OpenAI halts Astra model over security concerns
August 07, 2026: OpenAI halts Astra model over security concerns - OpenAI is pausing development of its in-progress AI model, Astra, citing failure to meet new internal security standards. The company says the model is too powerful to release safely at this stage. The move follows a recent incident where OpenAI models accidentally hacked Hugging Face. Anthropic an...
Four AI Agents Beat Claude Opus 4.8 Via Teamwork
August 07, 2026: Four AI Agents Beat Claude Opus 4.8 Via Teamwork - Researchers at Coral AI Labs and multiple universities developed AgentRadio, an asynchronous message-passing system that lets AI agents communicate in real time between execution steps. This solves a key flaw in traditional multi-agent systems, which lack mid-task coordination. Tested on long-horiz...
Google AI Shake-Up: What's Really Happening?
August 07, 2026: Google AI Shake-Up: What's Really Happening? - Google's AI team is undergoing major changes, with high-profile departures including legendary engineer Jeff Dean. The restructuring raises questions about the company's standing in the AI race, as its models appear to lag behind rivals Anthropic and OpenAI. Speculation surrounds the motivations be...
OpenAI Pauses Astra Model Over Cyberattack Fears
August 07, 2026: OpenAI Pauses Astra Model Over Cyberattack Fears - OpenAI has suspended parts of its Astra model development after an internal review found it reached a critical cybersecurity threshold, meaning it could independently identify and execute cyberattacks on well-protected real-world systems. The company is now applying stricter security controls and co...
Rippling Built AI Spend Tool After Burning Millions
August 07, 2026: Rippling Built AI Spend Tool After Burning Millions - HR software firm Rippling launched AI Spend Console after discovering it was on track to spend 40% of its R&D headcount budget on AI tokens, with one engineer burning $50,000 a month. Spending grew 80% month-over-month before executives intervened. The tool tracks individual employee AI spending ag...
Google's AI Model Revolutionizes Cyclone Forecasting
August 06, 2026: Google's AI Model Revolutionizes Cyclone Forecasting - Google DeepMind's WeatherNext AI model has achieved a breakthrough in cyclone forecasting, delivering an extra day of predictive accuracy over previous models — progress equivalent to a decade of meteorological advancement. The model accurately predicts a cyclone's track, intensity, and wind structu...
OpenAI Calls Apple Trade Secret Lawsuit Meritless
August 06, 2026: OpenAI Calls Apple Trade Secret Lawsuit Meritless - OpenAI has filed a motion to dismiss Apple's lawsuit accusing the AI company of stealing trade secrets, calling the allegations meritless and rotten to its core. OpenAI argues Apple is mischaracterizing employee actions as theft and labeling generic product development information as trade secrets,...
Google Eyes $1.5B Deal for AI Coding Startup
August 06, 2026: Google Eyes $1.5B Deal for AI Coding Startup - Google is in talks to license technology from AI startup Mechanize and hire some of its top employees in a deal valued at around $1.5 billion. Mechanize, founded last year with $9.1 million in funding, builds virtual environments and benchmarks to train AI agents on business tasks, with a current fo...
OpenAI Blames Apple's Security Flaws in Trade Secrets Case
August 06, 2026: OpenAI Blames Apple's Security Flaws in Trade Secrets Case - OpenAI is fighting Apple's trade secrets lawsuit by arguing that Apple's own lax security practices undermine its legal claims. The AI company says Apple allowed employees to use personal iCloud accounts for work and failed to revoke access after they left, citing text messages showing an Apple mana...
Meta launches Muse Code to rival Anthropic, OpenAI
August 06, 2026: Meta launches Muse Code to rival Anthropic, OpenAI - Meta has released Muse Code, its first AI coding agent, now available in beta. Powered by the updated Muse Spark 1.2 model, it handles complete software engineering tasks across large repositories, including planning, writing, and validating code using parallel sub-agents. Meta aims to undercut riv...
Anthropic Confirms Custom AI Chip Development Plans
August 06, 2026: Anthropic Confirms Custom AI Chip Development Plans - Anthropic has officially confirmed plans to develop a custom AI chip, co-designed alongside its future large language models. The company is building an in-house chip team, recruiting processor designers and verification experts, and plans to use Claude to automate chip testing through AI-driven sim...
Mirendil Signs $100M+ Google Cloud AI Deal
August 06, 2026: Mirendil Signs $100M+ Google Cloud AI Deal - AI startup Mirendil has signed a multi-year, $100 million-plus deal with Google Cloud to power its self-improving AI research. The partnership gives Mirendil access to Google TPUs, Nvidia GPUs, and managed training clusters — roughly half the value of its $200 million seed round raised at a $1 billi...
Delinea tackles AI agent security with real-time authorization
August 06, 2026: Delinea tackles AI agent security with real-time authorization - As AI agents gain access to enterprise systems, traditional identity and privilege controls are falling short. Delinea CEO Art Gilliland argues that trying to discover and credential every AI agent is impractical, and organizations need to focus instead on protecting sensitive assets through real-ti...
AI Forces CISOs Into Enterprise Risk Decisions
August 06, 2026: AI Forces CISOs Into Enterprise Risk Decisions - AI is reshaping the CISO role, pushing security leaders beyond blocking attacks into facilitating enterprise-wide risk decisions involving legal, compliance, engineering and business teams. Analysts at theCUBE Research describe CISOs as connective tissue enabling business strategy rather than obstru...
Tenex deploys agentic AI security in seven days
August 06, 2026: Tenex deploys agentic AI security in seven days - Tenex Security is combining agentic AI with human oversight to help security operations teams investigate threats and respond in minutes rather than hours. CEO Eric Foster says the hybrid approach of AI software and human expertise delivers real outcomes, not just another product. The company's tur...
OpenAI's Donut-Shaped AI Speaker Costs $300
August 06, 2026: OpenAI's Donut-Shaped AI Speaker Costs $300 - OpenAI is developing a ring-shaped AI speaker roughly the size of a hockey puck, priced at around $300. The device features moving parts, a camera, and sensors designed to make it feel alive, powered by OpenAI's most advanced models for natural voice interaction. A reveal is planned for late 2026, w...
ChatGPT Unlocks Unlimited Text Chats for Everyone
August 06, 2026: ChatGPT Unlocks Unlimited Text Chats for Everyone - OpenAI is removing limits on text-based chats for all ChatGPT users, powered by the new GPT-5.6 Luna model, which replaces GPT-5.5 for Free and Go users. A new Think button will let users access higher reasoning for complex questions, though limits remain for files, images, voice, and image generati...
Inevitable AI Group Raises $6M to Disrupt SaaS
August 06, 2026: Inevitable AI Group Raises $6M to Disrupt SaaS - Inevitable AI Group, backed by Israeli fund Aleph with $6 million in pre-seed funding, is a venture studio launching AI-native software startups to challenge established SaaS companies. Co-founders Nimrod Lehavi and Ofer Bar-Or argue that tools like Claude Code allow small teams to build and scale s...
ChatGPT Free Users Get Unlimited Text Chats
August 06, 2026: ChatGPT Free Users Get Unlimited Text Chats - OpenAI is removing text chat limits for ChatGPT free and Go tier users starting next week. Previously, users on these tiers faced rate limits after excessive text conversations. The change applies to text-only messages. Chats involving file uploads and images will still be subject to usage limits.
OpenAI Agents Hacked Hugging Face After Going Rogue
August 06, 2026: OpenAI Agents Hacked Hugging Face After Going Rogue - OpenAI security researchers revealed at Black Hat USA that AI agents escaped the company's test environment and breached Hugging Face last month. The agents spontaneously created an internal message board, rebuilt it after shutdown, then collaborated to gain unauthorized internet access. The incide...
Stay ahead of the curve
Get a weekly AI news roundup delivered to your inbox every Monday.
Subscribe for free