLM-based queries in your category.

Using emerging inferential tools

New platforms are beginning to track query data through API-level monitoring and machine learning models. 

Accuracy isn’t perfect yet, but these tools are improving quickly. Expect major advancements in inferential LLM query modeling within the next year or two.

---

Optimizing for LLM visibility

The technologies that help companies identify what to improve are evolving quickly. 

While still imperfect, they’re beginning to form a framework that parallels early SEO development, where better tracking and data gradually turned intuition into science.

Optimization breaks down into two main questions:

• What content should you create or update, and should you focus on quality content, entities, schema, FAQs, or something else?
• How should you align these insights with broader brand and SEO strategies?

Identify what content to create or update

One of the most effective ways to assess your current position is to take a representative sample of high-intent queries that people might ask an LLM and see how your brand shows up relative to competitors. This is where the Share of Voice tracking tools we discussed earlier become invaluable.

These same tools can help answer your optimization questions:

• Track who is being cited or mentioned for each query, revealing competitive positioning.
• Identify which queries your competitors appear for that you don’t, highlighting content gaps.
• Show which of your own queries you appear for and which specific assets are being cited, pinpointing what’s working.

From this data, several key insights emerge:

• Thematic visibility gaps: By analyzing trends across many queries, you can identify where your brand underperforms in LLM responses. This paints a clear picture of areas needing attention. For example, you’re strong in SEO but not in PPC content. 
• Third-party resource mapping: These tools also reveal which external resources LLMs reference most frequently. This helps you build a list of high-value third-party sites that contribute to visibility, guiding outreach or brand mention strategies. 
• Blind spot identification: When cross-referenced with SEO performance, these insights highlight blind spots; topics or sources where your brand’s credibility and representation could improve.

Understand the overlap between SEO and LLM optimization

LLMs may be reshaping discovery, but SEO remains the foundation of digital visibility.

Across five competitive categories, brands ranking on Google’s first page appeared in ChatGPT answers 62% of the time – a clear but incomplete overlap between search and AI results.

That correlation isn’t accidental. 

Many retrieval-augmented generation (RAG) systems pull data from search results and expand it with additional context. 

The more often your content appears in those results, the more likely it is to be cited by LLMs.

Brands with the strongest share of voice in LLM responses are typically those that invested in SEO first. 

Strong technical health, structured data, and authority signals remain the bedrock for AI visibility.

What this means for marketers:

• Don’t over-focus on LLMs at the expense of SEO. AI systems still rely on clean, crawlable content and strong E-E-A-T signals.
• Keep growing organic visibility through high-authority backlinks and consistent, high-quality content.
• Use LLM tracking as a complementary lens to understand new research behaviors, not a replacement for SEO fundamentals.

Redefine on-page and off-page strategies for LLMs

Just as SEO has both on-page and off-page elements, LLM optimization follows the same logic – but with different tactics and priorities.

Off-page: The new link building

Most industries show a consistent pattern in the types of resources LLMs cite:

• Wikipedia is a frequent reference point, making a verified presence there valuable.
• Reddit often appears as a trusted source of user discussion.
• Review websites and “best-of” guides are commonly used to inform LLM outputs.

Citation patterns across ChatGPT, Gemini, Perplexity, and Google’s AI Overviews show consistent trends, though each engine favors different sources.

This means that traditional link acquisition strategies, guest posts, PR placements, or brand mentions in review content will likely evolve. 

Instead of chasing links anywhere, brands should increasingly target:

• Pages already being cited by LLMs in their category.
• Reviews or guides that evaluate their product category.
• Articles where branded mentions reinforce entity associations.

The core principle holds: brands gain the most visibility by appearing in sources LLMs already trust – and identifying those sources requires consistent tracking.

On-page: What your own content reveals

The same technologies that analyze third-party mentions can also reveal which first-party assets, content on your own website, are being cited by LLMs. 

This provides valuable insight into what type of content performs well in your space.

For example, these tools can identify:

• What types of competitor content are being cited (case studies, FAQs, research articles, etc.).
• Where your competitors show up but you don’t.
• Which of your own pages exist but are not being cited.

From there, three key opportunities emerge:

• Missing content: Competitors are cited because they cover topics you haven’t addressed. This represents a content gap to fill.
• Underperforming content: You have relevant content, but it isn’t being referenced. Optimization – improving structure, clarity, or authority – may be needed.
• Content enhancement opportunities: Some pages only require inserting specific Q&A sections or adding better-formatted information rather than full rewrites.

Leverage emerging technologies to turn insights into action

The next major evolution in LLM optimization will likely come from tools that connect insight to action.

Early solutions already use vector embeddings of your website content to compare it against LLM queries and responses. This allows you to:

• Detect where your coverage is weak.
• See how well your content semantically aligns with real LLM answers.
• Identify where small adjustments could yield large visibility gains.

Current tools mostly generate outlines or recommendations.

The next frontier is automation – systems that turn data into actionable content aligned with business goals.

Timeline and expected results

While comprehensive LLM visibility typically builds over 6-12 months, early results can emerge faster than traditional SEO. 

The advantage: LLMs can incorporate new content within days rather than waiting months for Google’s crawl and ranking cycles. 

However, the fundamentals remain unchanged.

Quality content creation, securing third-party mentions, and building authority still require sustained effort and resources. 

Think of LLM optimization as having a faster feedback loop than SEO, but requiring the same strategic commitment to content excellence and relationship building that has always driven digital visibility.

From SEO foundations to LLM visibility

LLM traffic remains small compared to traditional search, but it’s growing fast.

A major shift in resources would be premature, but ignoring LLMs would be shortsighted. 

The smartest path is balance: maintain focus on SEO while layering in LLM strategies that address new ranking mechanisms.

Like early SEO, LLM optimization is still imperfect and experimental – but full of opportunity. 

Brands that begin tracking citations, analyzing third-party mentions, and aligning SEO with LLM visibility now will gain a measurable advantage as these systems mature.

In short:

• Identify the third-party sources most often cited in your niche and analyze patterns across AI engines.
• Map competitor visibility for key LLM queries using tracking tools.
• Audit which of your own pages are cited (or not) – high Google rankings don’t guarantee LLM inclusion.
• Continue strong SEO practices while expanding into LLM tracking – the two work best as complementary layers.

Approach LLM optimization as both research and brand-building.

Don’t abandon proven SEO fundamentals. Rather, extend them to how AI systems discover, interpret, and cite information.

Seattle is looking to celebrate and accelerate its leadership in artificial intelligence at the very moment the first wave of the AI economy is crashing down on the region’s tech workforce.

That contrast was hard to miss Monday evening at the opening reception for Seattle AI Week 2025 at Pier 70. On stage, panels offered a healthy dose of optimism about building the AI future. In the crowd, buzz about Amazon’s impending layoffs brought the reality of the moment back to earth.

A region that rose with Microsoft and then Amazon is now dealing with the consequences of Big Tech’s AI-era restructuring. Companies that hired by the thousands are now thinning their ranks in the name of efficiency and focus — a dose of corporate realism for the local tech economy.

The double-edged nature of this shift is not lost on Washington Gov. Bob Ferguson.

“AI, and the future of AI, and what that means for our state and the world — each day I do this job, the more that moves up in my mind in terms of the challenges and the opportunities we have,” Ferguson told the AI Week crowd. He touted Washington’s concentration of AI jobs, saying his goal is to maximize the benefits of AI while minimizing its downsides.

Seattle AI Week, led by the Washington Technology Industry Association, was started last year after a Forbes list of the nation’s top 50 AI startups included none from Seattle, said the WTIA’s Nick Ellingson, opening this year’s event. That didn’t seem right. Was it a messaging problem?

“A bunch of us got together and said, let’s talk about all the cool things happening around AI in Seattle, and let’s expand the tent beyond just tech things that are happening,” Ellingson explained.

So maybe that’s the best measuring stick: how many startups will this latest shakeout spark, and how can the Seattle region’s startup and tech leaders make it happen? Can the region become less dependent on the whims of the Microsoft and Amazon C-suites in the process? 

“Washington has so much opportunity. It’s one of the few capitals of AI in the world,” said WTIA’s Arry Yu in her opening remarks. “People talk about China, people talk about Silicon Valley — there are a few contenders, but really, it’s here in Seattle. … The future is built on data, on powerful technology, but also on community. That’s what makes this place different.”

And yet, “AI is a sleepy scene in Seattle, where people work at their companies, but there’s very little activity and cross-pollinating outside of this,” said Nathan Lambert, senior research scientist with the Allen Institute for AI, during the opening panel discussion.

No, we don’t want to become San Francisco or Silicon Valley, Lambert added. But that doesn’t mean the region can’t cherry-pick some of the ingredients that put Bay Area tech on top.

Whether laid-off tech workers will start their own companies is a common question after layoffs like this. In the Seattle region at least, that outcome has been more fantasy than reality. 

This is where AI could change things, if not with the fabled one-person unicorn then with a bigger wave of new companies born of this employment downturn. Who knows, maybe one will even land on that elusive Forbes AI 50 list. (Hey, a region can dream!)

But as the new AI reality unfolds in the regional workforce, maybe the best question to ask is whether Seattle’s next big thing can come from its own backyard again.

Microsoft and OpenAI announced the long-awaited details of their new partnership agreement Tuesday morning — with concessions on both sides that keep the companies aligned but not in lockstep as they move into their next phases of AI development.

Under the arrangement, Microsoft gets a 27% equity stake in OpenAI’s new for-profit entity, the OpenAI Group PBC (Public Benefit Corporation), a stake valued at approximately $135 billion. That’s a decrease from 32.5% equity but not a bad return on an investment of $13.8 billion.

At the same time, OpenAI has contracted to purchase an incremental $250 billion in Microsoft Azure cloud services. However, in a significant concession in return for that certainty, Microsoft will no longer have a “right of first refusal” on new OpenAI cloud workloads.

Microsoft, meanwhile, will retain its intellectual property rights to OpenAI models and products through 2032, an extension of the timeframe that existed previously. 

A key provision of the new agreement centers on Artificial General Intelligence (AGI), with any declaration of AGI by OpenAI now subject to verification by an independent expert panel. This was a sticking point in the earlier partnership agreement, with an ambiguous definition of AI potentially triggering new provisions of the prior arrangement. 

Microsoft and OpenAI had previously announced a tentative agreement without providing details. More aspects of the deal are disclosed in a joint blog post from the companies.

Shares of Microsoft are up 2% in early trading after the announcement. The company reports earnings Wednesday afternoon, and some analysts have said the uncertainty over the OpenAI arrangement has been impacting Microsoft’s stock. 

In cybersecurity, speed isn’t just a win — it’s a multiplier. The faster you learn about emerging threats, the faster you adapt your defenses, the less damage you suffer, and the more confidently your business keeps scaling. Early threat detection isn’t about preventing a breach someday: it’s about protecting the revenue you’re supposed to earn every day. Companies that treat cybersecurity as a

HBO Max is giving fans an extra Halloween treat this year as IT: Welcome to Derry episode 2 will arrive sooner than expected.

Discover how digital tools can modernize the UK’s energy grid and accelerate the journey to net zero.

Recently fixed ASP.NET Core bug affects QNAP NetBak PC Agent, so needs fixing now.

Huge job cuts on the way at Amazon as it looks to shed thousands of employees.

Amazon is slashing prices on its best-selling TVs ahead of Black Friday, and I'm rounding up my 15 favorite deals from Samsung, LG, TCL, and more.

You can count on Coco Belliveau, Kevin Jacobs and Ty McDonald to liven up Manoir Rouville-Campbell. Here's how to watch The Traitors Canada season 3 online.

Uniting the innovation agenda with a strong foundation of trust and resilience offers immense opportunities.

Here's how to download the Battlefield 6 REDSEC battle royale mode.

AI is driving explosive data growth and tape storage is powering the exabyte era.

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.

Making music? These are the headphones to look out for on Black Friday

Travel eSIM revenue is rising sharply as mobile operators prepare to challenge smaller competitors and protect their international roaming profits.

The IT firm was not the source of the breach, but it’s been replaced nonetheless.

A zero-day in Google Chrome was used to deploy commercial spyware against targets in Russia.

This month's Blu-ray Bounty explores 5 new titles that could make their way into your 4k Blu-ray collection.

Gaming accessory brand KontrolFreek has unveiled its brand new Call of Duty: Black Ops 7 thumbsticks.

TechRadar reveals when The Witcher's fourth season will make its debut on Netflix.

Lenovo's early Black Friday sale includes everything from super-cheap Windows 11 machines to high-eng gaming laptops: here are today's best deals.

Firaxis has shared its plans for Civilization 7's post-launch content, which includes an update that has been highly requested since the game's release.

Nissan says it has made major leaps in solid-state battery chemistries, hinting that we could see cars with the tech in the coming years.

Searching for the best dash cam for your car? We've rounded up the finest 4K, front-and-rear and budget choices around.

Don't get scared now...

Do Windows 11 driver updates sometimes confuse or even freak you out? Microsoft assures us the OS can handle driver oddities.

Heaps of Hue lights have hit record-low prices, so why wait for Black Friday?

The JBL Flip 7 is the best Bluetooth speaker you can buy, and it just hit a record-low price ahead of Black Friday.

Sensational stories like mass Gmail password leaks might not always be what they seem.

We love the Garmin Venu 4, but we think this Venu 3 deal is much better value for a smartwatch with similar smarts.

Perfect for PS5, Xbox, or Switch 2, the SteelSeries Arctis GameBuds are our favorite gaming earbuds and are down to their lowest-ever price right now.

Score a great deal on the Beats Pill, now down to just $99.95 in all colorways (even Champagne gold)

AccessGrid builds APIs that companies can use to manage digital key fobs directly within Apple and Google's wallet platforms.

The idea of memory for AI isn’t new, but it’s quickly becoming a critical battleground.

This is the first time Apple's market capitalization has crossed the $4 trillion mark, making it the third company to ever cross the milestone after Nvidia and Microsoft.

OpenAI has completed its recapitalization, splitting the AI lab into a for-profit corporation nested inside a non-profit foundation.

Second day of TechCrunch Disrupt 2025 at San Francisco's Moscone West. Here's the rundown to what to expect, including the 50% discount on passes for remainder of the event.

Amazon said on Tuesday that it plans to reduce its corporate workforce by 14,000 jobs as it seeks to reduce bureaucracy, remove layers and invest more in its AI strategy.

Taiwan Semiconductor Industry Association pressures the government for plentiful and green energy

Nvidia’s DGX Spark, the company’s new $4,000 developer box powered by the Grace Blackwell GB10 superchip, is under fire after questions were raised about real-world performance and power draw.

Nvidia is hosting GTC and its keynote to reveal what's next in agentic AI, robotics, and accelerated computing.

OpenAI has called on the US to build out more power-generating infrastructure, claiming that it is needed to help provide the backbone for the AI race the US is now in with China. With enormous infrastructure projects planned, it wants the US to build an additional 100 gigawatts of new energy capacity every year.

The AirSense UAV drone is constructed from eco-friendly packaging boards and paper fiber.

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

PNY launches the CS3250, a PCIe 5.0 SSD with speeds up to 14.9 GB/s at a starting price of $125.99

The DRAM supply chain is choking, and server memory is taking the first hit, with major U.S. and Chinese hyperscalers now receiving just 70% of the server DRAM they order.

The post AMD CEO on new $1 billion AI supercomputer partnership with the Department of Energy appeared first on StartupHub.ai.

“We are super excited to announce a new partnership with the Department of Energy,” stated Lisa Su, Chair and CEO of AMD, during a CNBC interview. This monumental $1 billion collaboration will usher in the development of two advanced supercomputers, designed to tackle some of the most complex scientific challenges facing humanity. The partnership signifies […]

The post AMD CEO on new $1 billion AI supercomputer partnership with the Department of Energy appeared first on StartupHub.ai.

The post Bitcoin Miners’ AI Pivot: A Strategic Masterclass in Energy and Compute appeared first on StartupHub.ai.

The burgeoning demand for artificial intelligence, a computational arms race among hyperscalers, has illuminated a critical bottleneck: access to reliable, scalable power. This very challenge, as discussed by CleanSpark CEO Matthew Schultz with CNBC’s Jordan Smith, is precisely where Bitcoin miners like CleanSpark find their strategic advantage. Their conversation unveils a nuanced pivot, not merely […]

The post Bitcoin Miners’ AI Pivot: A Strategic Masterclass in Energy and Compute appeared first on StartupHub.ai.

The post OpenAI Recapitalization Reshapes AI Landscape with Microsoft at the Helm appeared first on StartupHub.ai.

The recent finalization of OpenAI’s recapitalization plan marks a pivotal moment in the trajectory of artificial intelligence, not just for the involved parties but for the entire tech ecosystem. On CNBC, David Faber broke down the intricate details of this agreement, joined by Jim Cramer, who offered his characteristic sharp market commentary. Their discussion illuminated […]

The post OpenAI Recapitalization Reshapes AI Landscape with Microsoft at the Helm appeared first on StartupHub.ai.

The post Wild Moose Emerges from Stealth with $7 Million Seed Round to Redefine Site Reliability Engineering with AI appeared first on StartupHub.ai.

Wild Moose, the AI-powered Site Reliability Engineering (SRE) platform acting as a first responder for production incidents, today announced its emergence from stealth with $7 million in seed funding. The round was led by iAngels, with participation from Y Combinator, F2 Venture Capital, Maverick Ventures, and others. The company is also backed by a distinguished […]

The post Wild Moose Emerges from Stealth with $7 Million Seed Round to Redefine Site Reliability Engineering with AI appeared first on StartupHub.ai.

The post Gemini for Education: Google’s AI Dominates Higher Ed appeared first on StartupHub.ai.

Google's Gemini for Education is rapidly integrating into higher education, offering no-cost AI tools to over 1000 institutions and 10 million students.

The post Gemini for Education: Google’s AI Dominates Higher Ed appeared first on StartupHub.ai.

The post Securitize IPO to bring tokenization to Nasdaq at $1.25B appeared first on StartupHub.ai.

The Securitize IPO is a bellwether moment, creating the first publicly-traded company focused purely on the infrastructure for tokenizing real-world assets.

The post Securitize IPO to bring tokenization to Nasdaq at $1.25B appeared first on StartupHub.ai.

The post FriendliAI Expands Ultra-Fast AI Inference Platform with Nebius AI Cloud Integration appeared first on StartupHub.ai.

Enterprises can now deploy large-scale AI inference with FriendliAI’s optimized stack on Nebius AI infrastructure, combining top performance with cost efficiency.

The post FriendliAI Expands Ultra-Fast AI Inference Platform with Nebius AI Cloud Integration appeared first on StartupHub.ai.

Seems like we've been on a northern lights roll lately, haven't we? There's one happening again tonight and the NOAA is even promising how the spectacular aurora display will be visible as far south as New York and Wisconsin. Forecasts say that a significant burst of solar activity colliding with Earth’s magnetic field will be creating the

After kids, teens, and even some adults are finished knocking on your door in hopes of scoring pieces of candy (or full-size candy bars, if you're really looking to impress), it won't be long until the Black Friday and Cyber Monday sales season is in full force. If you want to get a jump on the mad rush and are in the market for a Fire TV,

Cybersecurity experts are sounding the alarm over a new Android Trojan dubbed Herodotus, which is designed to deliberately slow down its own malicious activity to mimic the casual, imperfect behavior of a human user. Such behavior allows the malware to slip past a generation of security systems built to flag more rapid, robotic actions of

Windows Insiders are getting early access to an upcoming Windows 11 feature that is intended to improve system reliability by scanning system memory following a system crash. The optional feature is related to those frustrating BSOD errors that, aside from letting you know that your PC tripped over itself and bonked its head (as was the case

Samsung is expanding its range of storage card solutions with the release of the P9 Express, a microSD Express memory card with enough speed to keep up with gaming handhelds like Nintendo's Switch 2. While not exclusively aimed at the Switch 2, Nintendo's newest console is a big reason why we're starting to see more microSD Express memory

The ESA is trying to raise awareness of the Earth's alarming lack of preparation for the worst possible solar events.

Developed by Eta Space, the proof-of-concept satellite is part of the development of in-orbit cryogenic refueling stations set to come online by 2030.

TSMC has enough stock to last for a couple of years, if needed.

The Trend Micro Zero Day Initiative has uncovered three more security vulnerabilities affecting the X.Org Server and the derived XWayland source code...

An improvement to Intel SGX slated for Linux 6.18 is supporting the EUPDATESVN found on Intel CPUs since the Ice Lake generation. EUPDATESVN allows for updating the security SVN version after run-time patching for addressing any Intel SGX vulnerabilities to avoid having to carry out a platform reboot...

Earlier this month Intel announced Crescent Island as a Xe3P graphics card with 160GB of vRAM optimized for AI inferencing at the enterprise scale. Crescent Island isn't expected to begin sampling until H2'2026, but already for the upcoming Linux 6.19 kernel initial Crescent Island support is being submitted for the Xe kernel graphics driver...

Server memory has become the scarcest commodity in tech. According to DigiTimes, Samsung and SK Hynix quietly circulated fourth-quarter contract appendix that retroactively increased RDIMM prices by 40-50%. Even hyperscalers that signed agreements in August must now pay the new rate or risk losing their queue position. The two Korean manufacturers simultaneously reduced confirmed allocations by 30%, pushing Tier-1 U.S. and Chinese cloud order books to an effective 70% fill rate and eliminating the safety stock most buyers believed they had secured. Module manufacturers such as Kingston and ADATA now pay $13 for 16 GB DDR5 chips that cost $7 six weeks ago, an increase large enough to erase entire gross margin.

Smaller OEMs and channel distributors have been told to expect 35-40% fulfillment through the first quarter of 2026, forcing them either to gamble on the spot market or idle production lines. Even the older DDR4, now reduced to 20% of global DRAM output, is affected. Switches, routers, and set-top boxes that still use DDR4 are suddenly facing very long lead times because no fabrication plant wants to allocate wafers to trailing nodes. Analysts at TrendForce now forecast that the DRAM shortfall will outlast the 2026 hyperscaler build-out, meaning the industry's next relief valve may not be new capacity but a potential demand contraction, an outcome no manufacturer is willing to budget for.

Corsair today announced two new PCIe 5.0 NVMe M.2 SSDs that extend high‑end performance across multiple form factors: the MP700 PRO XT, built for uncompromising Gen 5 speed, and the MP700 MICRO, a power‑efficient M.2 2242 drive sized for thin‑and‑light devices and space‑constrained builds. Utilizing a next-generation controller with cutting-edge internals, the Corsair MP700 PRO XT pushes PCIe Gen 5 storage to its limits. Featuring sequential read speeds up to 14,900 MB/s and sequential write speeds up to 14,500 MB/s, it offers incredible boot, load, and transfer times. Its power-efficient design ensures top-tier performance with lower power draw and less heat.

With support for Microsoft DirectStorage, this SSD can communicate directly with a GPU for unbeatable gaming speed, enabling snappier game loads and smoother in‑game transitions. Thanks to the versatility of Corsair SSD Toolbox software and a five-year warranty, the MP700 PRO XT will deliver cutting-edge storage performance for years to come for performance desktop and laptop platforms. Initially available in 1 TB, 2 TB and 4 TB capacities, additional capacity options are expected to become available in early 2026.

Corsair, maker of award-winning gaming peripherals, today revealed a collaboration with CD Projekt RED celebrating the 10th anniversary of The Witcher 3: Wild Hunt. The Corsair x The Witcher 3: Wild Hunt 10th Anniversary Collection commemorates a decade of killing monsters with a limited run of officially licensed gaming peripherals. Fans can relive their monster slaying adventures with a new set of gear, thoughtfully recrafted with iconic touches from the game.

"Fans worldwide have been spellbound by Geralt of Rivia's quest to find Ciri, and we're thrilled to help them complete their own quest to find the ideal gear to celebrate 10 years of this legendary game," said Tobias Brinkmann, Vice President and General Manager of Gaming Peripherals at Corsair. "We're proud to have the opportunity to work with CD Projekt RED to bring this collaboration to life and to let fans showcase their love for this iconic game with our high-performance peripherals featuring a captivating new design."

Kensington, a worldwide leader of desktop computing and mobility solutions for IT, business, and home office professionals, today announced the Expert Mouse TB800 EQ Trackball, a next-generation, finger-operated ergonomic trackball that delivers premium performance and precise control for creative professionals and consumers.

A recipient of a prestigious 2025 iF Design Award, the TB800 enhances user comfort and productivity through advanced customization options, an adjustable scroll ring, exceptional tracking, an ambidextrous design, and seamless wireless connectivity. Featuring an innovative symmetrical design that is ideal for use by left- and right-handed users, the TB800 incorporates a slight angle that enables users to operate the trackball with their hands in a natural resting position that maximizes comfort and potentially reduces fatigue and repetitive strain injuries.

LG Electronics (LG) today announced the signing of a memorandum of understanding (MOU) with SK Enmove (SKEN) and Green Revolution Cooling (GRC) to jointly develop and expand next-generation liquid immersion cooling solutions optimized for artificial intelligence data centers (AIDCs). Under the agreement, the three companies will collaborate to explore new business opportunities, conduct joint marketing and deliver proof of concept (PoC) demonstrations for integrated solutions and business models in liquid immersion cooling for AIDCs.

SKEN, a leading company in advanced lubricating base oils and premium lubricants, provides next-generation thermal management solutions, including immersion cooling fluids, and is strengthening its global partnerships to foster an immersion cooling ecosystem. In 2022, SKEN became the first company in Korea to develop immersion cooling fluids. Through an equity investment in GRC, SKEN is also leading the way in innovation in the future cooling market, jointly developing data center immersion cooling systems with the company.

AMD will power two new AI supercomputers in the US, using MI355X and MI430X accelerators The US Department of Energy has announced a $1 billion deal under which AMD will deliver two next-generation supercomputers to the Oak Ridge National Laboratory (ORNL). These systems are designed to expand the US’s leadership in artificial intelligence (AI) and […]

The post US cuts a $1 billion deal with AMD to build two new AI Supercomputers appeared first on OC3D.

OpenAI has provided some internal numbers related to users who seek help from GPT-5, and they're a lot bigger than I was expecting. The AI firm has laid out details as to how it's combating the issue.

Will you need to buy Battlefield 6 in order to play its new battle royale mode, Redsec? The answer lies within.

Newegg is offering a special 71% discount for the first game in The Outer Worlds series, giving players an affordable means to experience this grim yet funny Sci-fi universe, in case they want to experience what the original game is like before diving into The Outer Worlds 2, or can't afford it.

Copilot is more than just an app on Windows 11, and in this guide, I'll outline the steps to get started with the AI chatbot and actions with the app and the integration across the desktop.

Newegg is offering a limited-time 31% discount for the Logitech G515 Lightspeed, along with a $15 discount code for it, giving PC gamers a chance to own this slickly designed, well-performing TKL keyboard for less.

Speaking with Halo Studios, the developers working on Halo: Campaign Evolved confirmed that the game's new prequel missions are entirely new.

By Jeff Seibert

I’ve been building products and companies my entire career — Increo, Box, Crashlytics, Twitter and now, Digits — and I’ve had the privilege of speaking with some of the sharpest minds in venture and entrepreneurship along the way.

One recent conversation with a legendary investor really crystallized for me a set of truths about startups: what success really is, why some founders thrive while others burn out, and how to navigate the inevitable chaos of building something from nothing.

Here are some of the lessons I’ve internalized from years of building, observing and learning.

Success has no finish line

In the startup world, we talk a lot about IPOs, acquisitions and valuations. But those are milestones, not destinations.

The companies that endure don’t “win” and stop — they keep creating, adapting and pushing forward. They’re playing an infinite game, where the only goal is to remain in the game.

When you’re building something truly generative — driven by a purpose greater than yourself — there’s no point at which you can say “done.” If your company has a natural stopping point, you may be building the wrong thing.

You don’t choose the work — the work chooses you

The best founders I’ve met — and the best moments I’ve had as a founder — come from an almost irrational pull toward solving a specific problem I myself experienced.

You may want to start a company, but if you have to talk yourself into your idea, it probably won’t survive contact with reality. The founders who succeed are often the ones who can’t not work on their thing.

Starting a company shouldn’t be a career move — it should be the last possible option after every other path fails to scratch the itch.

The real killer: founder fatigue

Most companies don’t die because of one bad decision or one tough competitor. They die because the founders run out of energy.

Fatigue erodes vision, motivation and creativity. Protecting your own drive — keeping it clean and focused — may be the single most important survival skill you have.

That means staying close to the product, protecting time for customer work, and avoiding the slow drift into managing around problems instead of solving them.

Customer > competitor

It’s easy to get caught up in competitor moves, investor chatter or market gossip. But the most important question is always: Are we delivering joy to the customer?

If you’re losing focus, sign up for your own product as a brand-new user. Feel the friction. Fix it. Repeat.

At Digits, we run our own signup and core flows every week. It’s uncomfortable — it surfaces flaws we’d rather not see — but it keeps us anchored to the only metric that matters: customer delight.

Boards should ask questions, not give answers

Over the years, I’ve learned the most effective boards aren’t presentation theaters — they’re discussion rooms.

The best structure I’ve seen:

• No slides;
• A narrative pre-read sent in advance; and
• A deep dive into one essential question.

Good directors help you widen your perspective. They don’t hand you a to-do list. Rather, they help you see the problem in a way that makes the answer obvious.

Twitter: lessons from a phenomenon

When I think back to my time at Twitter, the most enduring lesson is that not all companies are built top-down. Some — like Twitter — are shaped more by their users than their executives.

Features like @mentions, hashtags and retweets didn’t come from a product roadmap — they came from the community.

That’s messy, but it’s also powerful. Sometimes your job isn’t to control the phenomenon, rather it’s to keep it healthy without smothering what made it magical in the first place.

Why now is a great time to start

If you’re building today, you have an advantage over the so-called “unicorn zombies” that raised massive rounds pre-AI and are now locked into defending old business models.

Fresh founders can design from scratch for the new reality; there’s no legacy to protect, no sacred cows to defend.

The macro environment? Irrelevant. The only timing that matters is when the problem calls you so strongly that not working on it feels impossible.

If there’s one takeaway from all of this, it’s that success is continuing. The real prize is the ability to keep playing, keep serving and keep creating.

If you’re standing at the edge, wondering if you should start — start. Take one step. See if it grows. And if it does, welcome to the infinite game.

---

 Jeff Seibert is the founder and CEO of Digits, the world’s first AI-native accounting platform. He previously served as Twitter‘s head of consumer product and starred in the Emmy Award-winning Netflix documentary “The Social Dilemma.”

Illustration: Dom Guzman

While startup investment has been climbing lately, not all industries are partaking in the gains.

Cleantech is one of the spaces that’s been mostly left out. Overall funding to the space is down this year, despite some pockets of bullishness in areas like fusion and battery recycling.

The broad trend: Cleantech- and sustainability-related startup investment has been on a downward trajectory for several years now. And so far, 2025 is on track to be another down year.

On the bright side, however, there’s been some pickup in recent months, boosted by big rounds for companies in energy storage, fusion and other cleantech subsectors.

The numbers: Investors put an estimated $20 billion into seed- through growth-stage funding to companies in cleantech, EV and sustainability-related categories so far this year.

That puts 2025 funding on track to come in well below last year’s levels, which were already at a multiyear low.

Still, quarter by quarter, the pattern looks more encouraging. Investment hit a low point in Q1 of this year and recovered some in the subsequent two quarters. The current quarter is also off to a strong start.

Noteworthy recent rounds

The largest cleantech-related round of the year closed this month. Base Power, a provider of residential battery backup systems and electricity plans, raised $1 billion in Series C funding. The Austin, Texas-based company says its systems allow energy providers to more efficiently harness renewable power.

The second-largest round was Commonwealth Fusion Systems’ $863 million Series B2 financing. The Devens, Massachusetts-based company says it is moving closer to being the first in the world to commercialize fusion power.

For a bigger-picture view, below we put together a list of 10 of the year’s largest cleantech- and sustainability-related financings.

The broad takeaway: Startups innovating for an era of rising power consumption

Not to over-generalize, but if there was one big takeaway from recent cleantech and sustainability startup funding, it would be that founders and investors recognize that these are times of ever-escalating energy demand. They’re planning accordingly, looking to tap new sources of power, fusion in particular, as well as better utilize and scale existing clean energy sources.

Related Crunchbase query and list:

• Cleantech, EV & Sustainability-Related Funding, 2025 YTD
• Large 2025 Financings For Cleantech, EV And Sustainability Startups

Illustration: Dom Guzman

Elon Musk has introduced Grokipedia, a digital encyclopedia that relies on artificial intelligence rather than human editors to compile and update entries. The new platform, developed through his artificial intelligence company xAI, marks his most direct challenge yet to Wikipedia, a site written and curated by human volunteers that he...

Read Entire Article

We thought the RTX 5060 Ti 16GB was a better-than-average product, giving it a score of 75 and calling it not great but certainly not terrible. The $150 8GB model, however, earned a pitiful score of 30 due to its memory buffer, which is no longer satisfactory for PC gaming...

Read Entire Article

Battlefield 6's battle royale mode will launch as a free, separate executable on Tuesday at 11am ET. A launch trailer detailing the new content will unlock at the same time, and there is no preload.

Read Entire Article

Speaking in the Shacknews documentary 24 'Til Launch: The Making of Borderlands 4, Pitchford talked about the massive amounts of money that go into developing big games, which have budgets that match or even exceed Hollywood blockbusters.

Read Entire Article

YouTuber Max Tech conducted a series of controlled comparisons using identical 14-inch MacBook Pro configurations running the M4 and M5 chips. Both machines featured 512GB of storage and identical cooling assemblies, each featuring a single fan, one heat pipe, and two NAND flash modules.

Read Entire Article

A new report exploring the potential for the Pacific Northwest to stake its claim as the global leader in responsible AI offers a paradoxical view. The Cascadia region, which includes Seattle, Portland and Vancouver, B.C., is described as a proven, promising player in the sphere — but with significant risks that threaten its success.

“We created companies that transformed global commerce,” writes former Gov. Chris Gregoire in a forward to the document. “Now we have the chance to add another chapter — one where Cascadia becomes the world’s standard-bearer for innovation that uplifts both people and planet.”

The Cascadia Innovation Corridor, which Gregoire chairs, released the report this morning as it kicks off its two-day conference. The economic advocacy group’s eighth annual event is being held in Seattle.

The study is built on an analysis by the Boston Consulting Group that ranks Cascadia’s three metro areas against 15 comparable regions in the U.S. and Canada for their economic competitiveness, including livability, workforce, and business and innovation climate. Seattle came in fourth behind Boston, Austin and Raleigh, while Portland ranked 13th and Vancouver 14th.

Over the past decade, the region’s gross domestic product and populations have both grown significantly, and when combined, their economies approach the 18th largest in the world.

Cascadia’s strengths, the report explains, include tech engines such as cloud giants Microsoft and Amazon in Washington, silicon chip manufacturing in Oregon, and quantum innovation in Vancouver, as well as academic excellence from the University of Washington, University of British Columbia and Oregon State University.

But as time goes on and as business and civic leaders aim for the prize of AI dominance, cracks in the system are increasingly troubling.

• Business costs are rising and there are mounting regulatory concerns — but it’s a tricky picture. Seattle, for example, often turns to B&O and headcount taxes to cover costs, while the state struggles to balance budgets in the absence of an income tax.
• Housing affordability is continuing to decline for many residents in these metro areas.
• Skilled tech workers are leaving Portland, in particular, and Seattle relies heavily on foreign workers receiving H1-B visas, which are less certain under the Trump administration.
• The clean, affordable energy that was once abundant in the Pacific Northwest is decreasingly available as droughts reduce river flows that drive hydropower dams and electricity demand increases with rapid data center growth.

The report notes that multiple regions around the U.S. and Canada have created AI-focused hubs with hundreds of millions of dollars in public and private funding to bolster their hold on the sector.

New Jersey has a half-billion dollar “AI Moonshot” program including tax incentives and public-worker AI training programs; New York’s “Empire AI Consortium” has an AI computing training center at the University of Buffalo and startup supports; and California has a public-private task force to increase AI adoption within government services and connecting tech leaders with state agencies.

For its part, Seattle Mayor Bruce Harrell announced a “responsible AI plan” this fall that provides guidelines for the municipality’s use of artificial intelligence and its support of the AI tech sector as an economic driver, which includes the earlier launches of the startup-focused AI House and Foundations.

But what the region really needs to succeed is a collaborative effort tapping all of the metro areas’ assets.

“For Cascadia, the lesson is clear: without a coordinated strategy that links our strengths in cloud computing, semiconductors, and research, we risk falling behind,” states the Cascadia Innovation Corridor report. “Acting together, we can position Cascadia not just to keep pace, but to lead.”

With the iPhone 17 lineup now in the hands of consumers, the legendary rumor mill that typically gyrates towards Apple's new products is naturally pivoting towards the next year's iPhone 18 lineup as well as the much-anticipated iPhone 20, which is due in 2027 and would commemorate 20 years since the first iPhone launched all the way back in 2007. Now, a new rumor suggests that Apple is transitioning towards simplified buttons in stages, with the iPhone 18 lineup transitioning towards a less complicated mechanical button for camera control, which will be replaced entirely by solid-state buttons in the iPhone […]

Read full article at https://wccftech.com/apple-iphone-18-to-use-a-simplified-camera-control-button-iphone-20-to-feature-haptics-instead-of-mechanical-buttons/

Death Stranding 2: On the Beach now supports the PlayStation 5's power saver mode, and its implementation is among the most interesting to date, according to a new technical analysis. In the latest episode of their weekly podcast, the tech experts at Digital Foundry examined how the two entries in the Kojima Productions series support Power Saver Mode, a newly introduced operating mode for the PlayStation 5 console that cuts CPU resources in half, halves the memory bandwidth, and reduces CPU and GPU clocks to reduce the system's power consumption. While the implementation in Death Stranding: Director's Cut was not […]

Read full article at https://wccftech.com/death-stranding-2-on-the-beach-most-interesting-power-saver-mode/

The INSPIRE series RTX 5050 is probably the smallest RTX 5050 editions, which offer a single fan design and weigh just 551 grams. MSI Launches Small Form-Factor RTX 5050 INSPIRE ITX and OC GPUs, Boasting Dual-Slot Thickness MSI has officially launched two new GeForce RTX 5050 cards in the INSPIRE series. These are probably the smallest RTX 5050 cards on the market, boasting a dual-slot design and a single-fan cooler to ensure compatibility with very small ITX cases. Apart from MSI, PNY also has a similarly compact GeForce RTX 5050, which measures just 147mm. The INSPIRE ITX RTX 5050 cards […]

Read full article at https://wccftech.com/msi-intros-geforce-rtx-5050-inspire-itx-and-oc-cards-measuring-just-147mm/

EA is pushing its employees to use AI for basically every task, but the results can be flawed, resulting in more work for developers. Business Insider recently talked with current EA staff, who confirmed that the company's leadership has spent the past year or so pushing its 15,000 employees to use AI for virtually every task, from producing code and concept art for games to advising managers how to speak to staff about a certain number of topics, including pay or promotions. The AI tools used to produce code are among those creating the most issues for developers. It is […]

Read full article at https://wccftech.com/ea-is-pushing-employees-to-use-ai-for-everything-including-producing-code-requiring-manual-fixing/

Intel's CEO, Lip-Bu Tan, has discussed the stake taken by the US government in the company, claiming that it was a necessary step to ensure that the American chipmaker could compete with Taiwan's TSMC. Intel's CEO Also Tells Specifics About His Meeting With President Trump, Calling It a Massive Success Well, the interest from the Trump administration in Intel was indeed a surprise for many of us, but for CEO Lip-Bu Tan, this initiative was "good to have", as he claims that it is similar to how Taiwan supports TSMC or South Korea backs the likes of Samsung Foundry. In […]

Read full article at https://wccftech.com/intel-ceo-says-us-government-stake-was-a-deliberate-move/

OnexPlayer has officially launched its flagship handheld, the OneXfly Apex, with a liquid-cooled AMD Ryzen AI MAX+ 395 SoC. AMD Ryzen AI MAX+ 395 Gets Liquid-Cooled Inside A Handheld With OneXPlayer's OneXfly Apex The OneXfly Apex handheld was teased last month and is positioned to be a flagship device featuring the AMD Ryzen AI MAX+ 395 SoC. This SoC has already been featured in other handhelds such as GPD Win 5 and Ayaneo Next 2. Now, OneXPlayer is rolling out its own high-end handheld, offering a nice upgrade vs the Ryzen AI 300 "Strix Point" stack. Just to recap the […]

Read full article at https://wccftech.com/onexfly-apex-handheld-launch-amd-ryzen-ai-max-395-liquid-cooling-128-gb-85wh-battery/

The popular PS3 emulator has updated its latest GPU recommendation list to AMD's RDNA and NVIDIA's Turing series. RPCS3 Announces Updated GPU Requirements for the Emulator; Recommends At Least AMD RX 5000 or NVIDIA RTX 2000 Series RPCS3 has just announced the new recommended GPU requirements for its popular PS3 emulator, which comes as a result of major GPU manufacturers ending support for some of its older generation GPU series. RPCS3 announced on X that it no longer "recommends" the AMD RX 400 and NVIDIA GTX 900 series GPUs as the recommended GPUs. The newer GPU recommendations now start with […]

Read full article at https://wccftech.com/rpcs3-removes-amd-rx-400-500-and-nvidia-gtx-900-1000-series-from-recommended-gpu-requirements/

A vapor chamber will make a significant difference to the overall temperatures of the M6 iPad Pro, with Apple previously reported to bring this cooling upgrade to its flagship tablet lineup. The California-based giant is often known to commence product development several months in advance, and according to the latest report, Apple is already in talks with two suppliers that could manufacture this crucial component. The M6 iPad Pro’s vapor chamber is reported to be provided by a Chinese and Taiwanese manufacturer Considering that the M6 iPad Pro launch will materialize approximately 18 months after the M5 iPad Pro’s inception, […]

Read full article at https://wccftech.com/apple-shortlisting-m6-ipad-pro-vapor-chamber-suppliers/

Mockin is a tool for simulating professional real-time AI interviews with personalized feedback and resume matching.

We’ve gathered 200+ real interview questions to practice, get personalized feedback, and master the STAR method to ace your next interview! Built for designers by designers.

View startup

While enterprises looking to sell goods and services online wait for the backbone of agentic commerce to be hashed out, PayPal is hoping its new features will bridge the gap.

The payments company is launching a discoverability solution that allows enterprises to make its product available on any chat platform, regardless of the model or agent payment protocol. 

PayPal, which is one of the participants for Google’s Agent Payments Protocol (AP2), found that it can leverage its relationship with merchants and enterprises to help pave the way for an easier transition into agentic commerce and offer the kind of flexibility they learned will benefit the ecosystem. 

Michelle Gill, PayPal general manager for small business and financial services, told VentureBeat that AI-powered shopping will continue to grow, so enterprises and brands need to start laying the groundwork early. 

“We think that merchants who've historically sold through web stores, particularly in the e-commerce space, are really going to need a way to get active on all of these large language models,” Gill said. “The challenge is that no one really knows how fast all of this is going to move. The issue that we’re trying to help merchants think through is how to do all of this as low-touch as possible while using the infrastructure you already have without doing a bazillion integrations.”

She added AI shopping would also bring about “a resurgence from consumers trying to ensure their investment is protected.”

PayPal partnered with website builder Wix, Cymbio, Commerce and Shopware to bring products to chat platforms like Perplexity. 

Agent-powered shopping 

PayPal’s Agentic Commerce Services include two features. The first is Agent Ready, which would allow existing PayPal merchants to accept payments on AI platforms. The second is called Shop Sync, which will enable companies’ product data to be discoverable through different AI chat interfaces. It takes a company’s catalog information and plug its inventory and fulfillment data to chat platforms. 

Gill said the data goes into a central repository where AI models can ingest the information. 

Right now, companies can access shop sync with Agent Ready coming in 2026. 

Gill said Agentic Commerce Services is a one-to-many solution, that would be helpful right now, as different LLMs scrape different data sources to surface information. 

Other benefits include:

• Fast integration with current and future partners

• More product discovery over the traditional search, browse and cart experiences

• Preserved customer insights and relationships where the brand continues to have control over their records and communications with customers. 

Right now, the service is only available through Perplexity, but Gill said more platforms will be added soon. 

Fragmented AI platforms 

Agentic commerce is still very much in the early stages. AI agents are just beginning to get better at reading a browser. while platforms like ChatGPT, Gemini and Perplexity can now surface products and services based on user queries, people cannot technically buy things from chat yet.

There’s a race right now to create a standard to enable agents to transact on behalf of users and pay for items. Other than Google’s AP2, OpenAI and Stripe have the Agentic Commerce Protocol (ACP) and Visa launched its Trusted Agent Protocol. 

Other than enabling a trust layer for agents to transact, another issue enterprises face with agentic commerce is fragmentation. Different chat platforms use different models which also interpret information in slightly different ways. Gill said PayPal learned that when it comes to working with merchants, flexibility is important. 

“How do you decide if you're going to spend your time integrating with Google, Microsoft, ChatGPT or Perplexity? And each one of them right now has a different protocol, a different catalog, config, a different everything. That is a lot of time to make a bet as to like where you should spend your time,” Gill said. 

Google Arts & Culture is launching a new collection to honor the rich heritage of the province of North Gyeongsang in South Korea.

The Mill, the storied production studio that shut down abruptly earlier this year, is relaunching under new ownership across Europe and Asia. 

While the industry obsesses over automation and cost savings, they’re overlooking three huge shifts in how consumers use AI that marketers must understand in order to harness the technology.

Longtime Haworth Marketing + Media chair Gary Tobey is stepping down after three decades

2025 is the year of testing agentic commerce.

A leaked deck shows paid ads on Reddit can boost AI search visibility as the platform becomes more relevant than ever.

AI tools can help teams move faster than ever – but speed alone isn’t a strategy.

As more marketers rely on LLMs to help create and optimize content, credibility becomes the true differentiator. 

And as AI systems decide which information to trust, quality signals like accuracy, expertise, and authority matter more than ever.

It’s not just what you write but how you structure it. AI-driven search rewards clear answers, strong organization, and content it can easily interpret.

This article highlights key strategies for smarter AI workflows – from governance and training to editorial oversight – so your content remains accurate, authoritative, and unmistakably human.

Create an AI usage policy

More than half of marketers are using AI for creative endeavors like content creation, IAB reports.

Still, AI policies are not always the norm. 

Your organization will benefit from clear boundaries and expectations. Creating policies for AI use ensures consistency and accountability.

Only 7% of companies using genAI in marketing have a full-blown governance framework, according to SAS.

However, 63% invest in creating policies that govern how generative AI is used across the organization. 

Even a simple, one-page policy can prevent major mistakes and unify efforts across teams that may be doing things differently.

As Cathy McPhillips, chief growth officer at the Marketing Artificial Intelligence Institute, puts it: 

• “If one team uses ChatGPT while others work with Jasper or Writer, for instance, governance decisions can become very fragmented and challenging to manage. You’d need to keep track of who’s using which tools, what data they’re inputting, and what guidance they’ll need to follow to protect your brand’s intellectual property.” 

So drafting an internal policy sets expectations for AI use in the organization (or at least the creative teams).

When creating a policy, consider the following guidelines: 

• What the review process for AI-created content looks like. 
• When and how to disclose AI involvement in content creation. 
• How to protect proprietary information (not uploading confidential or client information into AI tools).
• Which AI tools are approved for use, and how to request access to new ones.
• How to log or report problems.

Logically, the policy will evolve as the technology and regulations change. 

Keep content anchored in people-first principles

It can be easy to fall into the trap of believing AI-generated content is good because it reads well. 

LLMs are great at predicting the next best sentence and making it sound convincing. 

But reviewing each sentence, paragraph, and the overall structure with a critical eye is absolutely necessary.

Think: Would an expert say it like that? Would you normally write like that? Does it offer the depth of human experience that it should?

“People-first content,” as Google puts it, is really just thinking about the end user and whether what you are putting into the world is adding value. 

Any LLM can create mediocre content, and any marketer can publish it. And that’s the problem. 

People-first content aligns with Google’s E-E-A-T framework, which outlines the characteristics of high-quality, trustworthy content.

E-E-A-T isn’t a novel idea, but it’s increasingly relevant in a world where AI systems need to determine if your content is good enough to be included in search.

According to evidence in U.S. v. Google LLC, we see quality remains central to ranking:

• “RankEmbed and its later iteration RankEmbedBERT are ranking models that rely on two main sources of data: [redacted]% of 70 days of search logs plus scores generated by human raters and used by Google to measure the quality of organic search results.” 

It suggests that the same quality factors reflected in E-E-A-T likely influence how AI systems assess which pages are trustworthy enough to ground their answers.

So what does E-E-A-T look like practically when working with AI content? You can:

• Review Google’s list of questions related to quality content: Keep these in mind before and after content creation.
• Demonstrate firsthand experience through personal insights, examples, and practical guidance: Weave these insights into AI output to add a human touch.
• Use reliable sources and data to substantiate claims: If you’re using LLMs for research, fact-check in real time to ensure the best sources. 
• Insert authoritative quotes either from internal stakeholders or external subject matter experts: Quoting internal folks builds brand credibility while external sources lend authority to the piece.
• Create detailed author bios: Include:
  • Relevant qualifications, certifications, awards, and experience.
  • Links to social media, academic papers (if relevant), or other authoritative works.
• Add schema markup to articles to clarify the content further: Schema can clarify content in a way that AI-powered search can better understand.
• Become the go-to resource on the topic: Create a depth and breadth of material on the website that’s organized in a search-friendly, user-friendly manner. You can learn more in my article on organizing content for AI search.

Dig deeper: Writing people-first content: A process and template

Train the LLM 

LLMs are trained on vast amounts of data – but they’re not trained on your data. 

Put in the work to train the LLM, and you can get better results and more efficient workflows. 

Here are some ideas.

Maintain a living style guide

If you already have a corporate style guide, great – you can use that to train the model. If not, create a simple one-pager that covers things like:

• Audience personas.
• Voice traits that matter.
• Reading level, if applicable.
• The do’s and don’ts of phrases and language to use. 
• Formatting rules such as SEO-friendly headers, sentence length, paragraph length, bulleted list guidelines, etc. 

You can refresh this as needed and use it to further train the model over time. 

Build a prompt kit  

Put together a packet of instructions that prompts the LLM. Here are some ideas to start with: 

• The style guide
  • This covers everything from the audience personas to the voice style and formatting.
  • If you’re training a custom GPT, you don’t need to do this every time, but it may need tweaking over time. 
• A content brief template
  • This can be an editable document that’s filled in for each content project and includes things like:
    • The goal of the content.
    • The specific audience.
    • The style of the content (news, listicle, feature article, how-to).
    • The role (who the LLM is writing as).
    • The desired action or outcome.
• Content examples
  • Upload a handful of the best content examples you have to train the LLM. This can be past articles, marketing materials, transcripts from videos, and more. 
  • If you create a custom GPT, you’ll do this at the outset, but additional examples of content may be uploaded, depending on the topic. 
• Sources
  • Train the model on the preferred third-party sources of information you want it to pull from, in addition to its own research. 
  • For example, if you want it to source certain publications in your industry, compile a list and upload it to the prompt.  
  • As an additional layer, prompt the model to automatically include any third-party sources after every paragraph to make fact-checking easier on the fly.
• SEO prompts
  • Consider building SEO into the structure of the content from the outset.  
  • Early observations of Google’s AI Mode suggest that clearly structured, well-sourced content is more likely to be referenced in AI-generated results.

With that in mind, you can put together a prompt checklist that includes:

• Crafting a direct answer in the first one to two sentences, then expanding with context.
• Covering the main question, but also potential subquestions (“fan-out” queries) that the system may generate (for example, questions related to comparisons, pros/cons, alternatives, etc.).
• Chunking content into many subsections, with each subsection answering a potential fan-out query to completion.
• Being an expert source of information in each individual section of the page, meaning it’s a passage that can stand on its own.
• Provide clear citations and semantic richness (synonyms, related entities) throughout. 

Dig deeper: Advanced AI prompt engineering strategies for SEO

Create custom GPTs or explore RAG 

A custom GPT is a personalized version of ChatGPT that’s trained on your materials so it can better create in your brand voice and follow brand rules. 

It mostly remembers tone and format, but that doesn’t guarantee the accuracy of output beyond what’s uploaded.

Some companies are exploring RAG (retrieval-augmented generation) to further train LLMs on the company’s own knowledge base. 

RAG connects an LLM to a private knowledge base, retrieving relevant documents at query time so the model can ground its responses in approved information.

While custom GPTs are easy, no-code setups, RAG implementation is more technical – but there are companies/technologies out there that can make it easier to implement. 

That’s why GPTs tend to work best for small or medium-scale projects or for non-technical teams focused on maintaining brand consistency.

RAG, on the other hand, is an option for enterprise-level content generation in industries where accuracy is critical and information changes frequently.

Run an automated self-review

Create parameters so the model can self-assess the content before further editorial review. You can create a checklist of things to prompt it.

For example:

• “Is the advice helpful, original, people-first?” (Perhaps using Google’s list of questions from its helpful content guidance.) 
• “Is the tone and voice completely aligned with the style guide?” 

Have an established editing process 

Even the best AI workflow still depends on trained editors and fact-checkers. This human layer of quality assurance protects accuracy, tone, and credibility.

Editorial training

About 33% of content writers and 24% of marketing managers added AI skills to their LinkedIn profiles in 2024.

Writers and editors need to continue to upskill in the coming year, and, according to the Microsoft 2025 annual Work Trend Index, AI skilling is the top priority.  

Professional training creates baseline knowledge so your team gets up to speed faster and can confidently handle outputs consistently.

This includes training on how to effectively use LLMs and how to best create and edit AI content.

In addition, training content teams on SEO helps them build best practices into prompts and drafts.

Editorial procedures

Ground your AI-assisted content creation in editorial best practices to ensure the highest quality. 

This might include:

• Identifying the parts of the content creation workflow that are best suited for LLM assistance.
• Conducting an editorial meeting to sign off on topics and outlines. 
• Drafting the content.
• Performing the structural edit for clarity and flow, then copyediting for grammar and punctuation.
• Getting sign-off from stakeholders.  

The AI editing checklist

Build a checklist to use during the review process for quality assurance. Here are some ideas to get you started:

• Every claim, statistic, quote, or date is accompanied by a citation for fact-checking accuracy.
• All facts are traceable to credible, approved sources.
• Outdated statistics (more than two years) are replaced with fresh insights. 
• Draft meets the style guide’s voice guidelines and tone definitions. 
• Content adds valuable, expert insights rather than being vague or generic.
• For thought leadership, ensure the author’s perspective is woven throughout.
• Draft is run through the AI detector, aiming for a conservative percentage of 5% or less AI. 
• Draft aligns with brand values and meets internal publication standards.
• Final draft includes explicit disclosure of AI involvement when required (client-facing/regulatory).

Grounding AI content in trust and intent

AI is transforming how we create, but it doesn’t change why we create.

Every policy, workflow, and prompt should ultimately support one mission: to deliver accurate, helpful, and human-centered content that strengthens your brand’s authority and improves your visibility in search. 

Dig deeper: An AI-assisted content process that outperforms human-only copy

The New Reality for Lean Security Teams If you’re the first security or IT hire at a fast-growing startup, you’ve likely inherited a mandate that’s both simple and maddeningly complex: secure the business without slowing it down. Most organizations using Google Workspace start with an environment built for collaboration, not resilience. Shared drives, permissive settings, and constant

The zero-day exploitation of a now-patched security flaw in Google Chrome led to the distribution of an espionage-related tool from Italian information technology and services provider Memento Labs, according to new findings from Kaspersky. The vulnerability in question is CVE-2025-2783 (CVSS score: 8.3), a case of sandbox escape which the company disclosed in March 2025 as having come under

We'd almost given up on new Fitbit hardware, but Google says we can look forward to new hardware in 2026.

Everest is claiming responsibility and says it stole hundreds of gigabytes of files.

Leaked code suggests Apple might launch four ‘pro’ apps on iPad following their success on the Mac.

The new Insta360 X4 Air has many of the great features of the X4 and X5 360 degree cameras, squeezed into a more portable package.

EU attempts to lessen US hyperscaler work might be doing more harm to local firms, CISPE says.

These stunning Xbox Wireless Controller designs are enjoying their best discount yet at Amazon.

You can currently grab a top rated pair of gaming headphones at a massive discount at Best Buy, proving that you don't need to wait until Black Friday for good deals.

The Google Pixel Watch 3 drops down to $199.99 at Amazon – a fantastic record-low price for the excellent smartwatch.

The Bing Wallpaper app gets awful new feature – and I fear Windows 11's becoming more about the 'Microsoft experience' than the user experience.

These three DualSense Edge competitors really offer a wide range of premium features. Definitely ones to look for in the upcoming Black Friday sales.

The rumors were true! Epic Games has confirmed that The Simpsons is officially coming to Fortnite for the very first time as the game's annual Fortnitemares Halloween event draws to a close.

OpenAI is reportedly training an AI music generator with the help of students from the prestigious Juilliard School

Refurbished tech done right

The next Pixel Feature Drop could land soon with theme packs, custom GIFs, and prioritized notifications.

In travel, even seconds of IT downtime can spiral into crises - resilience makes survival possible.

IT: Welcome to Derry is an entertaining prequel to Stephen King's iconic story and there's plenty for fans to enjoy.

Samsung has revealed the design of the Galaxy tri-fold, while elsewhere a seemingly official render has leaked.

A gripping three-part docuseries based on the Ex-Wives Undercover podcast, this guide explains how to watch Don’t Date Brandon online and potentially for free.

If you want to buy a powerful, sleek, and smart new Android phone in the upcoming Black Friday sales, I’d recommend the Google Pixel 9 Pro.

Here's our live coverage of the launch of Battlefield 6 Season One and Battlefield REDSEC release time. This includes the start time, what's being added, as well as info on the new Battlefield battle royale mode.

Adobe said on Tuesday that it is launching the latest iteration of its image generation model, Firefly Image 5. The company is also adding features to the Firefly website, support for more third-party models, and the ability to generate speech and soundtracks.

Adobe today released new AI assistants Creative Cloud products, Express and Photoshop, that can help users with image creation and editing.

Paypal said on Tuesday that it is adopting a protocol in combination with OpenAI's "Instant Checkout" feature to let users pay for their shopping directly within ChatGPT, starting in 2026.

Google's John Mueller said case sensitivity matters and SEOs shouldn't just hope it works.

The post Google’s Advice On Canonicals: They’re Case Sensitive appeared first on Search Engine Journal.

Can SanDisk’s Extreme Pro USB4 SSD compete with impressive options from Corsair and LaCie? We ran the drive through our benchmark suite to find out.

The latest data shows that almost 90% of Windows games now run on Linux.

A user on Reddit shunt-modded their Zephyrus M16's RTX 4090 laptop GPU, which led to a 20% bump in performance compared to stock, while even beating RTX 5090 mobile on average. This was achieved by just stacking one resistor atop the existing one to trick the GPU into consuming way more power than it thinks it is.

An unfortunate user on Reddit bought a 1TB Seagate Backup Plus Slim external hard drive, only to find out that they've been scammed. Inside the enclosure was an SD card for storage, with metal blocks attached to a fake weight.

SK hynix introduced its AI NAND lineup — AIN D, AIN P, and AIN B — at the 2025 Global Summit, outlining a new strategy to deliver high-density, high-performance, and high-bandwidth storage tailored for AI servers and workloads.

China's GDIIST research institute has announced the development and soon release of the BIE-1, an AI supercomputer inspired by the operation of the human brain. This neuromorphic computing tech is one of the first standalone, non-rack-based brain-based computers we've ever seen.

A CPU collector/enthusiast has taken the top off of an Intel 4004 processor and taken pictures of the chip's insides, showing the 10,000nm chip in all its glory.

Researchers at the University of Cambridge and Meta Reality Labs have conducted a new study on just how many pixels the human eye can take in at certain distances, and determined it's fewer than we might think. They claim in their results that it means most humans wouldn't be able to tell the difference between 1440p and 4K on a 50-inch screen at 10 feet distance.

The post Beyond LLMs: Crafting Robust AI with Multi-Method Agentic Architectures appeared first on StartupHub.ai.

“Large language models have well-known issues and constraints. And so if you want to solve complex problems, you’re going to want to adopt what’s called multi-method agentic AI, which combines large language models with other kinds of proven automation technologies so that you can build more adaptable, more transparent systems that are much more likely […]

The post Beyond LLMs: Crafting Robust AI with Multi-Method Agentic Architectures appeared first on StartupHub.ai.

The post Google Arts & Culture Elevates Virtual Travel with AI Tours appeared first on StartupHub.ai.

Google Arts & Culture is redefining virtual exploration with new AI tours for North Gyeongsang, South Korea, featuring interactive, Gemini-powered commentary.

The post Google Arts & Culture Elevates Virtual Travel with AI Tours appeared first on StartupHub.ai.

The post Nvidia’s AI Imperative: Beyond Moore’s Law, Network is the New Compute appeared first on StartupHub.ai.

Michael Kagan, CTO of Nvidia and co-founder of Mellanox, recently engaged in a candid discussion with Sonya Huang and Pat Grady at Sequoia’s Europe100 event, offering profound insights into Nvidia’s meteoric rise as the architect of AI infrastructure. His commentary illuminated the pivotal role of the Mellanox acquisition in transforming Nvidia from a mere chip […]

The post Nvidia’s AI Imperative: Beyond Moore’s Law, Network is the New Compute appeared first on StartupHub.ai.

It's Fedora 43 release day! This latest installment of Fedora Linux is now available for download with Fedora Workstation 43 using the GNOME 49 desktop, the modern Linux 6.17 kernel powering this distribution release, and many exciting improvements and other leading-edge software updates powering this Red Hat sponsored Linux distribution...

The upcoming Linux 6.19 kernel cycle is expected to land initial support for USB3 with Apple Silicon devices...

For those making use of Device Mapper's DM-VERITY target for transparent integrity checking of block devices, the upcoming Linux 6.19 kernel has an enticing performance optimization en route that for some processors can lead to nearly doubling the hashing performance...

Linaro engineer Adhemerval Zanella recently sent out a set of 59 patches to allow building the GNU C Library "glibc" with the LLVM Clang compiler as an alternative to GCC...

Stride 4.2.1.2485 is now available as the latest feature release for this open-source and cross-platform game engine written in C# while still having first-rate Linux support. Stride is formerly known as Xenko and offers realistic rendering and virtual reality (VR) support...

The land of Windows 11 is finally getting a feature most users will appreciate, with the introduction of the new memory scanning for issues after a blue screen of death (BSOD) happens. "We're introducing a new feature that helps improve system reliability. If your PC experiences a bug check (unexpected restart), you may see a notification when signing in suggesting a quick memory scan," noted Windows Insider Program lead Amanda Langowski. Additionally, the "If you choose to run it, the system will schedule a Windows Memory Diagnostic scan to run during your next reboot (taking 5 minutes or less on average) and then continue to Windows. If a memory issue is found and mitigated, you will see a notification post-reboot."

Microsoft notes that this first wave flags every bug check so they can watch how memory glitches turn into blue screens, and they will refine targeting of these issues in the later updates. At the moment the preview will not run on Arm64 PCs, machines that have Administrator Protection turned on, or any BitLocker setup that boots without Secure Boot enabled. Users that are part of the Windows Insiders Dev and Beta channels will be able to access this feature. Windows 11 Insider Preview Build 26220.6982 (KB5067109) and Windows 11 Insider Preview Build 26120.6982 (KB5067109) are the first in the rollout, so they can beta test the feature before it hits the main stable Windows 11 branch.

NVIDIA's DGX Spark machine, designed as the ultimate AI box for local and fast AI prototyping, is reportedly operating at half the expected power and performance levels. John Carmack, founder of AGI-focused Keen Technologies and former CTO of Oculus VR, claims that the DGX Spark mini PC is not meeting its specified performance. NVIDIA rates the DGX Spark mini PC at 240 W of system power, but Carmack's benchmarks indicate it only draws about 100 W, effectively halving the power draw and performance. The DGX Spark's peak throughput is approximately 31 TeraFLOPS for FP32 and around 1,000 TOPS with NVIDIA's NVFP4 reduced-precision format. At BF16 dense compute, it is supposed to achieve 125 TeraFLOPS, but these targets are not being met. The measured compute is about 480 TeraFLOPS at FP4 and only about 60 TeraFLOPS at BF16.

After facing multiple delays, NVIDIA's DGX Spark has finally reached developers. However, many are reporting software and firmware issues on NVIDIA's end. There may also be thermal throttling problems, causing the chip to reduce frequency and voltage to prevent overheating. In some cases, the system has rebooted, potentially due to inadequate cooling. The GB10 SoC is rated for a 140 W TDP, and the 128 GB configuration of LPDDR5X could add several dozen additional watts. Therefore, a 100 W power draw doesn't seem feasible for the DGX Spark. It remains to be seen whether a software or firmware update will address these issues, or if NVIDIA will provide an aftermarket cooling solution for its $3,999 machine if it continues to overheat.

Samsung Electronics today announced the launch of its new microSD Express card lineup, the P9 Express, designed to deliver next-gen gaming experiences and optimized for leading platforms, including the Nintendo Switch 2. Based on the PCIe interface and NVMe protocol, SD Express technology significantly enhances data
transfer performance compared to UHS-I cards, making it ideal for environments that demand high-capacity processing and fast data transmission.

The P9 Express is especially valuable for hardcore console gamers who frequently enjoy a diverse range of games and Downloadable Content (DLC), often demanding additional storage capacity beyond the internal storage. To meet different gaming needs, it is available in both 256 GB and 512 GB options. It also provides an ideal solution for multiple users sharing a single console, where ample capacity is required for several different game installations, helping gamers overcome limited internal storage and enjoy a wide variety of titles without compromise. When used with a dedicated SD Express interface, the P9 Express delivers sequential read speeds up to four times faster than UHS-I, enabling creators and professionals to efficiently move large volumes of data from devices to PCs, laptops, or workstations.

ASUS today announced the shipment of the XA NB3I-E12 AI server, built on the NVIDIA HGX B300 platform. Delivering next-generation AI performance and reliability, XA NB3I-E12 gives enterprises and cloud-service providers (CSPs) early access to cutting-edge computing capabilities for the AI era. Accelerated by eight NVIDIA Blackwell Ultra GPUs and dual Intel Xeon 6 Scalable processors, ASUS XA NB3I-E12 is engineered for intensive AI workloads. With eight NVIDIA ConnectX-8 InfiniBand SuperNICs, five PCIe expansion slots, 32 DIMMs, 10 NVMe drives, and dual 10 Gb LAN, it transforms data into intelligence for real-world automation. The system is ideal for enterprises and CSPs running large language models (LLMs), research institutions and universities performing scientific computing, and the financial and automotive sectors focused on AI model training and inference.

KLEVV, the leading consumer memory and storage brand introduced by Essencore, proudly unveils a striking new colorway for its award-winning URBANE V RGB DDR5 Gaming/OC memory. The sleek new Jet Black Edition joins the popular Brilliant White, broadening the lineup and giving enthusiasts more freedom to personalize their setups without compromise.

The URBANE V RGB Gaming/OC series is designed with both style and function in mind, featuring a 2 mm-thick aluminium heatsink with refined, curved edges and precision linear grooves that ensure durability and efficient cooling. With a low-profile height of just 42.5 mm, the modules fit seamlessly into diverse builds while maintaining optimal thermal performance. A distinctive dual-beam RGB light guide delivers smooth, customizable illumination across 16 million colors, fully compatible with major motherboard lighting software. This proven design, recognized with the prestigious iF Design Award, is now available in an elegant Jet Black finish that complements today's modern gaming and creator setups.

Amazon confirmed Tuesday that it is cutting about 14,000 corporate jobs, citing a need to reduce bureaucracy and become more efficient in the new era of artificial intelligence.

In a message to employees, posted on the company’s website, Amazon human resources chief Beth Galetti signaled that the cutbacks are expected to continue into 2026, while indicating that the company will also continue to hire in key strategic areas.

Reuters reported Monday that the number of layoffs could ultimately total as many as 30,000 people, which is still a possibility as the cutbacks continue into next year. At that scale, the overall number of job cuts could eventually be the largest in Amazon’s history, exceeding the 27,000 positions that the company eliminated in 2023 across multiple rounds of layoffs.

“This generation of AI is the most transformative technology we’ve seen since the Internet, and it’s enabling companies to innovate much faster than ever before,” wrote Galetti, senior vice president of People Experience and Technology. Amazon needs “to be organized more leanly, with fewer layers and more ownership, to move as quickly as possible for our customers and business,” she explained.

Amazon’s corporate workforce numbered around 350,000 people in early 2023, the last time the company provided a public number. At that scale, the initial reduction of 14,000 represents about 4% of Amazon’s corporate workforce. However, the number is a much smaller fraction of its overall workforce of 1.55 million people, which includes workers in its warehouses.

Although the cuts are expected to be global, they are likely to hit especially hard in the Seattle region, home to the company’s first headquarters and its largest corporate workforce. The tech hub has already felt the impact of major layoffs by Microsoft and many other companies in recent months.

The cuts come two days before Amazon’s third quarter earnings report. Amazon and other cloud giants have been pouring billions into capital expenses to boost AI capacity. Cutting jobs is one way of showing operating-expense discipline to Wall Street.

In a memo to employees in June, Amazon CEO Andy Jassy wrote that he expected Amazon’s total corporate workforce to get smaller over time as a result of efficiency gains from AI. 

Jassy took over as Amazon CEO from founder Jeff Bezos in mid-2021. In recent years he has been pushing to reduce management layers and eliminate bureaucracy inside the company, saying he wants Amazon to operate like the “world’s largest startup.” 

Bloomberg News reported this week that Jassy has told colleagues that parts of the company remain “unwieldy” despite the 2023 layoffs and other efforts to streamline operations. 

Reuters cited sources saying the magnitude of the cuts is also a result of Amazon’s strict return-to-office policy failing to cause enough employees to quit voluntarily. Amazon brought workers back five days a week earlier this year.

Impacted teams and people will be notified of the layoffs today, Galetti wrote.

Amazon is offering most impacted employees 90 days to find a new role internally, though the timing may vary based on local laws, according to the message. Those who do not find a new position at Amazon or choose to leave will be offered severance pay, outplacement services, health insurance benefits, and other forms of support.

John Carmack reports performance issues with Nvidia’s DGX Spark AI system John Carmack, ID Software founder and former CTO of Oculus VR, has been testing an Nvidia DGX Spark AI system. So far, he is not impressed by the performance the system has delivered. His system appears to be maxing out at 100 watts, which […]

The post Nvidia DGX Spark delivers half quoted performance for John Carmack appeared first on OC3D.

Battlefield is getting a free-to-play Battle Royale mode EA has confirmed that Battlefield REDSEC will launch on October 28th at 3 PM GMT, a free-to-play Battle Royale game that debuts alongside Battlefield 6’s Season 1 content. Battlefield REDSEC acts as EA’s counter to Call of Duty: Warzone. Currently, exact details for the new game are […]

The post Battlefield REDSEC is launching today – Here’s what you need to know appeared first on OC3D.

Update: Amazon has now said it will reduce its corporate workforce by about 14,000 people, with more cuts to follow next year. It has partly blamed the move on AI adoption, which some say has become a scapegoat for companies making layoffs.

Read Entire Article

There's little doubt that The Matrix franchise is criminally underserved when it comes to videogame adaptations, despite being theoretically a perfect fit for the medium. In the 26 years since the original movie's theatrical debut, we only got two decent games: 2003's single player action/adventure game Enter the Matrix and 2005's MMORPG The Matrix Online. More recently, the interactive experience The Matrix Awakens was released in late 2021, but it was really just a tech demo for Unreal Engine 5 and a tease at the level of quality that gaming fans of the IP never really got to fully experience. […]

Read full article at https://wccftech.com/the-matrix-creators-wanted-kojima-make-a-game-on-the-ip-konami-refused/

Today at GTC 2025, NVIDIA's CEO, Jensen Huang, will deliver the opening keynote live from Washington, US, for the first time. NVIDIA GTC Comes To Washington, D.C, US: CEO Jensen Huang To Talk About Next Chapter of AI, Watch It Live Here NVIDIA's GTC 2025 is just a few hours away, and while you might be wondering, didn't GTC already happen a few months back? Well, it should be mentioned that while GTC used to be a one-time per annum affair in the past, the recent growth and success have turned NVIDIA's GTC into more of a quarterly event. As […]

Read full article at https://wccftech.com/watch-nvidia-gtc-2025-ceo-jensen-huang-keynote-live-washington-us/

The iPhone 17 lineup is expected to be Apple’s last to ship with Qualcomm’s 5G modems as the company prepares its transition to ship all of its iPhone 18 models with the C2 baseband chip. This in-house solution was said to be in development shortly after the iPhone 16e was announced, and while we will witness its materialization in 2026, a new report states that, unlike other Apple chipsets like the A20 and A20 Pro, it will not leverage TSMC’s newest 2nm process, but a lithography that is a couple of generations old. The C2 5G modem will reportedly be mass […]

Read full article at https://wccftech.com/apple-c2-to-be-mass-produced-on-older-tsmc-process-says-report/

A team of modders is working on Bully Online, a modification for the PC version of Bully: Scholarship Edition that promises to allow players to roam the grounds of Bullworth Academy and the nearby town with their friends. The Wii and Xbox 360 versions of Scholarship Edition did have a multiplayer mode, but it was limited to two players and only allowed them to face off in the class minigames. According to community creator SWEGTA, Bully Online promises much more, including free roam support, solo and group minigames, and even a role-playing system. They were able to add a 'fully […]

Read full article at https://wccftech.com/bully-online-mod-promises-let-you-roam-rockstars-classic-with-friends/

This morning, indie Chinese developer ChillyRoom unveiled Loulan: The Cursed Sand, one of the games funded through the PlayStation China Hero Project. The game is a hack 'n' slash action RPG viewed from a Diablo-like camera. The setting is the ancient Silk Road, in China's Western Regions. Loulan: The Cursed Sand tells the tragic love story of an exiled royal guard who returns to the titular fallen kingdom amidst the chaos of war in search of his beloved princess. Players will step into the game as the skeletal warrior known as 'The Cursed Sand', mastering the power of sand as […]

Read full article at https://wccftech.com/loulan-the-cursed-sand-chinese-hack-n-slash-arpg/

Samsung looks to be all set to announce its first triple-folding smartphone, the Galaxy Z TriFold, and even though the device is expected to be limited to a few markets, it was high time that we saw smartphones gravitate to a new form factor. Just before the official announcement happens, a series of images provides a first look at the Galaxy Z TriFold, showing a dual-infolding structure that can transform into a large-screen tablet. The Galaxy Z TriFold was on display at the Samsung booth at the K-Tech Showcase, with one report stating that the prototype did not display any […]

Read full article at https://wccftech.com/samsung-galaxy-z-trifold-first-look-image-gallery/

Watch free tennis streams today as Sinner and Alcaraz compete in the season's last Masters event – 2025 Paris Masters TV channels, broadcasters and streams.

DJI’s new Romo vacuum brings advanced AI and drone-like autonomy to floor cleaning.

Featuring a powerful CPU/GPU combo, bright, high-resolution AMOLED screen, quad cameras, powerful camping light and huge capacity battery, this rugged smartphone is about as refined as they come. Designed for those who need a mobile device that can withstand the elements and go days, if not weeks, between charges.

While it's been succeeded in recent months by the X50 Ultra, this Dreame robot vacuum is one of the smartest robovacs you can buy – now down a whopping AU$1,259.

ChatGPT Go is now free for Indian users under a limited-time promotional offer.

The most powerful way to build professional mobile apps.

Discussion | Link

The workspace for music creators

Discussion | Link

Turn your AI tool into a personal email marketing assistant

Discussion | Link

Your LinkedIn post studio from sources to impact

Discussion | Link

Private voice reminders for Mac with no cloud or accounts

Discussion | Link

Organize & edit local files with AI from your browser

Discussion | Link

Know What You Ship. Secure What You Depend On.

Discussion | Link

Compare API models by benchmarks, cost & capabilities

Discussion | Link

Secure file uploads for Intercom, Crisp & Zendesk

Discussion | Link

Personal health coach built with Gemini

Discussion | Link

Elon's answer to Wikipedia

Discussion | Link

The first add-on to manage Google Sheets tabs efficiently

Discussion | Link

Automatically adds short poems to your photos, for free.

Discussion | Link

Vibe Your Agentic Workflows

Discussion | Link

The post On-Policy Distillation LLMs Redefine Post-Training Efficiency appeared first on StartupHub.ai.

On-policy distillation LLMs from Thinking Machines Lab offer a highly efficient and cost-effective method for post-training specialized smaller models, combining direct learning with dense feedback.

The post On-Policy Distillation LLMs Redefine Post-Training Efficiency appeared first on StartupHub.ai.

The post Tensormesh exits stealth with $4.5M to slash AI inference caching costs appeared first on StartupHub.ai.

Tensormesh's AI inference caching technology eliminates redundant computation, promising to make enterprise AI cheaper and faster to run at scale.

The post Tensormesh exits stealth with $4.5M to slash AI inference caching costs appeared first on StartupHub.ai.

News recently broke in a Rolling Stones interview that Halo Studios had relied on AI in the development of Halo: Campaign Evolved, with the game director, Greg Hermann, commented about "how integrated AI is becoming" in the "tooling" of the studio's development pipeline. Following this and other comments that implied AI was being heavily used in the remake of Halo: Combat Evolved, fans started assuming that varying degrees of the creative workload involved in the development of Campaign Evolved was being handled by generative AI. This comes after EA and Krafton both announced an increase in their reliance on generative AI in both the game development process and overall corporate management processes.

More recently, however, an Xbox representative responded directly to Rolling Stones, clarifying that "There is no mandate to use generative AI in our game development, and that includes Halo: Campaign Evolved," contrary to the situation facing many EA workers, who have allegedly been facing pressure to use AI tools for over a year. This response echoed the game director's prior insistence that generative AI is being used merely as a tool by the developers to improve general workflows, and that game development "really is about that creative spark that comes from people." Mentions of generative AI in video games are only becoming more frequent, and many online discussions surrounding AI indicate that it is fuelling an erosion of trust in game studios and developers.

Self Hosted Project is platform to launch your own VPN server in minutes—no tech skills needed. Full root access, 32+ locations, total privacy. No KYC, no logs, 7 protocols. Just your server. VPN done right. Self-hosted. Self-controlled.

View startup

Meta plans to lay off more than 100 employees in Washington state as part of a broader round of cuts within its artificial intelligence division.

A new filing with the state’s Employment Security Department shows 101 employees impacted, including 48 in Bellevue, 23 in Seattle, and four in Redmond, along with 23 remote workers based in Washington.

The filing lists dozens of affected roles across Meta’s AI research and infrastructure units, including software engineers, AI researchers, and data scientists. Meta product managers, privacy specialists, and compliance analysts were also affected.

Meta is cutting around 600 positions in its AI unit, Axios reported last week. The company is investing heavily in AI and wants to create a “more agile operation,” according to an internal memo cited by Axios. Meta has just under 3,000 roles within its superintelligence lab, CNBC reported.

The separations at Meta in Washington take effect Dec. 22, according to the Worker Adjustment and Retraining Notification (WARN) notice filed Oct. 22.

Meta employs thousands of people across multiple offices in the Seattle region, one of its largest engineering hubs outside Menlo Park.

The latest reductions mark another contraction for Meta’s Pacific Northwest footprint following multiple rounds of layoffs over the past several years.

The company’s rapid expansion in Seattle over the past decade made it one of the emblems of the region’s tech boom, coinciding with Microsoft’s resurgence and Amazon’s rise.

Among the Bay Area titans, Google was among the first to establish a Seattle-area engineering office, way back in 2004. However, it was Facebook’s decision to open its own outpost across from Pike Place Market in 2010 that really got the attention of their Silicon Valley tech brethren.

In the decade that followed, out-of-town companies set up more than 130 engineering centers in the region.

However, more recently Meta has made moves to trim its Seattle-area footprint.

Apple earlier this year took over a building previously occupied by Meta in Seattle’s South Lake Union neighborhood, near Amazon’s headquarters. CoStar reported in April that Meta listed its other Arbor Blocks building for sublease.

Meta previously gobbled up much of the planned office space at the Spring District, a sprawling development northeast of downtown Bellevue, including a building that was originally going to be a new REI headquarters. But it has subleased some of the space since then to companies such as Snowflake, which recently took an entire building from Meta at the Spring District.

Meta’s office in Redmond, near Microsoft’s headquarters, is focused on its mixed reality development.

GeekWire has reached out to the company for an updated Seattle-area headcount.

Meta’s cuts come amid reported layoffs at Amazon that could impact up to 30,000 workers.

Tech companies have laid off more than 128,000 employees this year, according to Layoffs.fyi. Last year, companies cut nearly 153,000 positions.

A European embassy located in the Indian capital of New Delhi, as well as multiple organizations in Sri Lanka, Pakistan, and Bangladesh, have emerged as the target of a new campaign orchestrated by a threat actor known as SideWinder in September 2025. The activity "reveals a notable evolution in SideWinder's TTPs, particularly the adoption of a novel PDF and ClickOnce-based infection chain, in

Looking more like a peacock chair than an office chair, the ErgoTune Vesby combines design, comfort and ergonomics well, but its prebuilt bits need better fitting.

A 55-inch TV ticks multiple boxes, including ideal size, availability and most importantly, price. These are the sets we recommend based on our testing.

Streaming services like Xbox Game Pass and PlayStation Plus provide gamers with a wealth of games to play with the obvious drawback that you don't get to keep the games for an extended period of time. PlayStation Plus's free monthly games, however, skip this caveat and allow players to keep the game as long as it's added to their library during the month the game is featured. According to supposed leaks by DeaLabs, November's monthly free game for PS Plus subscribers will be Stray, and it will be available to claim from November 4. After that, as long as players have a PS Plus subscription, they will be able to play Stray. Stray was previously available to play via a PS Plus subscription, but it was subsequently removed around the game's Xbox launch.

Stray is a single-player indie adventure game that originally launched in 2022 for the PS5, PS4, and PC via Steam, later launching on the Xbox Series X|S, macOS, and Nintendo Switch. The game follows an orange cat as it explores an underground cyberpunk city occupied exclusively by robots in an effort to find its way back to the surface. It largely features puzzle-platformer mechanics, with a particular focus on environmental puzzles. The game will apparently be the flagship title for the month of November on PS Plus, but there will be two other as-yet unknown free games joining Stray at the same time.

According to the rumor mill, the next Xbox console will have access to all Windows software and support all games currently available on Xbox Series consoles. The information matches earlier rumors regarding Microsoft's radical shift in strategy.

Read Entire Article

Quietnet blocks ads, tracking, and harmful websites before they even reach your phone, laptop, or TV — no apps needed, and it works for every device in your home or office.

We make the internet faster, safer, and more private for families and small businesses — without the noise. We're not backed by big tech or VC money. We're privacy-focused, bootstrapped, and already seeing people pay for peace and quiet online. QuietNet is built by people who care — no ads, no tracking, just a cleaner, safer internet for your family or team.

View startup

We're going over the differences between two excellent Razer controllers, and why one is the clear winner.

We're in Los Angeles for Adobe Max 2025 - here's what we've seen.

Sequoia's managing partner Roelof Botha noted that there are currently 3,000 venture firms in the United States, while there were just 1,000 when he joined Sequoia 20 years ago.

Teller says X has a 2% hit rate, which means that most of the things the company tries don't work out, and that's okay.

Zoom CEO Eric Yuan says that, in a few years, we should be working a 3-4 day workweek because of AI.

Sequoia's managing partner Roelof Botha defended his colleague's controversial comments that sparked an online backlash, arguing that the VC firm needs "spiky" people.

EA announced a battle royale mode was in development for Battlefield 6 a while ago, and now we know what it'll be called. RedSec releases tomorrow, alongside the game's first season update, and it will be completely free to play. More details will be revealed tomorrow, but we already know that RedSec will feature an insta-kill zone system, far less lenient than any other game.

Routers

OpenRazer 3.11 is out as the newest version of these out-of-tree but open-source and community-maintained drivers for Razer devices on Linux. Plus OpenRazer also provides a user-space daemon for controlling Razer RGB lighting and other features. Paired with the likes of the Polychromatic app, OpenRazer makes for a pleasant Razer device experience for gamers and enthusiasts under Linux...

The 2018 soft reboot of the God of War franchise saw Kratos face off against the gods of Norse mythology, but, as reports from earlier this year claimed, Sony was working on a multiplayer God of War game that we now know would have taken place in Ancient Greece, thanks to MP1st's insider sources. The supposed leak of screenshots of the cancelled God of War multiplayer project would have seemingly taken place before the events of God of War 3, since Hades was still supposedly actively involved in the game's events. The unreleased God of War game was being developed by Bluepoint Games, the same development studio behind Demon Souls and a handful of game remasters, including Titanfall on the Xbox 360.

Further details about the game, including how a multiplayer God of War game would look, are still unknown, but the screenshots provided by MP1st's source reveal that the game would be a return to Ancient Greece—or a territory conquered by the Greek empire. The game was seemingly still in very early development, but it is evident from the screenshots that the halls, rooms, and caverns were being designed for multiple players. Curiously, it seems as though some assets—weapons and statues that bear an uncanny resemblance to warriors and weapons from the Norse arcs—were also borrowed from the 2018 God of War reboot, although it's unclear if these were simply substitutes that would later be replaced or if they'd actually make it into the final game.

AMD gains $3 billion by divesting from ZT Systems’ manufacturing business – retains key talent AMD has confirmed that it has officially divested from ZT Systems’ manufacturing business, selling it to Sanmina for $3 billion. This recoups most of AMD’s acquisition costs from its ZT Systems (ZTS) purchase earlier this year, and secures AMD a […]

The post AMD gains $3 billion divesting from ZT Systems’ manufacturing business appeared first on OC3D.

Researchers at Kyung Hee University in South Korea have unveiled a new ultraviolet photodetector designed to improve protection against excessive UVA exposure. The fully transparent device measures UVA levels in real time and transmits an electrical signal to an external monitoring system, such as a phone app.

Read Entire Article

Voxtara is your personal AI speech coach that helps you become a more confident and effective public speaker. Whether you're preparing for a big presentation, teaching a class, or pitching to investors, Voxtara provides instant, actionable feedback to help you improve.

Key Features: • AI-Powered Analysis: Get comprehensive feedback on clarity, pacing, confidence, engagement, and body language • Video Recording: Record practice sessions up to 5 minutes • Deep-Dive Reports: Detailed analysis • Progress Tracking: Watch your speaking skills improve over time with detailed analytics • Practice Reminders: Set custom reminders • Session History: Review past performances

View startup

Today, we announced a collaboration with NextEra Energy to accelerate nuclear deployment across the U.S., including the restart of the Duane Arnold Energy Center in Iowa…

These quick tips could help you get more traction with your email outreach. 

With cyber scams on the increase, TikTok's looking to help raise awareness among its user community.

Paramount Skydance's cuts will happen on Wednesday.

At Brandweek in Atlanta, we're honoring the year’s visionary leaders and creators at the 2025 Brand Genius Celebration, presented by Shutterstock.

Even vehicles with over 100,000 miles on the clock lose very little of their original battery capacity, study says.

The Alienware Aurora (ACT1250) features a sleek design, improved airflow, RTX 50-series performance, and some sweet RGB for good measure.

The Legion Tower 5 stays cool and composed no matter what you throw at it – but can it match the best gaming PCs for performance?

A secret Dell clearance page exposes a wider range of discounted products, challenging the curated limits of the company’s official deals portal.

An interesting take on the Mini PC, the Beelink SER9 Pro on the surface offers a powerful yet balanced computing solution, with a design perfectly equipped for high-demand applications inluding content creation. However, it's the AI Voice kit integration that really stands out, making it an ideal option for the Office where voice notes and meetings need to be documented with clarity.

Microsoft has a new plan to ensure your Windows 11 PC crashes less often: proactive memory diagnostics.

Asus’s 8K ProArt PA32KCX merges accuracy, brightness, and calibration technology to serve professionals demanding precise HDR and rendering workflows.

This bird is cooked

Congo seeks to expand the Inga hydro complex to power AI tools and data centers amid ongoing funding and governance hurdles.

Expect to score bargain prices on the iPhone 16, iPhone 16 Pro, and iPhone 16 Pro Max.

Pytho AI wants to compress mission planning from days to mere minutes.

Strong by Form has designed a structural floor piece that can span longer distances than existing engineered wood, making it a replacement for steel or concrete. At the same time, the product is lighter than all three.

MacroCycle's approach to recycling dramatically reduces the amount of energy needed to produce new material, potentially lowering costs to the point where it could compete with virgin plastic.

Mbodi users prompt the software with natural language, and its cluster of AI agents works to make training easier.

Damoa pinpointed the problem: the complex, multi-step process moving a container from a ship to a freight train. He founded Glīd Technologies to try and solve it.

Waymo co-CEO Tekedra Mawakana believes the robotaxi company can increase road safety by reaching scale.

Waymo co-CEO Tekedra Mawakana opened up at TechCrunch Disrupt 2025 about the vandalism of its cars and how the company pushes back on government surveillance requests.

Waymo co-CEO Tekedra Mawakana urged rival autonomous vehicle companies to be more transparent about their safety data.

Amazon could be planning to cut up to 30,000 corporate positions starting on Tuesday.

OpenAI released data on just how many of ChatGPT's users are facing mental health challenges, and how it's addressing them.

Called “Coach,” this feature is designed to be your all-in-one fitness trainer, sleep coach, and health and wellness advisor.

Two 401(k) management unicorns, Human Interest and Guideline, are squaring off in federal court with allegations so brazen they're embarrassing.

YouTube launched Ask Studio, an AI assistant in YouTube Studio that analyzes channel data to surface comment insights, performance analysis, and content ideas.

The post YouTube Introduces ‘Ask Studio’ AI For Channel Analytics appeared first on Search Engine Journal.

OpenAI is telling companies that “relationship building” with AI has limits. Emotional dependence on ChatGPT is considered a safety risk, with new guardrails in place.

The post OpenAI Flags Emotional Reliance On ChatGPT As A Safety Risk appeared first on Search Engine Journal.

The post Edge AI: The Key to Sustainable AI Energy Efficiency appeared first on StartupHub.ai.

Arm and SCSP's new paper highlights edge computing as the strategic imperative for achieving AI energy efficiency and securing U.S. competitiveness.

The post Edge AI: The Key to Sustainable AI Energy Efficiency appeared first on StartupHub.ai.

The post America’s AI Future: AMD Powers U.S. Sovereign AI appeared first on StartupHub.ai.

AMD and the DOE are launching Lux and Discovery supercomputers at ORNL, a $1 billion investment to establish secure U.S. Sovereign AI infrastructure.

The post America’s AI Future: AMD Powers U.S. Sovereign AI appeared first on StartupHub.ai.

The post ASEAN’s AI Ambition: Infrastructure, Innovation, and Tailored Governance appeared first on StartupHub.ai.

“Infrastructure is destiny,” declared James Hairston, Head of International Policy & Partnerships for Asia, Africa, & Latin America at OpenAI, encapsulating the strategic imperative facing Southeast Asia in the burgeoning age of artificial intelligence. This powerful statement set the stage for a compelling discussion at the Bloomberg Business Summit at ASEAN in Kuala Lumpur, where […]

The post ASEAN’s AI Ambition: Infrastructure, Innovation, and Tailored Governance appeared first on StartupHub.ai.

Microsoft’s efforts to include its Copilot AI in as many of its services and products as possible has landed the company in some hot water. The Australian Competition and Consumer Commission (ACCC) is taking the company to court, accusing the software giant of misleading customers with its integration of Copilot into Microsoft 365 plans. This

Retroid is making waves in the handheld gaming space—a year after the September release of the Retroid Pocket 5, Retroid is following up with both a proper Retroid Pocket 6 and a Retroid Pocket G2, the latter of which serves as a souped-up refresh of Pocket 5 with the same external shell. Both handhelds are targeting a similar $200-$300 USD

Since the bombshell announcement of new Halo installment coming to PlayStation just two days ago, the wider public seems convinced that Xbox's battle in the console war is over. And now, comments made by Xbox Game Studios head Matt Booty seem to be lending credence to that, declaring TikTok a bigger competitor to Xbox than PlayStation. There's

It's leaner, lighter, and more capable, with the potential for additional missions after cargo transfer.

The child-sized Bumi can walk, dance, and assist with educational tasks for merely $1,400.

Meanwhile, many users say it's still tricky to find businesses on Apple Maps in the first place.

A big focus for the FreeBSD 15.0 development was on supporting reproducible builds as has been a growing trend in the open-source ecosystem in recent years. One month out from the official FreeBSD 15.0 release, the FreeBSD project is today celebrating having crossed the milestone of being able to be built reproducibly and as well now building FreeBSD without requiring root privileges...

Thermals don't take a winter break—that's why your PC cooling setup matters more than ever. Ocypus, a fast-emerging name in PC cooling, has made waves at trade shows over the past decade and is now widely available across the US and EU. TechPowerUp has teamed up with Ocypus for a pre-winter giveaway, open worldwide (wherever legal). It's your chance to score one of two standout Ocypus products—or both. The grand prize winner takes home an Ocypus Iota C70 case and Sigma Pro AIO CPU cooler, pictured below. The first runner-up gets the Sigma Pro, and the second runner-up scores the Iota C70 case.

The Ocypus Sigma Pro is the star attraction here, with its floating true-color display on top of its cuboidal pump-block that can be programmed to display just about anything. Ocypus bundles software that can display system monitoring info on this display. You can also program it with cool visualizations and game stats. The Iota C70 is a spacious, wide-format ATX mid-tower case with ample room for builds with high-end graphics cards and elaborate liquid coolers, such as the Sigma Pro. It comes with a pillarless front-left corner letting you revel on your creation. The Giveaway is open worldwide and runs until November 6th.

For more information, and to participate, visit this page.

MSI has officially announced the latest entries in its Inspire GPU line-up, the NVIDIA GeForce RTX 5050 Inspire ITX and the RTX 5050 Inspire ITX OC. As the name suggests, the RTX 5050 Inspire ITX line-up is designed for small form-factor PC builds, so both models feature the same compact two-slot design as the RTX 5060 version we recently reported on, measuring in at just 147 × 120 × 45 mm. Both versions of the GPU also feature the same cooling solution, with a single-fan cooler design with a Torx Fan 5.0 and a single heat pipe.

The only notable difference between the two versions is the clock speed, with the vanilla RTX 5050 Inspire ITX topping out at 2587 MHz in Extreme Performance mode (2572 MHz regular boost clocks), while the OC version is capable of 2617 MHz in Extreme Performance mode while boosting to 2602 MHz ordinarily. Both ITX cards feature three DisplayPort 2.1b ports and a single HDMI 2.1b port and have a rated power consumption of 130 W, supplied via a single 8-pin power connector. Users can select between Gaming and Silent modes in MSI Center or dive deeper into customization and overclocking in MSI Afterburner. No pricing information has been announced for the new ITX GPUs.

Obsidian's long-awaited Xbox Game Pass RPG The Outer Worlds 2 is almost upon us, but can you run it? Here are the requirements you need to know.

Microsoft has confirmed that Satya Nadella will be skipping Ignite 2025, with newly appointed CEO of commercial Judson Althoff leading the main keynote instead.

The Resident Evil Remake Trilogy, Resident Evil 7 and Village, and the timeless Resident Evil HD Remaster are all enjoying a special Halloween sale for PC at Loaded.

Amazon notes that last week's AWS outage began at 11:48 PM PDT on October 19, as users reported widespread errors and latency across services in the US-EAST-1 region (Northern Virginia). Amazon confirmed the disruption stemmed from DNS resolution problems with the DynamoDB API endpoint. While engineers mitigated the issue within...

Read Entire Article

Qualcomm is expanding beyond its roots in mobile technology, directly challenging the companies dominating artificial intelligence hardware. The semiconductor firm has announced that it will enter the high-end data center market with two new AI accelerator chips, marking its most ambitious move yet into the computing infrastructure that underpins the...

Read Entire Article

According to tipster Gray (@Olrak29_), the Zen 2 lineup includes four SKUs: Ryzen 5 40, Ryzen 3 30, Athlon Gold 20, and Athlon Silver 10. The Zen 3+ parts include Ryzen 7 170, Ryzen 7 160, Ryzen 5 150, Ryzen 5 130, and Ryzen 3 110. All SKUs feature RDNA 2 integrated graphics.

Read Entire Article

The Australian Competition & Consumer Commission (ACCC) is taking Microsoft to the country's federal court. The agency is suing Microsoft Australia and Microsoft Corporation, alleging that the company misled approximately 2.7 million Australian customers over subscription price changes and alternative renewal options.

Read Entire Article

OCCT 15 delivers a major update with a new storage test modeled after CrystalDiskMark, plus enhanced GPU diagnostics through an improved 3D adaptive test that detects errors more precisely and adds a coil-whine detection feature.

Read Entire Article

So, I stumbled upon this interesting portable monitor, which is unusually large to be called "portable", but considering there are users who would want something that can be useful for both travel and regular use, the UMax 24 looked interesting. I have reviewed a few UPERFECT portable monitors, including the Dual-Stack UStation Delta Max, which is one of the best options for work and gaming. However, UMax's big 24.5-inch screen size makes it an interesting option for daily usage if you are considering a versatile option for your desktop and travel. Personally, I wanted to see if I could replace […]

Read full article at https://wccftech.com/review/uperfect-umax-24-portable-monitor-review/

Apple will eventually introduce the M5 MacBook Air in a few months now that it has officially started selling the M5 MacBook Pro, but you will not immediately see the discounts on the company’s newer portable Macs, making the M4 MacBook Air models a more attractive proposition. Why? Because Amazon has slashed both the 13-inch and 15-inch versions by $200, and best of all, you can configure these machines up to 24GB unified RAM and a 512GB SSD. The base model starts from $799, making it an instant steal. The M5 MacBook Air will likely adopt the same ‘fanless’ cooling […]

Read full article at https://wccftech.com/do-not-wait-for-the-m5-macbook-air-because-amazon-offers-200-off-on-m4-macbook-air/

Halo: Campaign Evolved marked the official confirmation of something that felt like it would never happen just a short few years ago, with Halo officially coming to PlayStation. Following that announcement, GameStop, the retailer that was initially known for selling physical video games that's now more known for turning into a glorified Pop Funko and merch store called the 'Console Wars' over. Of all entities, the White House, which would normally have more important things to post about, responded with an AI-generated image of President Trump in Spartan armor with the caption "Power to the players." GameStop's original post comes […]

Read full article at https://wccftech.com/gamestop-console-wars-over-halo-on-ps-white-house-trump-ai-spartan-armor/

Gearbox founder Randy Pitchford was recently interviewed by Shacknews alongside a few other colleagues to discuss the making of Borderlands 4, the studio's latest game, which was released on September 12. The video runs for 73 minutes, and right toward the end, Pitchford goes into exactly what is needed to create such a big videogame like Borderlands 4. Interestingly, he then adds that the gaming industry as a whole is just getting started and 'figuring out' videogames, which haven't yet received their 'Citizen Kane' moment yet. To make a game like Borderlands 4 takes a big investment. It's a massive, […]

Read full article at https://wccftech.com/borderlands-boss-gaming-hasnt-even-produced-single-masterpiece-yet/

Qualcomm has announced its latest AI chips, which are designed to scale up to a purpose-built rack-level AI inference solution, but interestingly, they employ mobile memory onboard. Qualcomm's New AI Chips Take a 'Daring' Pivot Away From HBM To Target Efficient Inferencing Workloads Qualcomm has come a long way from being a mobile-focused firm, and in recent years, the San Diego chipmaker has expanded into new segments, including consumer computing and AI infrastructure. Now, the firm has announced its newest AI200 and AI250 chip solutions, which are reportedly designed for rack-scale configurations. This not only marks the entry of a […]

Read full article at https://wccftech.com/qualcomm-new-ai-rack-scale-solution-actually-uses-lpddr-mobile-memory-onboard/

Watch out, DeepSeek and Qwen! There's a new king of open source large language models (LLMs), especially when it comes to something enterprises are increasingly valuing: agentic tool use — that is, the ability to go off and use other software capabilities like web search or bespoke applications — without much human guidance.

That model is none other than MiniMax-M2, the latest LLM from the Chinese startup of the same name. And in a big win for enterprises globally, the model is available under a permissive, enterprise-friendly MIT License, meaning it is made available freely for developers to take, deploy, retrain, and use how they see fit — even for commercial purposes. It can be found on Hugging Face, GitHub and ModelScope, as well as through MiniMax's API here. It supports OpenAI and Anthropic API standards, as well, making it easy for customers of said proprietary AI startups to shift out their models to MiniMax's API, if they want.

According to independent evaluations by Artificial Analysis, a third-party generative AI model benchmarking and research organization, M2 now ranks first among all open-weight systems worldwide on the Intelligence Index—a composite measure of reasoning, coding, and task-execution performance.

In agentic benchmarks that measure how well a model can plan, execute, and use external tools—skills that power coding assistants and autonomous agents—MiniMax’s own reported results, following the Artificial Analysis methodology, show τ²-Bench 77.2, BrowseComp 44.0, and FinSearchComp-global 65.5.

These scores place it at or near the level of top proprietary systems like GPT-5 (thinking) and Claude Sonnet 4.5, making MiniMax-M2 the highest-performing open model yet released for real-world agentic and tool-calling tasks.

What It Means For Enterprises and the AI Race

Built around an efficient Mixture-of-Experts (MoE) architecture, MiniMax-M2 delivers high-end capability for agentic and developer workflows while remaining practical for enterprise deployment.

For technical decision-makers, the release marks an important turning point for open models in business settings. MiniMax-M2 combines frontier-level reasoning with a manageable activation footprint—just 10 billion active parameters out of 230 billion total.

This design enables enterprises to operate advanced reasoning and automation workloads on fewer GPUs, achieving near-state-of-the-art results without the infrastructure demands or licensing costs associated with proprietary frontier systems.

Artificial Analysis’ data show that MiniMax-M2’s strengths go beyond raw intelligence scores. The model leads or closely trails top proprietary systems such as GPT-5 (thinking) and Claude Sonnet 4.5 across benchmarks for end-to-end coding, reasoning, and agentic tool use.

Its performance in τ²-Bench, SWE-Bench, and BrowseComp indicates particular advantages for organizations that depend on AI systems capable of planning, executing, and verifying complex workflows—key functions for agentic and developer tools inside enterprise environments.

As LLM engineer Pierre-Carl Langlais aka Alexander Doria posted on X: "MiniMax [is] making a case for mastering the technology end-to-end to get actual agentic automation."

Compact Design, Scalable Performance

MiniMax-M2’s technical architecture is a sparse Mixture-of-Experts model with 230 billion total parameters and 10 billion active per inference.

This configuration significantly reduces latency and compute requirements while maintaining broad general intelligence.

The design allows for responsive agent loops—compile–run–test or browse–retrieve–cite cycles—that execute faster and more predictably than denser models.

For enterprise technology teams, this means easier scaling, lower cloud costs, and reduced deployment friction. According to Artificial Analysis, the model can be served efficiently on as few as four NVIDIA H100 GPUs at FP8 precision, a setup well within reach for mid-size organizations or departmental AI clusters.

Benchmark Leadership Across Agentic and Coding Workflows

MiniMax’s benchmark suite highlights strong real-world performance across developer and agent environments. The figure below, released with the model, compares MiniMax-M2 (in red) with several leading proprietary and open models, including GPT-5 (thinking), Claude Sonnet 4.5, Gemini 2.5 Pro, and DeepSeek-V3.2.

MiniMax-M2 achieves top or near-top performance in many categories:

• SWE-bench Verified: 69.4 — close to GPT-5’s 74.9

• ArtifactsBench: 66.8 — above Claude Sonnet 4.5 and DeepSeek-V3.2

• τ²-Bench: 77.2 — approaching GPT-5’s 80.1

• GAIA (text only): 75.7 — surpassing DeepSeek-V3.2

• BrowseComp: 44.0 — notably stronger than other open models

• FinSearchComp-global: 65.5 — best among tested open-weight systems

These results show MiniMax-M2’s capability in executing complex, tool-augmented tasks across multiple languages and environments—skills increasingly relevant for automated support, R&D, and data analysis inside enterprises.

Strong Showing in Artificial Analysis’ Intelligence Index

The model’s overall intelligence profile is confirmed in the latest Artificial Analysis Intelligence Index v3.0, which aggregates performance across ten reasoning benchmarks including MMLU-Pro, GPQA Diamond, AIME 2025, IFBench, and τ²-Bench Telecom.

MiniMax-M2 scored 61 points, ranking as the highest open-weight model globally and following closely behind GPT-5 (high) and Grok 4.

Artificial Analysis highlighted the model’s balance between technical accuracy, reasoning depth, and applied intelligence across domains. For enterprise users, this consistency indicates a reliable model foundation suitable for integration into software engineering, customer support, or knowledge automation systems.

Designed for Developers and Agentic Systems

MiniMax engineered M2 for end-to-end developer workflows, enabling multi-file code edits, automated testing, and regression repair directly within integrated development environments or CI/CD pipelines.

The model also excels in agentic planning—handling tasks that combine web search, command execution, and API calls while maintaining reasoning traceability.

These capabilities make MiniMax-M2 especially valuable for enterprises exploring autonomous developer agents, data analysis assistants, or AI-augmented operational tools.

Benchmarks such as Terminal-Bench and BrowseComp demonstrate the model’s ability to adapt to incomplete data and recover gracefully from intermediate errors, improving reliability in production settings.

Interleaved Thinking and Structured Tool Use

A distinctive aspect of MiniMax-M2 is its interleaved thinking format, which maintains visible reasoning traces between <think>...</think> tags.

This enables the model to plan and verify steps across multiple exchanges, a critical feature for agentic reasoning. MiniMax advises retaining these segments when passing conversation history to preserve the model’s logic and continuity.

The company also provides a Tool Calling Guide on Hugging Face, detailing how developers can connect external tools and APIs via structured XML-style calls.

This functionality allows MiniMax-M2 to serve as the reasoning core for larger agent frameworks, executing dynamic tasks such as search, retrieval, and computation through external functions.

Open Source Access and Enterprise Deployment Options

Enterprises can access the model through the MiniMax Open Platform API and MiniMax Agent interface (a web chat similar to ChatGPT), both currently free for a limited time.

MiniMax recommends SGLang and vLLM for efficient serving, each offering day-one support for the model’s unique interleaved reasoning and tool-calling structure.

Deployment guides and parameter configurations are available through MiniMax’s documentation.

Cost Efficiency and Token Economics

As Artificial Analysis noted, MiniMax’s API pricing is set at $0.30 per million input tokens and $1.20 per million output tokens, among the most competitive in the open-model ecosystem.

Provider	Model (doc link)	Input $/1M	Output $/1M	Notes
MiniMax	MiniMax-M2	$0.30	$1.20	Listed under “Chat Completion v2” for M2.
OpenAI	GPT-5	$1.25	$10.00	Flagship model pricing on OpenAI’s API pricing page.
OpenAI	GPT-5 mini	$0.25	$2.00	Cheaper tier for well-defined tasks.
Anthropic	Claude Sonnet 4.5	$3.00	$15.00	Anthropic’s current per-MTok list; long-context (>200K input) uses a premium tier.
Google	Gemini 2.5 Flash (Preview)	$0.30	$2.50	Prices include “thinking tokens”; page also lists cheaper Flash-Lite and 2.0 tiers.
xAI	Grok-4 Fast (reasoning)	$0.20	$0.50	“Fast” tier; xAI also lists Grok-4 at $3 / $15.
DeepSeek	DeepSeek-V3.2 (chat)	$0.28	$0.42	Cache-hit input is $0.028; table shows per-model details.
Qwen (Alibaba)	qwen-flash (Model Studio)	from $0.022	from $0.216	Tiered by input size (≤128K, ≤256K, ≤1M tokens); listed “Input price / Output price per 1M”.
Cohere	Command R+ (Aug 2024)	$2.50	$10.00	First-party pricing page also lists Command R ($0.50 / $1.50) and others.

Notes & caveats (for readers):

• Prices are USD per million tokens and can change; check linked pages for updates and region/endpoint nuances (e.g., Anthropic long-context >200K input, Google Live API variants, cache discounts).

• Vendors may bill extra for server-side tools (web search, code execution) or offer batch/context-cache discounts.

While the model produces longer, more explicit reasoning traces, its sparse activation and optimized compute design help maintain a favorable cost-performance balance—an advantage for teams deploying interactive agents or high-volume automation systems.

Background on MiniMax — an Emerging Chinese Powerhouse

MiniMax has quickly become one of the most closely watched names in China’s fast-rising AI sector.

Backed by Alibaba and Tencent, the company moved from relative obscurity to international recognition within a year—first through breakthroughs in AI video generation, then through a series of open-weight large language models (LLMs) aimed squarely at developers and enterprises.

The company first captured global attention in late 2024 with its AI video generation tool, “video-01,” which demonstrated the ability to create dynamic, cinematic scenes in seconds. VentureBeat described how the model’s launch sparked widespread interest after online creators began sharing lifelike, AI-generated footage—most memorably, a viral clip of a Star Wars lightsaber duel that drew millions of views in under two days.

CEO Yan Junjie emphasized that the system outperformed leading Western tools in generating human movement and expression, an area where video AIs often struggle. The product, later commercialized through MiniMax’s Hailuo platform, showcased the startup’s technical confidence and creative reach, helping to establish China as a serious contender in generative video technology.

By early 2025, MiniMax had turned its attention to long-context language modeling, unveiling the MiniMax-01 series, including MiniMax-Text-01 and MiniMax-VL-01. These open-weight models introduced an unprecedented 4-million-token context window, doubling the reach of Google’s Gemini 1.5 Pro and dwarfing OpenAI’s GPT-4o by more than twentyfold.

The company continued its rapid cadence with the MiniMax-M1 release in June 2025, a model focused on long-context reasoning and reinforcement learning efficiency. M1 extended context capacity to 1 million tokens and introduced a hybrid Mixture-of-Experts design trained using a custom reinforcement-learning algorithm known as CISPO. Remarkably, VentureBeat reported that MiniMax trained M1 at a total cost of about $534,700, roughly one-tenth of DeepSeek’s R1 and far below the multimillion-dollar budgets typical for frontier-scale models.

For enterprises and technical teams, MiniMax’s trajectory signals the arrival of a new generation of cost-efficient, open-weight models designed for real-world deployment. Its open licensing—ranging from Apache 2.0 to MIT—gives businesses freedom to customize, self-host, and fine-tune without vendor lock-in or compliance restrictions.

Features such as structured function calling, long-context retention, and high-efficiency attention architectures directly address the needs of engineering groups managing multi-step reasoning systems and data-intensive pipelines.

As MiniMax continues to expand its lineup, the company has emerged as a key global innovator in open-weight AI, combining ambitious research with pragmatic engineering.

Open-Weight Leadership and Industry Context

The release of MiniMax-M2 reinforces the growing leadership of Chinese AI research groups in open-weight model development.

Following earlier contributions from DeepSeek, Alibaba’s Qwen series, and Moonshot AI, MiniMax’s entry continues the trend toward open, efficient systems designed for real-world use.

Artificial Analysis observed that MiniMax-M2 exemplifies a broader shift in focus toward agentic capability and reinforcement-learning refinement, prioritizing controllable reasoning and real utility over raw model size.

For enterprises, this means access to a state-of-the-art open model that can be audited, fine-tuned, and deployed internally with full transparency.

By pairing strong benchmark performance with open licensing and efficient scaling, MiniMaxAI positions MiniMax-M2 as a practical foundation for intelligent systems that think, act, and assist with traceable logic—making it one of the most enterprise-ready open AI models available today.

Anthropic is making its most aggressive push yet into the trillion-dollar financial services industry, unveiling a suite of tools that embed its Claude AI assistant directly into Microsoft Excel and connect it to real-time market data from some of the world's most influential financial information providers.

The San Francisco-based AI startup announced Monday it is releasing Claude for Excel, allowing financial analysts to interact with the AI system directly within their spreadsheets — the quintessential tool of modern finance. Beyond Excel, select Claude models are also being made available in Microsoft Copilot Studio and Researcher agent, expanding the integration across Microsoft's enterprise AI ecosystem. The integration marks a significant escalation in Anthropic's campaign to position itself as the AI platform of choice for banks, asset managers, and insurance companies, markets where precision and regulatory compliance matter far more than creative flair.

The expansion comes just three months after Anthropic launched its Financial Analysis Solution in July, and it signals the company's determination to capture market share in an industry projected to spend $97 billion on AI by 2027, up from $35 billion in 2023.

More importantly, it positions Anthropic to compete directly with Microsoft — ironically, its partner in this Excel integration — which has its own Copilot AI assistant embedded across its Office suite, and with OpenAI, which counts Microsoft as its largest investor.

Why Excel has become the new battleground for AI in finance

The decision to build directly into Excel is hardly accidental. Excel remains the lingua franca of finance, the digital workspace where analysts spend countless hours constructing financial models, running valuations, and stress-testing assumptions. By embedding Claude into this environment, Anthropic is meeting financial professionals exactly where they work rather than asking them to toggle between applications.

Claude for Excel allows users to work with the AI in a sidebar where it can read, analyze, modify, and create new Excel workbooks while providing full transparency about the actions it takes by tracking and explaining changes and letting users navigate directly to referenced cells.

This transparency feature addresses one of the most persistent anxieties around AI in finance: the "black box" problem. When billions of dollars ride on a financial model's output, analysts need to understand not just the answer but how the AI arrived at it. By showing its work at the cell level, Anthropic is attempting to build the trust necessary for widespread adoption in an industry where careers and fortunes can turn on a misplaced decimal point.

The technical implementation is sophisticated. Claude can discuss how spreadsheets work, modify them while preserving formula dependencies — a notoriously complex task — debug cell formulas, populate templates with new data, or build entirely new spreadsheets from scratch. This isn't merely a chatbot that answers questions about your data; it's a collaborative tool that can actively manipulate the models that drive investment decisions worth trillions of dollars.

How Anthropic is building data moats around its financial AI platform

Perhaps more significant than the Excel integration is Anthropic's expansion of its connector ecosystem, which now links Claude to live market data and proprietary research from financial information giants. The company added six major new data partnerships spanning the entire spectrum of financial information that professional investors rely upon.

Aiera now provides Claude with real-time earnings call transcripts and summaries of investor events like shareholder meetings, presentations, and conferences. The Aiera connector also enables a data feed from Third Bridge, which gives Claude access to a library of insights interviews, company intelligence, and industry analysis from experts and former executives. Chronograph gives private equity investors operational and financial information for portfolio monitoring and conducting due diligence, including performance metrics, valuations, and fund-level data.

Egnyte enables Claude to securely search permitted data for internal data rooms, investment documents, and approved financial models while maintaining governed access controls. LSEG, the London Stock Exchange Group, connects Claude to live market data including fixed income pricing, equities, foreign exchange rates, macroeconomic indicators, and analysts' estimates of other important financial metrics. Moody's provides access to proprietary credit ratings, research, and company data covering ownership, financials, and news on more than 600 million public and private companies, supporting work and research in compliance, credit analysis, and business development. MT Newswires provides Claude with access to the latest global multi-asset class news on financial markets and economies.

These partnerships amount to a land grab for the informational infrastructure that powers modern finance. Previously announced in July, Anthropic had already secured integrations with S&P Capital IQ, Daloopa, Morningstar, FactSet, PitchBook, Snowflake, and Databricks. Together, these connectors give Claude access to virtually every category of financial data an analyst might need: fundamental company data, market prices, credit assessments, private company intelligence, alternative data, and breaking news.

This matters because the quality of AI outputs depends entirely on the quality of inputs. Generic large language models trained on public internet data simply cannot compete with systems that have direct pipelines to Bloomberg-quality financial information. By securing these partnerships, Anthropic is building moats around its financial services offering that competitors will find difficult to replicate.

The strategic calculus here is clear: Anthropic is betting that domain-specific AI systems with privileged access to proprietary data will outcompete general-purpose AI assistants. It's a direct challenge to the "one AI to rule them all" approach favored by some competitors.

Pre-configured workflows target the daily grind of Wall Street analysts

The third pillar of Anthropic's announcement involves six new "Agent Skills" — pre-configured workflows for common financial tasks. These skills are Anthropic's attempt to productize the workflows of entry-level and mid-level financial analysts, professionals who spend their days building models, processing due diligence documents, and writing research reports. Anthropic has designed skills specifically to automate these time-consuming tasks.

The new skills include building discounted cash flow models complete with full free cash flow projections, weighted average cost of capital calculations, scenario toggles, and sensitivity tables. There's comparable company analysis featuring valuation multiples and operating metrics that can be easily refreshed with updated data. Claude can now process data room documents into Excel spreadsheets populated with financial information, customer lists, and contract terms. It can create company teasers and profiles for pitch books and buyer lists, perform earnings analyses that use quarterly transcripts and financials to extract important metrics, guidance changes, and management commentary, and produce initiating coverage reports with industry analysis, company deep dives, and valuation frameworks.

It's worth noting that Anthropic's Sonnet 4.5 model now tops the Finance Agent benchmark from Vals AI at 55.3% accuracy, a metric designed to test AI systems on tasks expected of entry-level financial analysts. A 55% accuracy rate might sound underwhelming, but it is state-of-the-art performance and highlights both the promise and limitations of AI in finance. The technology can clearly handle sophisticated analytical tasks, but it's not yet reliable enough to operate autonomously without human oversight — a reality that may actually reassure both regulators and the analysts whose jobs might otherwise be at risk.

The Agent Skills approach is particularly clever because it packages AI capabilities in terms that financial institutions already understand. Rather than selling generic "AI assistance," Anthropic is offering solutions to specific, well-defined problems: "You need a DCF model? We have a skill for that. You need to analyze earnings calls? We have a skill for that too."

Trillion-dollar clients are already seeing massive productivity gains

Anthropic's financial services strategy appears to be gaining traction with exactly the kind of marquee clients that matter in enterprise sales. The company counts among its clients AIA Labs at Bridgewater, Commonwealth Bank of Australia, American International Group, and Norges Bank Investment Management — Norway's $1.6 trillion sovereign wealth fund, one of the world's largest institutional investors.

NBIM CEO Nicolai Tangen reported achieving approximately 20% productivity gains, equivalent to 213,000 hours, with portfolio managers and risk departments now able to "seamlessly query our Snowflake data warehouse and analyze earnings calls with unprecedented efficiency."

At AIG, CEO Peter Zaffino said the partnership has "compressed the timeline to review business by more than 5x in our early rollouts while simultaneously improving our data accuracy from 75% to over 90%." If these numbers hold across broader deployments, the productivity implications for the financial services industry are staggering.

These aren't pilot programs or proof-of-concept deployments; they're production implementations at institutions managing trillions of dollars in assets and making underwriting decisions that affect millions of customers. Their public endorsements provide the social proof that typically drives enterprise adoption in conservative industries.

Regulatory uncertainty creates both opportunity and risk for AI deployment

Yet Anthropic's financial services ambitions unfold against a backdrop of heightened regulatory scrutiny and shifting enforcement priorities. In 2023, the Consumer Financial Protection Bureau released guidance requiring lenders to "use specific and accurate reasons when taking adverse actions against consumers" involving AI, and issued additional guidance requiring regulated entities to "evaluate their underwriting models for bias" and "evaluate automated collateral-valuation and appraisal processes in ways that minimize bias."

However, according to a Brookings Institution analysis, these measures have since been revoked with work stopped or eliminated at the current downsized CFPB under the current administration, creating regulatory uncertainty. The pendulum has swung from the Biden administration's cautious approach, exemplified by an executive order on safe AI development, toward the Trump administration's "America's AI Action Plan," which seeks to "cement U.S. dominance in artificial intelligence" through deregulation.

This regulatory flux creates both opportunities and risks. Financial institutions eager to deploy AI now face less prescriptive federal oversight, potentially accelerating adoption. But the absence of clear guardrails also exposes them to potential liability if AI systems produce discriminatory outcomes, particularly in lending and underwriting.

The Massachusetts Attorney General recently reached a $2.5 million settlement with student loan company Earnest Operations, alleging that its use of AI models resulted in "disparate impact in approval rates and loan terms, specifically disadvantaging Black and Hispanic applicants." Such cases will likely multiply as AI deployment grows, creating a patchwork of state-level enforcement even as federal oversight recedes.

Anthropic appears acutely aware of these risks. In an interview with Banking Dive, Jonathan Pelosi, Anthropic's global head of industry for financial services, emphasized that Claude requires a "human in the loop." The platform, he said, is not intended for autonomous financial decision-making or to provide stock recommendations that users follow blindly. During client onboarding, Pelosi told the publication, Anthropic focuses on training and understanding model limitations, putting guardrails in place so people treat Claude as a helpful technology rather than a replacement for human judgment.

Competition heats up as every major tech company targets finance AI

Anthropic's financial services push comes as AI competition intensifies across the enterprise. OpenAI, Microsoft, Google, and numerous startups are all vying for position in what may become one of AI's most lucrative verticals. Goldman Sachs introduced a generative AI assistant to its bankers, traders, and asset managers in January, signaling that major banks may build their own capabilities rather than rely exclusively on third-party providers.

The emergence of domain-specific AI models like BloombergGPT — trained specifically on financial data — suggests the market may fragment between generalized AI assistants and specialized tools. Anthropic's strategy appears to stake out a middle ground: general-purpose models, since Claude was not trained exclusively on financial data, enhanced with financial-specific tooling, data access, and workflows.

The company's partnership strategy with implementation consultancies including Deloitte, KPMG, PwC, Slalom, TribeAI, and Turing is equally critical. These firms serve as force multipliers, embedding Anthropic's technology into their own service offerings and providing the change management expertise that financial institutions need to successfully adopt AI at scale.

CFOs worry about AI hallucinations and cascading errors

The broader question is whether AI tools like Claude will genuinely transform financial services productivity or merely shift work around. The PYMNTS Intelligence report "The Agentic Trust Gap" found that chief financial officers remain hesitant about AI agents, with "nagging concern" about hallucinations where "an AI agent can go off script and expose firms to cascading payment errors and other inaccuracies."

"For finance leaders, the message is stark: Harness AI's momentum now, but build the guardrails before the next quarterly call—or risk owning the fallout," the report warned.

A 2025 KPMG report found that 70% of board members have developed responsible use policies for employees, with other popular initiatives including implementing a recognized AI risk and governance framework, developing ethical guidelines and training programs for AI developers, and conducting regular AI use audits.

The financial services industry faces a delicate balancing act: move too slowly and risk competitive disadvantage as rivals achieve productivity gains; move too quickly and risk operational failures, regulatory penalties, or reputational damage. Speaking at the Evident AI Symposium in New York last week, Ian Glasner, HSBC's group head of emerging technology, innovation and ventures, struck an optimistic tone about the sector's readiness for AI adoption. "As an industry, we are very well prepared to manage risk," he said, according to CIO Dive. "Let's not overcomplicate this. We just need to be focused on the business use case and the value associated."

Anthropic's latest moves suggest the company sees financial services as a beachhead market where AI's value proposition is clear, customers have deep pockets, and the technical requirements play to Claude's strengths in reasoning and accuracy. By building Excel integration, securing data partnerships, and pre-packaging common workflows, Anthropic is reducing the friction that typically slows enterprise AI adoption.

The $61.5 billion valuation the company commanded in its March fundraising round — up from roughly $16 billion a year earlier — suggests investors believe this strategy will work. But the real test will come as these tools move from pilot programs to production deployments across thousands of analysts and billions of dollars in transactions.

Financial services may prove to be AI's most demanding proving ground: an industry where mistakes are costly, regulation is stringent, and trust is everything. If Claude can successfully navigate the spreadsheet cells and data feeds of Wall Street without hallucinating a decimal point in the wrong direction, Anthropic will have accomplished something far more valuable than winning another benchmark test. It will have proven that AI can be trusted with the money.

Some enterprises are best served by fine-tuning large models to their needs, but a number of companies plan to build their own models, a project that would require access to GPUs. 

Google Cloud wants to play a bigger role in enterprises’ model-making journey with its new service, Vertex AI Training. The service gives enterprises looking to train their own models access to a managed Slurm environment, data science tooling and any chips capable of large-scale model training. 

With this new service, Google Cloud hopes to turn more enterprises away from other providers and encourage the building of more company-specific AI models. 

While Google Cloud has always offered the ability to customize its Gemini models, the new service allows customers to bring in their own models or customize any open-source model Google Cloud hosts. 

Vertex AI Training positions Google Cloud directly against companies like CoreWeave and Lambda Labs, as well as its cloud competitors AWS and Microsoft Azure.  

Jaime de Guerre, senior director of product management at Gloogle Cloud, told VentureBeat that the company has been hearing from a lot of organizations of varying sizes that they need a way to better optimize compute but in a more reliable environment.

“What we're seeing is that there's an increasing number of companies that are building or customizing large gen AI models to introduce a product offering built around those models, or to help power their business in some way,” de Guerre said. “This includes AI startups, technology companies, sovereign organizations building a model for a particular region or culture or language and some large enterprises that might be building it into internal processes.”

De Guerre noted that while anyone can technically use the service, Google is targeting companies planning large-scale model training rather than simple fine-tuning or LoRA adopters. Vertex AI Services will focus on longer-running training jobs spanning hundreds or even thousands of chips. Pricing will depend on the amount of compute the enterprise will need. 

“Vertex AI Training is not for adding more information to the context or using RAG; this is to train a model where you might start from completely random weights,” he said.

Model customization on the rise

Enterprises are recognizing the value of building customized models beyond just fine-tuning an LLM via retrieval-augmented generation (RAG). Custom models would know more in-depth company information and respond with answers specific to the organization. Companies like Arcee.ai have begun offering their models for customization to clients. Adobe recently announced a new service that allows enterprises to retrain Firefly for their specific needs. Organizations like FICO, which create small language models specific to the finance industry, often buy GPUs to train them at significant cost. 

Google Cloud said Vertex AI Training differentiates itself by giving access to a larger set of chips, services to monitor and manage training and the expertise it learned from training the Gemini models. 

Some early customers of Vertex AI Training include AI Singapore, a consortium of Singaporean research institutes and startups that built the 27-billion-parameter SEA-LION v4, and Salesforce’s AI research team. 

Enterprises often have to choose between taking an already-built LLM and fine-tuning it or building their own model. But creating an LLM from scratch is usually unattainable for smaller companies, or it simply doesn’t make sense for some use cases. However, for organizations where a fully custom or from-scratch model makes sense, the issue is gaining access to the GPUs needed to run training.

Model training can be expensive

Training a model, de Guerre said, can be difficult and expensive, especially when organizations compete with several others for GPU space.

Hyperscalers like AWS and Microsoft — and, yes, Google — have pitched that their massive data centers and racks and racks of high-end chips deliver the most value to enterprises. Not only will they have access to expensive GPUs, but cloud providers often offer full-stack services to help enterprises move to production.

Services like CoreWeave gained prominence for offering on-demand access to Nvidia H100s, giving customers flexibility in compute power when building models or applications. This has also given rise to a business model in which companies with GPUs rent out server space.

De Guerre said Vertex AI Training isn’t just about offering access to train models on bare compute, where the enterprise rents a GPU server; they also have to bring their own training software and manage the timing and failures. 

“This is a managed Slurm environment that will help with all the job scheduling and automatic recovery of jobs failing,” de Guerre said. “So if a training job slows down or stops due to a hardware failure, the training will automatically restart very quickly, based on automatic checkpointing that we do in management of the checkpoints to continue with very little downtime.”

He added that this provides higher throughput and more efficient training for a larger scale of compute clusters. 

Services like Vertex AI Training could make it easier for enterprises to build niche models or completely customize existing models. Still, just because the option exists doesn’t mean it's the right fit for every enterprise. 

A new framework developed by researchers at Google Cloud and DeepMind aims to address one of the key challenges of developing computer use agents (CUAs): Gathering high-quality training examples at scale.

The framework, dubbed Watch & Learn (W&L), addresses the problem of training data generation in a way that doesn’t require human annotation and can automatically extract demonstrations from raw videos.

Their experiments show that data generated W&L can be used to train or fine-tune existing computer use and foundation models to improve their performance on computer-use tasks. But equally important, the same approach can be used to create in-context learning (ICL) examples for computer use agents, enabling companies to create CUAs for bespoke internal tasks without the need for costly training of specialized models.

The data bottleneck of CUA

The web is rich with video tutorials and screencasts that describe complex workflows for using applications. These videos are a gold mine that can provide computer use agents with domain knowledge and instructions for accomplishing different tasks through user interface interactions.

However, before they can be used to train CUA agents, these videos need to be transformed into annotated trajectories (that is, a set of task descriptions, screenshots and actions), a process that is prohibitively expensive and time-consuming when done manually.

Existing approaches to address this data bottleneck rely on annotating these videos through the use of multimodal language models, which usually result in low precision and faulty examples. A different approach uses self-play agents that autonomously explore user interfaces to collect trajectories. However, techniques using this approach usually create simple examples that are not useful in unpredictable real-world situations.

As the researchers note in their paper, “Overall, these approaches either rely on brittle heuristics, are costly as they rely on explorations in real environments or generate low-complexity demonstrations misaligned with human intent.”

Watch & Learn

The Watch & Learn framework tries to address the challenges of creating CUA demonstrations by rethinking the problem formulation.

Instead of directly generating trajectories or depending on complex multi-stage pipelines, the researchers frame the problem as an “inverse dynamics objective”: Given two consecutive observations, predict the intermediate action that produced the transition.

According to the researchers, this formulation is “easier to learn, avoids hand-crafted heuristics and generalizes robustly across applications.”

The W&L framework can be broken down into three key stages: Training an inverse dynamics model (IDM), retrieving raw videos, and training CUA agents.

In the first phase, the researchers used agents to interact with live web pages to create a large corpus of 500,000 state transitions (two consecutive observations and the action that resulted in the transition). They then used this data (along with 132,000 human-annotated transitions from existing open datasets) to train an inverse dynamics model (IDM) that takes in two consecutive observations and predicts the transition action. Their trained IDM, which is a small transformer model, outperformed off-the-shelf foundation models in predicting transition actions.

The researchers then designed a pipeline that retrieves videos from platforms such as YouTube and runs them through IDM to generate high-quality trajectories. The IDM takes in consecutive video frames and determines the actions (scroll, click) that caused the changes in the environment, which are then packaged into annotated trajectories. Using this method, they generated 53,125 trajectories with high-accuracy action labels.

These examples can be used to train effective computer use models for specific tasks. But the researchers also found that trajectories extracted through IDM can serve as in-context learning examples to improve the performance of CUAs on bespoke tasks at inference time. For ICL, they use Gemini 2.5 Flash to add additional reasoning annotations to the observation/action examples in the trajectories, which can then be inserted into the CUA agent’s prompt (usually 3-5 examples) during inference.

“This dual role (training and in-context guidance) enables flexible integration with both open-source models and general-purpose agents,” the researchers write.

W&L in action

To test the usefulness of W&L, the researchers ran a series of experiments with closed and open source models on the OSWorld benchmark, which evaluates agents in real desktop and operating system environments across different tasks, including productivity, programming and design.

For fine-tuning, they used their corpus of 53,000 trajectories to train two open source models: UI-TARS-1.5, a strong, open source vision-language-action model designed specifically for computer use, and Qwen 2.5-VL, an open-weight multimodal LLM. 

For in-context learning tests, they applied W&L examples to general-purpose multimodal models such as Gemini 2.5 Flash, OpenAI o3 and Claude Sonnet 4. 

W&L resulted in improvements on OSWorld in all model categories, including up to 3 points for ICL on general-purpose models and up to 11 points for fine-tuned open-source models.

More importantly, these benefits were achieved without any manual annotation, “demonstrating that web-scale human workflows can serve as a practical and scalable foundation for advancing CUAs towards real-world deployment,” the researchers write.

This could have important implications for real-world applications, enabling enterprises to turn their existing corpora of videos and conference recordings into training data for CUAs. It also makes it easier to generate new training trajectories. All you will need to do is record videos of performing different tasks and have them annotated by an IDM. And with frontier models constantly improving and becoming cheaper, you can expect to get more from your existing data and the field continues to progress.

Amazon is preparing to lay off as many as 30,000 corporate employees in a sweeping workforce reduction intended to reduce expenses and compensate for over-hiring during the pandemic, according to a report from Reuters on Monday.

GeekWire has contacted Amazon for comment.

Layoff notifications will start going out via email on Tuesday, according to Reuters, which cited people familiar with the matter. One employee at Amazon told GeekWire the workforce is on “pins and needles” in anticipation of cuts.

Bloomberg reported that cuts will impact several business units, including logistics, payments, video games, and Amazon Web Services.

Amazon’s corporate workforce numbered around 350,000 in early 2023. It has not provided an updated number since then.

The company’s last significant layoff occurred in 2023 when it cut 27,000 corporate workers in multiple stages. Since then the company has made a series of smaller layoffs across different business units.

Fortune reported this month that Amazon planned to cut up to 15% of its human resources staff as part of a wider layoff.

Amazon has taken a cautious hiring approach with its corporate workforce, following years of huge headcount growth. The company’s corporate headcount tripled between 2017 and 2022, according to The Information.

The reported cuts come as Amazon is investing heavily in artificial intelligence. The company said earlier this year it expects to increase capital expenditures to more than $100 billion in 2025, up from $83 billion in 2024, with a majority going toward building out capacity for AI in AWS.

Amazon CEO Andy Jassy also hinted at potential workforce impact from generative AI earlier this year in a memo to employees that was shared publicly.

“We will need fewer people doing some of the jobs that are being done today, and more people doing other types of jobs,” he wrote. “It’s hard to know exactly where this nets out over time, but in the next few years, we expect that this will reduce our total corporate workforce as we get efficiency gains from using AI extensively across the company.”

Amazon reported 1.54 million total employees as of June 30 — up 3% year-over-year. The majority of the company’s workforce is made up of warehouse workers.

Amazon employs roughly 50,000 corporate and tech workers in buildings across its Seattle headquarters, with another 12,000 in Bellevue.

The company reports its third quarter earnings on Thursday afternoon.

Fellow Seattle-area tech giant Microsoft has laid off more than 15,000 people since May as it too invests in AI and data center capacity. Microsoft has cut more than 3,200 roles in Washington this year.

Last week, The New York Times cited internal Amazon documents and interviews to report that the company plans to automate as much as 75% of its warehouse operations by 2033. According to the report, the robotics team expects automation to “flatten Amazon’s hiring curve over the next 10 years,” allowing it to avoid hiring more than 600,000 workers even as sales continue to grow.

GeekWire reporter Kurt Schlosser contributed to this story.