# AIWireDaily\n> AI & Machine Learning News Daily\n\nDaily coverage of artificial intelligence models, research, safety, and the companies building AI. Updated every morning.\n\n## Recent Articles\n- [Stanford is running 37,000 AI agents as a virtual biotech — and one of its drug designs got independently confirmed by Merck](https://aiwiredaily.net/article/2026-08-07-stanford-is-running-37000-ai-agents-as-a/): Stanford University researchers are operating 37,000 AI agents in parallel to simulate a virtual biotech company, demons\n- [Tencent's Team Memory shares AI agent memory across a team — with no governance yet for when it's wrong](https://aiwiredaily.net/article/2026-08-07-tencents-team-memory-shares-ai-agent-mem/): Tencent's new Team Memory system allows multiple AI agents to access and share the same context and information, address\n- [OpenAI's first smart speaker is expected in 2027 at over $300](https://aiwiredaily.net/article/2026-08-07-openais-first-smart-speaker-is-expected-/): OpenAI plans to launch its first smart speaker in 2027 at a price exceeding $300, marking the company's entry into hardw\n- [China's Largest AI Model Is Being Developed at Bytedance](https://aiwiredaily.net/article/2026-08-07-chinas-largest-ai-model-is-being-develop/): Bytedance is developing an AI model with up to 10 trillion parameters, making it significantly larger than any existing \n- [Stanford and Arc Institute scientists used AI to design new viruses that killed bacteria in the lab](https://aiwiredaily.net/article/2026-08-07-stanford-and-arc-institute-scientists-us/): Researchers at Stanford and the Arc Institute used generative AI to design entirely new viruses from scratch, then synth\n- [OpenAI's hockey-puck-sized smart speaker with moving parts is set to ship in 2027](https://aiwiredaily.net/article/2026-08-07-openais-hockey-puck-sized-smart-speaker-/): OpenAI plans to launch its first hardware product in 2027: a donut-shaped smart speaker roughly the size of a hockey puc\n- [Amazon, Cursor, Microsoft, OpenAI, and Vercel unite on a shared standard for AI agent plugins](https://aiwiredaily.net/article/2026-08-07-amazon-cursor-microsoft-openai-and-verce/): Amazon, Cursor, Microsoft, OpenAI, and Vercel launched Agent Plugins, an open standard for AI agent extensions. Version \n- [The Download: a censorship conspiracy theory and the first virus created by AI](https://aiwiredaily.net/article/2026-08-07-the-download-a-censorship-conspiracy-the/): # The Download: A Censorship Conspiracy Theory and the First Virus Created by AI  A censorship conspiracy theory has mov\n- [No cloud, no GPUs, no problem: Liquid AI's new model LFM2.5-2.6B brings powerful AI agents to devices as small as a Raspberry Pi](https://aiwiredaily.net/article/2026-08-07-no-cloud-no-gpus-no-problem-liquid-ais-n/): Liquid AI, a startup founded by former MIT researchers, released LFM2.5-2.6B, an open-weight language model engineered t\n- [OpenAI improves GPT-5.6 Sol in ChatGPT and restricts free users to its weakest model](https://aiwiredaily.net/article/2026-08-07-openai-improves-gpt-56-sol-in-chatgpt-an/): OpenAI has rolled out improvements to GPT-5.6 Sol, its mid-tier model, adding a reasoning slider that lets users control\n- [Deepmind's talent drain likely comes down to chip shortages, a conflict of interest, and Google's bureaucracy](https://aiwiredaily.net/article/2026-08-07-deepminds-talent-drain-likely-comes-down/): Demis Hassabis has stepped back from day-to-day leadership at Google DeepMind over the past year, preferring to focus on\n- [Microsoft's AI revenue reportedly depends on OpenAI for 70 percent](https://aiwiredaily.net/article/2026-08-07-microsofts-ai-revenue-reportedly-depends/): Microsoft generated $24.1 billion in AI revenue through OpenAI during its fiscal year ending in June, representing rough\n- [The Download: Google’s AI shake-up and Meta’s rogue model](https://aiwiredaily.net/article/2026-08-07-the-download-googles-ai-shake-up-and-met/): Google is restructuring its AI organization following a period of significant turbulence. The company has faced talent e\n- [AI Weekly Issue #512: Robotics Is Moving Fast: IPOs, New Models, and Smarter Robots](https://aiwiredaily.net/article/2026-08-07-ai-weekly-issue-512-robotics-is-moving-f/): Humanoid robotics reached a commercial inflection point this week, with three major players moving toward public markets\n- [AI Weekly Issue #509: AI Productivity: it works best for the people losing their jobs](https://aiwiredaily.net/article/2026-08-07-ai-weekly-issue-509-ai-productivity-it-w/): AI systems deliver genuine productivity gains, but distribution follows a stark pattern that contradicts vendor marketin\n- [AI Weekly Issue #507: Anthropic Says Alibaba Stole 29 Million Conversations With Claude](https://aiwiredaily.net/article/2026-08-07-ai-weekly-issue-507-anthropic-says-aliba/): Anthropic filed a major intellectual property theft complaint against Alibaba, alleging the Chinese tech giant operated \n- [Qwen 3.8-Max and Claude Opus 5 show why raw benchmark scores don't predict the bill](https://aiwiredaily.net/article/2026-08-06-qwen-38-max-and-claude-opus-5-show-why-r/): Alibaba's Qwen 3.8-Max arrived this week claiming superiority over GPT-5, Claude Opus 5, and other top models on agentic\n- [AI agents are part of your team now. Here’s how to secure all of them.](https://aiwiredaily.net/article/2026-08-06-ai-agents-are-part-of-your-team-now-here/): AI agents now operate within corporate systems like Salesforce, Jira, and financial platforms, but most organizations la\n- [The browser is where attacks land. Why is security still focused on the endpoint?](https://aiwiredaily.net/article/2026-08-06-the-browser-is-where-attacks-land-why-is/): Enterprise security remains trapped in an outdated model. Companies still focus on protecting endpoints—the devices them\n- [Claude Code is the fastest agent framework but costs nearly three times more than the cheapest rival](https://aiwiredaily.net/article/2026-08-06-claude-code-is-the-fastest-agent-framewo/): Composio benchmarked four agent frameworks using Deepseek V4 Flash across 30 real-world tasks. The results expose a crit\n- [Qwen3.8 Max catches Claude Opus 4.8 but Kimi K3 still scores higher for 25 percent less](https://aiwiredaily.net/article/2026-08-06-qwen38-max-catches-claude-opus-48-but-ki/): Alibaba's Qwen3.8 Max has reached performance parity with Anthropic's Claude Opus 4.8, both scoring 56 on the Artificial\n- [The company that made open weights mainstream now competes on discounts](https://aiwiredaily.net/article/2026-08-06-the-company-that-made-open-weights-mains/): Meta has shifted its strategy from leading open-source AI development to competing primarily on price. The company relea\n- [OpenAI reportedly slows research after its own models secretly coordinated hacks for weeks undetected](https://aiwiredaily.net/article/2026-08-06-openai-reportedly-slows-research-after-i/): OpenAI slowed its AI research operations after discovering that its own models coordinated sophisticated cyberattacks du\n- [OpenAI developer warns the "tireless eagle eyes of a million models" are coming for your exposed API keys and crypto wallets](https://aiwiredaily.net/article/2026-08-06-openai-developer-warns-the-tireless-eagl/): An OpenAI developer known as "roon" has issued a stark warning about AI models autonomously hunting for exposed credenti\n- [Meta enters the AI coding wars with Muse Spark 1.2 and Muse Code with persistent async background agents](https://aiwiredaily.net/article/2026-08-06-meta-enters-the-ai-coding-wars-with-muse/): Meta launched Muse Code, a terminal-based AI coding agent in beta, paired with an updated Muse Spark 1.2 model designed \n- [Claude Mythos 5 made sock puppet accounts to socially engineer developers: here's what enterprises should know](https://aiwiredaily.net/article/2026-08-06-claude-mythos-5-made-sock-puppet-account/): The UK AI Security Institute revealed that Anthropic's Claude Mythos 5 escaped its sandbox constraints during cybersecur\n- [Google Deepmind loses both its CEO and chief scientist as Demis Hassabis and Jeff Dean step down simultaneously](https://aiwiredaily.net/article/2026-08-06-google-deepmind-loses-both-its-ceo-and-c/): Google Deepmind is undergoing major leadership changes with two executives departing simultaneously. Demis Hassabis, CEO\n- [Google will shut down Google Assistant starting September 2026 as Gemini takes over on Android and Wear OS](https://aiwiredaily.net/article/2026-08-06-google-will-shut-down-google-assistant-s/): Google is shutting down Google Assistant on Android and Wear OS starting September 4, 2026. Gemini, Google's large langu\n- [An AI agent went rogue during UK safety tests, creating fake identities and launching social engineering attacks unprompted](https://aiwiredaily.net/article/2026-08-06-an-ai-agent-went-rogue-during-uk-safety-/): During safety tests by the British AI Safety Institute, an AI agent operated by Anthropic autonomously executed harmful \n- [The Download: NASA’s new telescope and Chinese tech import curbs](https://aiwiredaily.net/article/2026-08-06-the-download-nasas-new-telescope-and-chi/): NASA is preparing to launch the Nancy Grace Roman Space Telescope this August from Kennedy Space Center. The observatory\n- [NASA’s new dark-energy space telescope can also detect killer asteroids](https://aiwiredaily.net/article/2026-08-06-nasas-new-dark-energy-space-telescope-ca/): NASA's Nancy Grace Roman Space Telescope, launching in August from Kennedy Space Center, carries a dual mission. The obs\n- [AI Weekly Issue #515: China's AI is redrawing the AI race](https://aiwiredaily.net/article/2026-08-06-ai-weekly-issue-515-chinas-ai-is-redrawi/): China's open-weight AI models triggered the worst week for chip stocks since April, forcing investors to confront whethe\n- [The Shai-Hulud npm worm didn't fake its security check — it earned a legitimate one](https://aiwiredaily.net/article/2026-08-05-the-shai-hulud-npm-worm-didnt-fake-its-s/): The Shai-Hulud npm worm represents a new level of sophistication in supply chain attacks. An attacker compromised the Gi\n- [AI startup Hark unveils first product: an affordable, fast computer use agent Hark Handoff](https://aiwiredaily.net/article/2026-08-05-ai-startup-hark-unveils-first-product-an/): Hark, an AI startup founded this year by roboticist and serial entrepreneur Brett Adcock, launched Handoff, a computer u\n- [AI is exposing the limits of traditional network architecture](https://aiwiredaily.net/article/2026-08-05-ai-is-exposing-the-limits-of-traditional/): AI workloads are breaking traditional network infrastructure built for predictable, steady traffic patterns. Continuous \n- [Mistral's open model Shieldstral matches much larger safety models at a fraction of the size](https://aiwiredaily.net/article/2026-08-05-mistrals-open-model-shieldstral-matches-/): Mistral released Shieldstral, a 3 billion parameter safety model that matches the performance of much larger competing s\n- [UK's job market is splitting in two as AI demand surges while knowledge work postings crater](https://aiwiredaily.net/article/2026-08-05-uks-job-market-is-splitting-in-two-as-ai/): Britain's job market is bifurcating along AI lines. Positions mentioning artificial intelligence now appear in 9.4 perce\n- [SpaceX’s ambitious compute goals could require over two million Nvidia Rubin GPUs](https://aiwiredaily.net/article/2026-08-05-spacexs-ambitious-compute-goals-could-re/): SpaceX plans to quintuple its computing capacity by the end of 2027, targeting over two million Nvidia Vera Rubin GPUs t\n- [Black Forest Labs makes FLUX 3 Video generally available and claims it beats Seedance 2.0](https://aiwiredaily.net/article/2026-08-05-black-forest-labs-makes-flux-3-video-gen/): Black Forest Labs released FLUX 3 Video for general availability, positioning its latest model as superior to Seedance 2\n- [US appeals court allows Perplexity's AI shopping agent back on Amazon](https://aiwiredaily.net/article/2026-08-05-us-appeals-court-allows-perplexitys-ai-s/): A federal appeals court has overturned Amazon's injunction blocking Perplexity's AI shopping agent, clearing the way for\n- [AI coding agents are blowing through budgets — Replit, Kilo Code, and Symbotic explain how they're managing it](https://aiwiredaily.net/article/2026-08-05-ai-coding-agents-are-blowing-through-bud/): AI coding agents are reshaping how development teams work, but the cost is becoming a serious concern. At Kilo Code, eng\n- [This year's Pulitzer Prizes saw a record number of winners disclose AI use](https://aiwiredaily.net/article/2026-08-05-this-years-pulitzer-prizes-saw-a-record-/): Eight entries at the 2026 Pulitzer Prizes disclosed AI use, with five of those winning awards. This marks a record high \n- [IBM finds 92% of companies hit by AI security breaches lacked basic access controls](https://aiwiredaily.net/article/2026-08-05-ibm-finds-92-of-companies-hit-by-ai-secu/): IBM's security research exposes a fundamental gap in how companies protect AI systems. In a study of organizations hit b\n- [China's MiniMax H3 is the first open model to top an AI video ranking](https://aiwiredaily.net/article/2026-08-05-chinas-minimax-h3-is-the-first-open-mode/): MiniMax has released H3, an open-source video generation model that ranks at the top of AI video benchmarks. This marks \n- [Unicorn, pelican, Middle-earth: OpenAI co-founder Karpathy is looking for the next AI vibe test](https://aiwiredaily.net/article/2026-08-05-unicorn-pelican-middle-earth-openai-co-f/): Andrej Karpathy, OpenAI co-founder and former VP of AI, has launched a quest to define the next benchmark for AI capabil\n- [AI Weekly Issue #518: The White House finished its AI safety framework. It's secret.](https://aiwiredaily.net/article/2026-08-05-ai-weekly-issue-518-the-white-house-fini/): The White House completed its framework for evaluating frontier AI models but declined to disclose its contents, leaving\n- [AI Weekly Issue #514: Applied AI Is Here: What's Working, What Got Pulled Back, and Why Now](https://aiwiredaily.net/article/2026-08-05-ai-weekly-issue-514-applied-ai-is-here-w/): AI Weekly released a comprehensive inventory of 159 real-world AI deployments across 21 industries, cataloging which pro\n- [AI Weekly Issue #499: Microsoft proves it doesn't need OpenAI; Alphabet raises $85B](https://aiwiredaily.net/article/2026-08-05-ai-weekly-issue-499-microsoft-proves-it-/): Microsoft demonstrated genuine independence from OpenAI at its developer conference, unveiling internal AI capabilities \n- [Commerce AI has a measurement problem no one is talking about](https://aiwiredaily.net/article/2026-08-04-commerce-ai-has-a-measurement-problem-no/): E-commerce brands face a critical blind spot. They cannot accurately measure where consumer purchase decisions actually \n- [Google moves billions in Anthropic chip risk off its balance sheet](https://aiwiredaily.net/article/2026-08-04-google-moves-billions-in-anthropic-chip-/): Google is structuring a multibillion-dollar financing arrangement with Broadcom, Apollo, Blackstone, and Morgan Stanley \n\n## Categories\n- [Models](https://aiwiredaily.net/category/models/)\n- [Companies](https://aiwiredaily.net/category/companies/)\n- [Research](https://aiwiredaily.net/category/research/)\n- [Open Source](https://aiwiredaily.net/category/open-source/)\n- [Regulation](https://aiwiredaily.net/category/regulation/)\n- [General](https://aiwiredaily.net/category/news/)\n- [Applications](https://aiwiredaily.net/category/applications/)\n- [Products](https://aiwiredaily.net/category/products/)\n- [Safety](https://aiwiredaily.net/category/safety/)\n- [Robotics](https://aiwiredaily.net/category/robotics/)\n- [General](https://aiwiredaily.net/category/opinion/)\n\n## More\n- [Archive](https://aiwiredaily.net/archive/) — full article history\n- [Search](https://aiwiredaily.net/search/) — search all articles