Liverpool'un golü VAR'dan döndü




























OpenAI has announced its plans to acquire Promptfoo, an established AI security platform widely used by enterprises to identify and remediate vulnerabilities in AI systems during development. The company confirmed that once the acquisition is finalized, Promptfoo’s technology will be integrated directly into OpenAI Frontier, the platform designed for building and operating AI coworkers. The move reflects OpenAI’s growing focus on strengthening evaluation, security, and compliance capabilities as enterprises increasingly deploy AI agents into real‑world workflows.
According to OpenAI, organizations adopting AI coworkers require systematic methods to test agent behavior, detect risks before deployment, and maintain transparent records to support oversight and governance. Promptfoo, led by co‑founders Ian Webster and Michael D’Angelo, has built a suite of tools trusted by more than a quarter of Fortune 500 companies. Its open‑source CLI and library for evaluating and red‑teaming large language model applications have become widely used across the industry. OpenAI stated that it will continue supporting the open‑source project while expanding enterprise‑grade capabilities within Frontier.
Srinivas Narayanan, CTO of B2B Applications at OpenAI, said the acquisition brings deep engineering expertise in evaluating and securing AI systems at scale. He noted that Promptfoo’s work enables businesses to deploy secure and reliable AI applications, and integrating these capabilities into Frontier will strengthen the platform’s native security features. OpenAI highlighted that the integration will introduce automated security testing and red‑teaming directly into Frontier, enabling enterprises to identify risks such as prompt injections, jailbreaks, data leaks, tool misuse, and out‑of‑policy agent behaviors.
The company also emphasized that security and evaluation will be embedded into development workflows, allowing teams to identify, investigate, and remediate risks earlier in the lifecycle. Enhanced reporting and traceability will support governance, risk management, and compliance requirements as AI oversight expectations continue to rise globally.
Promptfoo CEO Ian Webster said the company was founded to give developers practical tools to secure AI systems, noting that the increasing connectivity of AI agents to real data and systems makes validation more critical than ever. He added that joining OpenAI will accelerate efforts to deliver stronger security, safety, and governance capabilities for teams building real‑world AI applications. The acquisition remains subject to customary closing conditions.
The post OpenAI to acquire AI security startup Promptfoo appeared first on My Startup World - Everything About the World of Startups!.
Luma announced the launch of Luma Agents, a new class of AI collaborators capable of executing end-to-end creative work across text, image, video, and audio. Designed for agencies, marketing teams, studios, and enterprise organizations that aspire to scale creative output without sacrificing quality, Luma Agents maintain full context from the initial brief to final delivery – coordinating tools, models, and iterations within a single unified system.
“Creative work has never lacked ambition; it’s lacked execution capacity,” said Amit Jain, Co-Founder and CEO of Luma. “Creative teams shouldn’t have to spend their time orchestrating tools. They should spend it creating. Agents aren’t shortcuts. They’re collaborators that maintain context, coordinate execution, and advance projects so teams can focus on taste, direction, and strategy.”
For the past several years, most AI systems have been assembled by chaining together separate models for language, vision, video, and reasoning — stitching outputs together through orchestration layers. While powerful in isolation, these systems fragment context and require increasingly complex workflows to produce reliable creative results.
Luma believes intelligence should not be assembled in pieces; it should be built as one coherent system.
Creative Agents That Make You Prolific
Luma Agents replace fragmented, multi-model workflows with coordinated, execution built on unified reasoning. Instead of switching between disconnected tools and rebuilding context at every step, teams work alongside Agents that:
Agents operate inside a collaborative, multiplayer environment where humans direct creative intent and Agents handle orchestration, routing, and execution – resulting in more output, greater consistency, and higher creative velocity.
Deployed at Global Scale
Luma Agents are already embedded across global agency operations.
Publicis Groupe and Serviceplan Group are deploying Luma Agents across strategy, creative development, and production workflows to increase throughput while maintaining brand consistency across markets.
“Luma is now part of our broader House of AI ecosystem and integrated directly into our creative workflows. It allows our teams across more than 20 countries to collaborate more smoothly and develop great work faster. For our clients, that means high-quality creative output delivered with greater speed and efficiency – without compromising craft,” says Alexander Schill, Global CCO at Serviceplan Group.
Built on Unified Intelligence
Luma Agents are built on Unified Intelligence, a new model architecture designed to move beyond the industry’s prevailing approach of assembling intelligence in pieces. Instead of chaining together separate models for language, vision, and generation, Unified Intelligence trains a single multimodal reasoning system capable of understanding and generating across formats within the same architecture.
For the past several years, most AI systems have been assembled as pipelines: one model writes text, another generates images, another processes video, and orchestration layers attempt to stitch their outputs together. While effective for narrow tasks, these systems fragment reasoning, lose context between steps, and require complex workflows to produce reliable results.
Unified Intelligence takes a different approach. Instead of connecting specialized models after the fact, it trains a single multimodal reasoning system capable of understanding and generating across formats within the same architecture.
Rather than separating thinking from creation, Unified Intelligence tightly couples reasoning and rendering, allowing the system to plan, imagine, and produce as part of one coherent cognitive process.
When a human architect sketches a building, they are not simply drawing lines – they are simultaneously simulating structure, light, spatial dynamics, and lived experience. Reasoning and imagination happen together. Unified Intelligence is built on the same principle.
The first model built on this architecture is Uni-1.
Uni-1 is a decoder-only autoregressive transformer operating over a shared token space that interleaves language and image tokens, allowing both modalities to function as first-class inputs and outputs in the same sequence. This design enables the model to reason in language while imagining and rendering in pixels within the same forward pass.
Rather than generating outputs step-by-step across disconnected systems, Uni-1 can plan, visualize, and produce creative artifacts as part of a single coherent reasoning process. The result is a foundation where thinking and creation are tightly coupled, much closer to how human intelligence works.
Built on top of this unified architecture, Luma Agents can coordinate complex creative workflows that previously required multiple tools and manual orchestration. They can:
Together, these capabilities allow Luma Agents to function not as isolated generation tools, but as collaborative AI creatives capable of executing end-to-end creative work.
“Intelligence shouldn’t be fragmented by modality,” added Jain. “Unified systems reason holistically. When the same model can think, imagine, and render, you move closer to intelligence that behaves coherently across the entire creative process.”
Enterprise-Ready by Design
Luma Agents are designed for enterprise environments where intellectual property protection, compliance, and operational scale are critical. Key enterprise safeguards include Full IP ownership retained by customers, automated content review to reduce copyright risk, legal trace documentation demonstrating human involvement, required human review workflows prior to public release, and cloud-based infrastructure with enterprise-grade guardrails.
The post Luma launches Luma Agents for creative works appeared first on My Startup World - Everything About the World of Startups!.
Reclaim Security, a preemptive exposure-remediation platform, announced $26 million in total funding, including a recent $20 million Series A round led by Acrew Capital, with participation from QP Ventures and Ibex Investors. The funding will accelerate the company’s mission to eliminate what many security leaders consider cybersecurity’s most persistent gap: remediation.
As attacker breakout times have fallen to as little as 27 seconds, enterprises still require an average of 27 days to remediate critical exposures. Over the past decade, organizations have invested heavily in detection tools to identify vulnerabilities and misconfigurations, yet resolving them remains largely manual, slow, and operationally risky. The result is an expanding backlog of exposures that security teams identify but struggle to safely close.
“There is a massive ‘Remediation Mirage’ in the market right now. Vendors are slapping an AI label on what is essentially just Prioritization 2.0 or faster ticket management,” says Barak Klinghofer, CEO and Co-founder of Reclaim Security.
”The recent launch of Claude Code, which wiped billions from the market value of traditional security giants, is a massive wake-up call. While such tools can identify hundreds of vulnerabilities in seconds, they also hand attackers an autonomous, high-speed engine for exploit generation. We’ve seen reports of AI-orchestrated espionage campaigns where 80-90% of tactical operations were executed autonomously. In this new reality, if your ‘remediation’ strategy still ends with a human reviewing a manual Jira ticket, you aren’t just slow, you’ve lost the race.
Reclaim is the only platform providing true Agentic Remediation. Through our PIPE engine, we’ve removed the fear of ‘breaking the business,’ allowing our AI to move from discovery to resolution in seconds. While others are perfecting the recommendation, we are perfecting the execution.”
Automating Cybersecurity’s “Last Mile”
Reclaim’s platform introduces the industry’s first AI Security Engineer, an autonomous system designed not only to identify exposures, but to resolve them safely and at scale.
At the core of the platform is PIPE (Productivity Impact Prediction Engine), a simulation engine that predicts the operational and business impact of a proposed security change before it is deployed. By accurately modeling how changes impact applications, workloads, user productivity and business processes, organizations can implement remediation without risking downtime or operational disruption.
This simulation-first approach enables organizations to:
Reclaim analyzes how real attack techniques would traverse a specific environment, evaluates how existing defenses would respond, and predicts the operational impact of remediation before changes are deployed. By combining advanced attack path modeling with business-aware remediation, the company eliminates exploitable pathways safely and at scale. This approach enables a shift away from reactive “assume breach” strategies toward proactively removing exposure without disrupting critical business operations.
Real World Impact
Early enterprise customers across financial services, healthcare, government, and critical infrastructure sectors report measurable results, including 80% increase in overall threat resilience, 75% increase in ROI from existing security stack and 90% reduction in manual effort when resolving critical exposures
“Security tools are excellent at explaining why something is risky,” said Mark Kraynak, Founding Partner at Acrew Capital. “What they don’t do is make remediation safe and practical. The real breakthrough isn’t more prioritization, it’s removing risk without breaking the business. Reclaim does exactly that, and that’s why it matters.”
The post Reclaim Security raises $26M led by Acrew Capital appeared first on My Startup World - Everything About the World of Startups!.































































































































































































































































































































































































































































