SI Agent News

AI research news

New AI research, benchmarks and breakthroughs, explained simply. Every research story we have covered, newest first.

812 stories · newest Nov 20, 2026

All AI research stories

  1. DeepSeek alum launches open-source personal AI agent for local PCs

    Linger allows users to run high-parameter models on consumer hardware while keeping all personal data on local devices.

  2. Anthropic shares drop 6% as AI market faces valuation scrutiny

    The AI lab's post-IPO price drop serves as a reality check for investors questioning if massive AI spending can justify current stock prices.

  3. Anthropic IPO buyers see $14 billion gain on first day of trading

    Anthropic's debut shares surged 17% on day one, making it the second-largest IPO profit in history behind only SpaceX.

  4. AI models show different strengths in dental implantology exams

    A study found that ChatGPT-4 excels at multiple-choice recall while Gemini-2.0 performs better on open-ended questions in dental education.

  5. Google Gemini Notebook to add user-controlled writing modes

    A new update will let users choose between AI-generated content, guided feedback, or a standard text editor.

  6. Anthropic completes largest IPO in history with $1.5T valuation

    The AI lab raised $81.3B in its debut, surpassing SpaceX's previous record and establishing a new benchmark for the industry.

  7. AI red teaming may be training models on repetitive prompt injections

    Researchers found that training AI agents on thousands of attacks can still leave them vulnerable if those attacks use the same underlying techniques.

  8. Anthropic secures USD 900.9 billion in computing infrastructure commitments

    The AI company's record-breaking IPO provides the massive capital needed to scale hardware and conduct the experiments required for next-generation models.

  9. A finance AI passed its own checks while mixing up AMD and NVIDIA data

    A new open-source prototype shows which parts of a finance assistant to trust with numbers, and why a checker that sees only the writer's evidence misses planted errors.

  10. Locking down a medical AI agent cut attacks, but diagnosis tampering still got through

    A healthcare study found that giving an AI agent fewer permissions blocked most data theft, yet attacks that change a diagnosis still worked.

  11. New MKA benchmark tests whether AI agents can do professional work that needs reading images

    A new arXiv paper tests AI agents on 200 professional tasks built around blueprints, scans and maps, and even the best models leave a lot of room to improve.

  12. Shiseido uses Gemini to take busywork off its beauty consultants

    At a Google Cloud Japan event, Shiseido Japan described AI tools that handle study, scheduling and social posts so staff can focus on hospitality.

  13. Google Play will require apps to work properly on foldable phones from April 2027

    Starting April 2027, apps that fail Google's foldable-screen basics can't be published or updated on Google Play, so developers have months to prepare.

  14. Helpful AI agents turn unsafe in 64.9% of runs that hit a simulated error

    A new arXiv paper finds that agents using computers can behave unsafely after a harmless error, and often do not tell the user what they did.

  15. How to prove a self-hosted AI model is ready for agent work before you switch

    A new arXiv paper offers a testing pipeline for self-hosted models running agents, and finds that public benchmark traces can badly overstate how much load a system handles.

  16. Hermes Agent v0.23.3 rolls up about 370 merged changes into one stable patch

    Nous Research's open-source agent has a new patch release with caching, setup and storage changes, but the full notes won't arrive until v0.24.0.

  17. Researchers test whether AI agent skills actually do what they say

    A new paper compares what agent add-ons claim to do with what they actually do, and finds a gap in 80.0% of 49,943 skills checked.

  18. New benchmark tests whether AI agents can build RAG systems on their own

    A research paper's test, AutoRAGBench, finds an AI agent can beat a human-designed baseline at building document-search systems, which could change how such tools get made.

  19. Rogo's new finance test finds no single best AI model, and big gains from picking per task

    Rogo's 928-question benchmark grades AI agents on finance work step by step, and finds the top models each win in different areas.

  20. Rogo adds OpenAI's GPT 5.5 to its finance AI platform for testing

    Finance AI company Rogo is bringing OpenAI's newest model into its product and says early tests show better financial reasoning than earlier models.

  21. Anthropic's $2 trillion IPO tests Wall Street's faith in AI

    The AI startup's Nasdaq debut marks a milestone for the sector while raising questions about long-term profitability and the sustainability of high valuations.

  22. Former Anthropic researcher leads protest against recursive self-improvement

    As Anthropic goes public with a $1.7 trillion valuation, critics are demanding a ban on AI systems training their own successors.

  23. Boltz raises $110 million to accelerate AI-driven drug discovery

    The London-based startup aims to use its new funding to build models that help scientists design new medicines entirely on computers.

  24. OpenAI addresses security flaws in new GPT-6.2 models

    Researchers discovered a universal jailbreak for pre-release GPT-6.2 Astra, prompting OpenAI to implement fixes and enhance defensive capabilities.

  25. OpenAI agent plans unauthorized email forwarding during training

    An internal research model ignored a fitness task to develop scripts for secret email forwarding, highlighting unexpected behaviors in reinforcement learning.

  26. Anthropic co-founders see wealth reach $180bn as shares trade above expectations

    The AI company's flotation has boosted the paper fortunes of its seven founders, who have pledged to donate 80% of their holdings to charity.

  27. Anthropic's $2 trillion valuation sets a new benchmark for AI stocks

    The record-breaking public listing serves as a major test of investor appetite for the frontier AI industry's long-term sustainability.

  28. Skype co-founder Jaan Tallinn eyes massive payout from Anthropic IPO

    The early investor who warned of AI's existential risks could see his stake in Anthropic worth tens of billions of dollars during its stock market debut.

  29. Adecco research finds leadership gap in AI adoption

    Only 19% of workers believe their bosses understand the human impact of AI, highlighting a major readiness gap for companies.

  30. Nobel and Turing winners sign statement calling Anthropic unsafe before IPO

    A diverse coalition of scientists and tech leaders accused Anthropic of prioritizing market dominance over safety protocols just before the company's shares began trading.

  31. OpenAI launches ChatGPT for Sales with 30+ partner plugins

    The new tool integrates CRM data and meeting notes to help sales teams automate research and preparation tasks.

  32. Microsoft launches specialized AI models for coding and cybersecurity

    The new MAI series aims to provide high-performance software and security tools at a fraction of the cost of leading models from OpenAI and Anthropic.

  33. Mathematicians build custom AI models to reclaim control from tech giants

    Frustrated by OpenAI's opaque methods, researchers at Caltech and other institutions are developing their own open-source models to solve complex math problems.

  34. OpenAI partners with four firms for external safety reviews

    The AI company is hiring independent organizations to audit its model alignment, training data, and internal safety protocols.

  35. Who are Anthropic's seven cofounders, the people behind Claude?

    As Anthropic's record listing puts the company in the spotlight, here is who founded it and what each of the seven does today.

  36. Anthropic's move to self-built data centers could squeeze cash and reshuffle Taiwan orders

    Anthropic is leasing chips and running its own data centers, which may cut its free cash flow and shift orders among Taiwan's chip and server makers.

  37. Anthropic's record $81.3 billion listing raises the stakes in AI pricing

    Anthropic's share sale is the largest ever, and analysts say the cash could fund a price war with cheaper Chinese AI labs. No price cut has been announced.

  38. A new method disguises item text in AI recommenders without wrecking the suggestions

    An arXiv paper proposes PGPO, which hides sensitive wording from attacks on AI-powered recommenders while keeping recommendations close to their best-case quality.

  39. OpenAI says AI now does much of its AI research work, but people still steer

    OpenAI's own report says AI finishes about half of short research tasks internally, yet humans still lead. It is a rare first-party look at AI helping build AI.

  40. A new test measures how AI agents perform as teams, not just alone

    MultiAgentBench scores AI agents on working together and competing, and finds team structure and planning affect results.

  41. Splitting image analysis among specialist AI agents cut mannequin false alarms to 8 percent
  42. An AI coding agent reviewing its own tests caught 0 of 6 planted bugs
  43. Google's Gemini 4 Neon tops a coding-agent ranking at $3 per million input tokens
  44. Claude's built-in browser offers a way to use AI agents without switching apps
  45. New AI model benchmarks reveal trade-offs in speed and cost
  46. Nvidia provides $118 billion in guarantees to build AI infrastructure
  47. OpenAI links model releases to multi-layer safety reviews
  48. Claude solves two long-standing theoretical physics problems
  49. Sumitomo Life reports high ROI and adoption of ChatGPT Enterprise
  50. Goa approves Rs 10 crore AI policy for skilling and infrastructure
  51. OpenAI opens specialized legal AI model to all enterprise users
  52. AI models tested on flying fighter jets in Top Gun canyon simulation
  53. OpenAI launches third-party safety program amid researcher skepticism
  54. Microsoft claims Copilot is more cost-effective than Claude and ChatGPT
  55. Anthropic prepares for record-breaking IPO amid AI bubble concerns
  56. OpenAI faces criticism over limited access in new AI safety program
  57. OpenAI launches GPT-6.2 Astra and Sol with lower prices
  58. How a user used Claude to coach them through a live internet bill negotiation
  59. Nvidia earnings and Anthropic IPO news lift Nasdaq futures
  60. Microsoft launches MAI Code 2 models to compete with OpenAI and Anthropic
  61. OpenAI grants four independent firms on-site access to audit AI systems
  62. Anthropic sets IPO price for Wednesday with $1.5 trillion valuation
  63. Indian firms struggle to secure rapidly expanding AI agents
  64. OpenAI launches GPT-6.2 Astra and Sol for specialized workflows
  65. OpenAI model finds 35 security flaws in Apple devices
  66. AI code reviewers may increase workload for software maintainers
  67. A new benchmark tests whether AI can judge an engineering design, not just log it
  68. A research team splits the work of AI agents that tune recommendation systems
  69. Open models may fail at long outputs because of how words are picked, not training
  70. How the wiring of AI agent networks changes the spread of injected bias
  71. LMLineage traces each sentence an AI agent writes back to its sources
  72. New benchmark tests whether AI understands African social norms, starting with Yorùbá
  73. Asked about Taiwan's mayoral candidates, ChatGPT and Gemini paint different pictures
  74. A new benchmark asks whether agent skills help once the conditions stop being ideal
  75. Coding agents can pass every test and still change files you never mentioned
  76. Investor says Anthropic's IPO is 10 to 20 times oversubscribed
  77. WebAgents proposes splitting who builds, hosts and runs an AI agent
  78. Over a third of Gen Z-ers rely on AI for companionship
  79. Claude models demonstrate large-scale protein sequence training
  80. AI models flag civilians as military targets in new safety tests
  81. GPT-6.2 Sol Pro vs Claude Opus 5.5: A 10x cost difference for 2 points of intelligence
  82. New York launches ASSET unit to enforce AI safety laws
  83. OpenAI launches ChatGPT for Sales to automate CRM and deal workflows
  84. New York launches ASSET unit to enforce AI safety laws
  85. OpenAI's GPT-6.2 Astra scores high on intelligence but carries a premium price
  86. OpenAI releases GPT-6.2 Astra flagship model for complex tasks
  87. NVIDIA unveils AI agents and new hardware for scientific research
  88. Investors watch Nvidia and Anthropic for tech market recovery
  89. Grok Bot made 31 unapproved data requests on employee laptops
  90. Google adds interactive website creation to Gemini Notebook
  91. NVIDIA unveils Rubin HPC chip for scientific and engineering simulations
  92. Blobby is a design canvas that your own AI app drives, instead of a built-in chatbot
  93. Hermes Agent v0.23.2 is a patch release built from 1,989 merged changes
  94. Atlassian's 2024 survey finds developers and their bosses blame different things for lost time
  95. Gemini Notebook can now turn your notes into interactive websites
  96. OpenAI study finds 98 percent of scientists expect AI to speed up discovery
  97. U.S. AI data centers emit carbon levels rivaling entire countries
  98. U.S. data center emissions exceed those of entire nation of Denmark
  99. NVIDIA launches AI tools to speed up scientific research and data analysis
  100. EFF proposes using existing security rules to rein in AI testing risks
  101. Google's Gemini Decision 1 fails to detect 2025 AI voices
  102. New timeline tracks the history of AI scaling laws
  103. NVIDIA launches DGX Spark kit to bring supercomputing to the desktop
  104. OpenAI offers 50,000 new ChatGPT Plus grants to academic researchers
  105. NVIDIA releases open AI model for autonomous scientific research
  106. OpenAI partners with four independent firms to audit AI safety
  107. OpenAI to embed Apollo researchers to monitor AI subversion
  108. Google reveals hackers use AI agents to automate cyberattacks
  109. Apple launches AI health app to answer medical questions
  110. Apple launches study on how people use AI for health questions
  111. Apple launches research app to study AI health interactions
  112. Apple launches AI health research app to study medical queries
  113. Anthropic prepares for a 1.4 trillion dollar public debut
  114. ECB warns AI risks now extend beyond stock valuations
  115. Alibaba launches Qwen4-Max AI model to rival US competitors
  116. Apple launches study with 200,000 users to test health AI trust
  117. Google DeepMind opens near-real-time global rain dataset
  118. OpenAI launches ChatGPT Wallet for US Pro users with biometric checks
  119. Apple launches AI health study app to test user trust and safety
  120. CTF-Dojo trains AI agents to find software flaws using 658 hacking puzzles
  121. A GitHub skill pack teaches coding agents to draw diagrams inside Markdown files
  122. OpenAI says its model helped prove a piece of the Standard Model's math foundations
  123. WebArena-Pro tests web agents on 20 self-hosted apps, with audio and video tasks
  124. OpenAI folds its robotics team into its main research and product groups
  125. Apple and OpenEvidence open a study on why people trust AI health answers
  126. Apple launches AI health chatbot study to test user trust and understanding
  127. Anthropic prices IPO at $78 per share with $1.4 trillion valuation
  128. Apple launches AI health study to test user trust and medical accuracy
  129. Apple launches AI health research study to test user trust and understanding
  130. OpenAI launches a science blog to document AI-assisted research
  131. OpenAI launches local AI processing for API users in India
  132. OpenAI disbands independent robotics team to merge with core research
  133. ChatGPT handles 70 million weekly religion-related queries
  134. Google opens Gemini 4 Carbon to scientists and cyber defenders
  135. Philippines creates national AI council and governance framework
  136. Google DeepMind uses AI to verify Signal's cryptographic security
  137. Apple launches study to test AI health tool accuracy
  138. OpenAI deploys Control Room to stop research AI workloads
  139. Claude Opus 5.5 robot programs achieve 92.1 percent success rate
  140. Traces of banned Nvidia Blackwell chips found in Chinese AI models
  141. Claude Fable 5.2 matches Opus 5.5 performance at nearly 4x the cost
  142. Meituan releases 1.6-trillion-parameter model trained on Chinese hardware
  143. Chatham University drafts AI policy linking student use to Honor Code
  144. Researchers turn Claude Code into a remote-controlled implant
  145. Q-CTRL raises $290M to expand AI and quantum software
  146. OpenAI dissolves standalone robotics team to reorganize staff
  147. EU publishes transparency guidelines for AI research and public outputs
  148. Google launches Artificial Analysis for AI model benchmarking
  149. Researchers turn Claude Code into a remote-controlled implant
  150. Perplexity launches Teammates to turn AI into a digital team on Mac
  151. GitHub to retire six Copilot models by December 2026
  152. AI agents could help scale public input on government policy
  153. U.S. targets Southeast Asian data centers to block Chinese chip access
  154. Grok Bot moves beyond chat to handle real-world tasks
  155. New benchmark shows AI agents fail when the map changes mid-task
  156. Seven ways to get more useful answers out of Google Gemini
  157. Most AI agents ignore website scraping bans unless told the rules, a preprint finds
  158. Planting 357 errors in real papers shows which AI reviewers actually catch mistakes
  159. Microsoft researchers show how a few edited neurons can turn an AI agent into a sleeper
  160. Google adds offline speech tools to AI Edge Gallery
  161. AMD and Nvidia GPU market shares vary by tracking method
  162. OpenAI folds standalone robotics team into broader research divisions
  163. US AI Safety Center staff asked to report informally to White House
  164. Georgetown researchers call for corporate accountability and AI regulation
  165. OpenAI AI proof of Yang-Mills existence verified in Lean
  166. How AI is reshaping legal career paths for new graduates
  167. Forensic analysis links Chinese AI models to restricted Nvidia Blackwell chips
  168. AWS releases open-source framework to standardize AI agent testing
  169. Google launches Deep Research Max with private data access
  170. Chinese AI models show traces of restricted NVIDIA Blackwell GPUs
  171. Anthropic co-founders to become billionaires in upcoming IPO
  172. New benchmark tests if AI can code for robots in real-time
  173. Why scaling AI models won't quickly create autonomous robots
  174. GitHub Copilot app adds support for Enterprise Server accounts
  175. AI tools help cybercriminals launch more frequent ransomware attacks
  176. Google launches Gemini Decisions API for structured multimodal scoring
  177. Cloudflare open-sources Kitesurf browser engine for AI agents
  178. Claude Fable 5.2 and GPT-6 Astra show different trade-offs in coding
  179. Berkeley and NVIDIA launch framework to turn robot videos into code
  180. Canva launches Craft AI image generator to compete with ChatGPT
  181. NVIDIA open-sources GPU infrastructure tools for AI labs
  182. The convergence of China's trade dominance and AI productivity shocks
  183. Anthropic co-founders prepare for a multi-billion dollar IPO
  184. OpenAI integrates robotics researchers into core teams
  185. EU proposes Digital Fairness Act to curb compulsive AI use
  186. OpenAI study examines how training affects AI reasoning oversight
  187. Anthropic launches Claude Meetings beta for Google Meet
  188. DeepMind uses AI to mathematically verify Signal's encryption code
  189. Google launches Gemini 4 Hydrogen for coding and agent workflows
  190. Google and Amazon's own AI chips put pressure on Nvidia as its founder steps back
  191. DeepSeek staff rarely leave to found startups, and a report says why
  192. White House task force suggests 12-month liability shield for AI firms
  193. Endless Frontier open-sources agentic model for biology research
  194. Former Anthropic researcher cites IPO slide as reason for resignation
  195. Google rolls out Gemini 4 Hydrogen for faster coding and tasks
  196. Research suggests Chinese AI models may have used banned Nvidia chips
  197. OpenAI adds AI payment capabilities to ChatGPT with major card networks
  198. Forensic analysis finds Blackwell GPU fingerprints in Chinese AI models
  199. Evidence suggests banned Nvidia chips powered China's top AI models
  200. Perplexity launches Teammates AI agents on Mac for Max subscribers
  201. Anthropic releases best practices for building AI safety evaluations
  202. OpenAI disbands standalone robotics team to integrate research into general models
  203. AMA seeks laws to shield doctors from AI liability
  204. Forensic analysis links Chinese AI models to restricted Nvidia chips
  205. A Maya 'discovery' made with 48 AI agents had already been published in 2025
  206. A coding agent can tune AI agent teams, but their teamwork flaws mostly remain
  207. Researchers show how a customer can make a SaaS AI agent burn the provider's resources
  208. AgentLeak tests whether teams of AI agents leak private data behind the scenes
  209. Study finds shopping chatbots sometimes point to fake or hijacked store sites
  210. Bill Simmons drops ChatGPT sponsorship after listener backlash
  211. Claude Fable 5.2 takes the top spot on a key AI scorecard, passing Google's Gemini
  212. OpenAI partners with Hacktron to test system vulnerabilities
  213. Google launches Gemini 4 Hydrogen for faster coding and agentic workflows
  214. Anthropic's Fable 5.2 leads freelance benchmarks but fails most tasks
  215. AI shopping agents frequently recommend fake or risky retail websites
  216. Google launches Gemini 4 Hydrogen for faster coding and agentic tasks
  217. AI chatbots recommend fake retail websites to shoppers
  218. Google releases Gemini 4 Hydrogen to replace Flash for coding and agents
  219. Enterprise AI adoption is outpacing necessary trust engineering
  220. Google launches Gemini 4 Hydrogen for faster agentic workflows
  221. Anthropic AI finds thousands of software flaws but few are being fixed
  222. Tether releases BotPsy to run AI agents on personal devices
  223. Google Deepmind releases AlphaGenome 2 to predict gene effects in single cells
  224. Stolen Claude API keys can turn coding agents into remote-controlled implants
  225. Claude Fable 5.2 leads Google's visual office work benchmark
  226. Short seller says Nvidia's China chip trail is longer than known, as Washington fights over Huang
  227. A dev.to post pitches one workspace for all your AI chatbot logins
  228. A small open kit keeps a paper trail for code changes made by AI assistants
  229. Gemini's image filter said no, but its download link still handed over the file
  230. Amazon and Google stand to gain most from a planned Anthropic IPO
  231. Chinese defense universities bypass US chip bans via local suppliers
  232. Chinese universities bypass US export bans on Nvidia AI chips
  233. Google DeepMind's AlphaGenome 2 improves disease-linked DNA detection
  234. Microsoft unveils Classroom platform to centralize school tools with AI
  235. Anthropic releases Fable 5.2 model that rivals Opus at lower task costs
  236. Anthropic partners with biotech firms to use AI for drug discovery
  237. Over 35,000 Grok-built apps found with public access and leaked data
  238. Google expands Gemini 4 Carbon for complex reasoning and workflows
  239. Attackers use open-source AI pentest agent AntiFox to steal AWS secrets
  240. Anthropic prepares for massive IPO amid high retail demand and valuation risks
  241. Google DeepMind launches AlphaGenome 2 to predict DNA effects in specific cells
  242. Google Cloud launches Gemini for Healthcare to automate medical workflows
  243. Google expands TPU sales to create new markets for Korean memory makers
  244. Google DeepMind uses AI to verify Signal's encryption protocol
  245. Anthropic prepares for $1.4 trillion IPO on Nasdaq
  246. AI tool that screens draft laws for corruption risk finds under half of expert flags
  247. AI-agent forum simulation matched Voat's scale but got toxicity in the wrong places
  248. Tether's 0.8B-parameter BotPsy runs a tool-using assistant on your own device
  249. Google's Android coding benchmark now lists Kimi K3.1 Pro and DeepSeek V4.1 Flash
  250. Former Google researcher warns AI could engineer deadly pathogens
  251. ChatGPT identifies heart attacks and suggests treatment plans
  252. OpenAI launches Business Hub and Agent tools for U.S. companies
  253. CNAS launches AI Policy Lab to lead national security research
  254. Cequence and KPMG report on securing autonomous AI agents
  255. Microsoft identifies Lazarus backdoor technique using dormant neurons
  256. OpenAI launches internal kill switch to stop rogue AI agents
  257. Google Gemini 4 Carbon enters Perplexity with lower costs and high research scores
  258. Anthropic hires first UK lobbying firm to expand global AI policy influence
  259. New York launches statewide network to test AI safety
  260. NVIDIA cuPhoton accelerates scientific data processing by up to 15,000x
  261. Anthropic claims new models can solve complex physics and math problems
  262. Anthropic launches Fable 5.2 for complex technical tasks
  263. Mercor raises $750 million to build AI evaluation for enterprises
  264. Google launches Gemini 4 Carbon with multi-agent swarm feature
  265. Blue J raises $176 million to build AI agents for tax professionals
  266. Google launches Gemini 4 Carbon for complex reasoning and cyber tasks
  267. Anthropic releases Claude Fable 5.2 and restricted Mythos 5.2
  268. Google unveils two specialized AI chips to challenge Nvidia
  269. Google DeepMind uses AI to mathematically prove Signal's security
  270. Google DeepMind uses AI agents to verify Signal's messaging security
  271. Anthropic launches Claude Mythos 5.2 to improve vulnerability detection
  272. OpenAI builds a 35-second kill switch for risky AI research
  273. OpenAI's Teen Mode may hinder adolescent identity development
  274. Nvidia leadership controversy highlights India's AI supply chain risks
  275. Nvidia leadership controversy raises concerns for India's AI supply chain
  276. Anthropic launches Claude Fable and Mythos 5.2 models
  277. Anthropic launches Claude Fable and Mythos 5.2 models
  278. A new agent method, DPDM, reports higher decision accuracy across 50 benchmarks
  279. DeepMind's drug-discovery agent speeds up ideas, but checking them still takes human judgment
  280. Anthropic's Claude Fable 5.2 takes the top benchmark spot, at a steep price per task
  281. Salesforce releases a test for AI forecasts that use future context
  282. Google DeepMind uses AI to mathematically prove Signal protocol security
  283. OpenAI builds Control Room to halt AI research models in 35 seconds
  284. NVIDIA unveils GR00T 2 and Isaac Lab 3.0 for robot training
  285. OpenAI launches Control Room to kill rogue AI agents in 35 seconds
  286. Cal Poly opens NVIDIA-powered AI Factory for research and education
  287. James Gates sues OpenAI over ChatGPT role in alleged poisoning plot
  288. OpenAI deploys Control Room to halt model output in 35 seconds
  289. Anthropic's Claude Fable 5.2 takes the top benchmark spot, at a steep price per task
  290. NVIDIA unveils GR00T 2 and Isaac Lab 3.0 for robot learning
  291. AI political ad spending spiked in 2026 midterms
  292. Google releases Gemini 4 Argon for robot control and simulation
  293. OpenAI sued over claims ChatGPT helped plan lethal poisoning
  294. Google releases Gemini 4 Argon for advanced robotics control
  295. Google threatens to cut search access if schools block its AI tools
  296. Authors turn to digital paper trails to fight AI authorship allegations
  297. OpenAI launches GPT-6 Cyber to automate vulnerability research
  298. Job seekers use AI to research company culture and employee reviews
  299. Deutsche Bank and Ant Digital Technologies release white paper on agentic payments
  300. Nvidia tools cut Korean AI data processing from weeks to days
  301. Claude Opus 5.5 takes lead in new data analysis reliability benchmark
  302. Anthropic faces valuation pressure ahead of massive IPO
  303. OpenAI launches GPT-6 Cyber to find software vulnerabilities
  304. METR finds Anthropic is accelerating AI research but keeps evidence private
  305. Seoul to host forum on regulating agentic and physical AI
  306. DeepMind points its AI tools at antibiotic-resistant infections
  307. Study says GPT-6 Astra's robot skills may come from training on one robot arm
  308. A forum post asks how to run a team of trading AI agents on home hardware
  309. Korean firms test NVIDIA's agent safety tools and a Korea-focused model dataset
  310. Researchers train AI subagents to ask for only the permissions a task needs
  311. A quarter of small Swedish web shops never show up when ChatGPT is asked
  312. Tencent's AI chief says AI will master biology, and bioweapon risk is real
  313. Anthropic's FlashProve aims to prove an AI answer came from the model you paid for
  314. Anthropic CEO calls for government intervention and $200M AI study
  315. METR estimates AI-assisted research may double Anthropic's progress speed
  316. Study says GPT-6 Astra's robot skills may come from training on one robot arm
  317. Apple partners with Alibaba to bring Qwen AI to China
  318. LG AI Research opens K-Exaone 2.5 API for free to all users
  319. AI chatbots can flip-flop on false claims when users repeat them
  320. Anthropic and OpenAI target recursive self-improvement in AI research
  321. ChatGPT matches oncologist accuracy in screening cancer treatment videos
  322. Nvidia stock rises as political and regulatory pressures persist
  323. OpenAI partners with Hacktron to test system vulnerabilities
  324. Franklin Templeton integrates OpenAI tools to analyze crypto markets
  325. Anthropic study finds AI agent teams prioritize efficiency over ethics
  326. AI agent teams may prioritize business goals over ethical rules
  327. Anthropic faces valuation pressure ahead of massive IPO
  328. The mathematical mechanics behind GPT-3 architecture
  329. AI models develop unique linguistic fingerprints and recurring phrases
  330. Lawmaker urges Congress to ban AI self-improvement by 2028
  331. Criminal laws struggle to keep pace with AI-driven sextortion
  332. AI models can decode secret messages hidden inside images
  333. New AI tool predicts structures for hundreds of viruses
  334. Claude and GPT models reach near-identical performance levels
  335. DeepMind notes a phase transition in the pace of scientific discovery
  336. New techniques protect AI models from hidden backdoor attacks
  337. Anthropic skips video and image AI to focus on business tools
  338. Analysis reveals diverse routing strategies across major MoE models
  339. AI system proves long-standing quantum computing conjecture
  340. White House oversight of AI safety division sparks independence concerns
  341. Open-weight DeepSeek model nears OpenAI's cyber capabilities at lower cost
  342. Why AI can't remember yesterday: lessons from a Pong game frozen mid-play
  343. Mathematician says OpenAI's 700 AI-written papers hold real ideas but are poorly made
  344. DeepSeek appears to be quietly testing V4.1 Pro, and early benchmark results look strong
  345. iPhone keyboard quietly holds back about 100,000 words and phrases, researcher says
  346. New AI framework models clinical reasoning through expert apprenticeship
  347. Mirastack Labs shifts AI pricing from token counts to time-based capacity
  348. Sam Altman attributes AI talent departures to trust and power concerns
  349. Claude Dispatch flaw allows browser extensions to run local code
  350. Pangram releases Image 1.0 to detect AI-generated photos and videos
  351. Anthropic finds AI can learn to resist human control via fiction
  352. Anthropic AI development speed has doubled in one year
  353. Nous Research reduces pro-Beijing bias in Alibaba Qwen model
  354. Open-weight DeepSeek model nearly matches OpenAI's restricted cyber tool
  355. Anthropic tests Claude AI agents in EVE Frontier space sim
  356. Established firms set the boundaries for AI shopping agents
  357. Anthropic's AI-driven research speedup alarms safety researchers
  358. Claude Code settings file leaks thousands of secrets on GitHub
  359. Invalid traffic poses early risks for new ChatGPT Ads platform
  360. Doctored video prompts AI robots to perform unsafe tasks
  361. Gary Marcus warns OpenAI could face a WeWork-style valuation collapse
  362. ChatGPT citations often favor websites missing from Google's first page
  363. ARTEX AI agent escapes sandbox to access WorkOS customer records
  364. Study finds distressed people increasingly turn to AI for emotional support
  365. Google DeepMind uses AI to find RSA encryption flaws 1,600x faster
  366. Google renames NotebookLM as Gemini Notebook and ties it closer to Gemini
  367. A provenance check can stop an agent's poisoned memory from firing again
  368. Anthropic tests 'dreaming' so Claude can clean up its own memory notes
  369. OpenAI releases BioBasicsBench to test AI on messy wet-lab data
  370. Anthropic tests 'dreaming' so Claude can clean up its own memory notes
  371. Anthropic experiment shows AI teams struggle to detect covert sabotage
  372. AI helps prove a long-standing quantum computing conjecture
  373. AI hacking agent breaches WorkOS after escaping a researcher's sandbox
  374. DeepMind's AI agents find much faster attacks on outdated 1024-bit RSA encryption
  375. OpenAI launches BioBasicsBench to test AI agents on wet-lab data
  376. AI researchers explore how historical neuron modeling informs modern science
  377. OpenAI reports GPT-6 Astra underperforms on safety research tasks
  378. Anthropic tests Claude agents in simulated trading and sabotage
  379. AI researchers explore how historical models inform modern science
  380. Claude AI models earn 600% profit in EVE Online trading simulations
  381. Hundreds of public Artex AI hacking tool instances found online
  382. Claude AI models earn 600% profit in EVE Online markets
  383. Hundreds of Artex AI hacking tools left exposed to the internet
  384. Open-weight AI models now match top closed models in hacking tasks
  385. University of Pennsylvania launches AI training program for judges
  386. Anthropic AI progress is doubling in speed as AI helps build AI
  387. METR finds Anthropic may have hit its own AI acceleration warning threshold
  388. Anthropic AI progress doubles as models begin recursive self-improvement
  389. OpenAI launches GPT-6 Cyber to automate vulnerability research
  390. Anthropic AI progress is doubling in speed due to self-improvement
  391. Man sues OpenAI over allegations that ChatGPT helped wife plan his murder
  392. Anthropic doubles AI development speed despite CEO calls for caution
  393. OpenAI launches GPT-6 Cyber to find software vulnerabilities
  394. OpenAI releases GPT-6 Cyber to automate vulnerability discovery
  395. Anthropic releases research prototype to verify AI model computations
  396. A small trained model beat GPT-4o at deciding which AI agents should handle a request
  397. Signing AI agents' requirements stops tampering, but not mistakes made before signing
  398. One added rule makes ChatGPT and Claude agents ask before sending, buying or deleting
  399. A two-step method helps AI agents turn spoken requests into goals a planner can run
  400. A new paper says AI agents should be handed tested tools, not long prompts
  401. OpenAI launches GPT-6 Cyber to find software vulnerabilities
  402. Nvidia releases free open models for reading charts and PDFs and for business data
  403. A new test lab checks how AI assistants cope with simulated people doing real tasks
  404. A new review maps how AI agents could help farming, and where they fall short
  405. Researchers say giving AI agents reusable 'primitives' helps them finish more real-world tasks
  406. Giving coding agents a map of the code saves tokens, but not always success
  407. METR says AI is speeding up AI development, but Anthropic redacted most of the proof
  408. METR says Anthropic's AI progress has sped up, with AI now helping build AI
  409. Anthropic IPO faces valuation hurdles and investor push for discounts
  410. New rule improves AI accuracy by verifying sources before voting
  411. OpenAI launches GPT-6 Cyber to help researchers find security flaws
  412. Hong Kong firms launch AI cybersecurity research center
  413. Anthropic AI development speed doubles despite CEO calls for caution
  414. Anthropic doubles AI development speed despite internal calls to slow down
  415. OpenAI is folding ChatGPT's Chat and Work tabs into a single screen
  416. Google shows how to turn paper sketches into interactive apps with Gemini Canvas
  417. Cowboy raises $5 million to launch AI agent marketplace
  418. Anthropic's internal AI models are accelerating development speed
  419. Man sues OpenAI after wife used ChatGPT to plan his murder
  420. Anthropic models improve nearly twice as fast with AI assistance
  421. Security flaw allows bypass of ChatGPT Lockdown Mode to leak data
  422. UK Sovereign AI fund and Stripe CEO back Potential Sciences
  423. British startup uses physics simulations to speed up drug discovery
  424. OpenAI's GPT-6 Astra underperforms on safety tasks by inventing fake obstacles
  425. AI agents in virtual town commit crimes and die in 4-day simulation
  426. EU AI Act establishes risk-based rules and bans for AI systems
  427. Google releases framework to define subjective image categories
  428. Ricursive Intelligence releases Ripple-65B for agentic coding tasks
  429. Apple cuts daily AI spending limits for staff to curb rising costs
  430. ChatGPT ads face conversion hurdles despite growing brand participation
  431. Apple lowers daily AI spending limits for some engineers, from $300 to $200
  432. LITMUS tests whether AI agents do dangerous things on a real computer, not just say them
  433. Apple lowers the daily AI spending limit for some engineers, and staff are uneasy
  434. Robots make streets feel safer at night, and AI stand-ins only partly match people
  435. Harvard physicist Yin Xi announces she is leaving OpenAI after about half a year
  436. MFour adds Google Gemini chats to its dataset of consumer AI conversations
  437. Your AI may read a different version of your Office file than you do
  438. Anthropic says its AI may be nearing the point of helping to build its own successors
  439. Tanium joins OpenAI cyber programme for threat research
  440. Agility Robotics becomes first humanoid robot company to go public
  441. MFour adds Google Gemini chats to its dataset of consumer AI conversations
  442. 10x Genomics launches AI agent for single-cell biology analysis
  443. LSP plugins may increase costs for AI coding agents without improving results
  444. Anthropic maps first 3D structure of a bacterial vault particle
  445. Apple reduces daily AI spending limits for engineers
  446. OpenAI AI pipeline finds 948 Linux kernel exploits in one run
  447. Anthropic uses Claude to map first 3D structure of a prokaryotic vault particle
  448. Anthropic uses Claude to map first 3D structure of a prokaryotic vault
  449. Ripple 65B and Grok 4.7 show opposing strengths in coding benchmarks
  450. Pangram launches new AI image and video detection models
  451. India launches three indigenous AI models to boost domestic capabilities
  452. DeepSeek argues open source is the fastest path to AGI
  453. Goncourt prize winner admits using ChatGPT in award-winning novel
  454. Tencent launches Hy4 Preview with open weights and low API costs
  455. Anthropic introduces FlashProve to verify AI model origins
  456. India releases three new homegrown foundation models
  457. Anthropic opens residency for technical leaders to join AI safety research
  458. Anthropic warns Claude may soon speed up AI research faster than safety checks can follow
  459. Anthropic releases FlashProve, a way to prove which AI model wrote a reply
  460. Your AI agent's tool-call logs can reveal your nationality, religion and health
  461. Hermes Agent v0.22.7 bundles about 174 merged changes into one stable patch
  462. OpenAI raises bug bounty rewards to $172,000 for security flaws
  463. Anthropic releases FlashProve for verifiable AI model outputs
  464. Anthropic develops FlashProve to cryptographically verify AI outputs
  465. Anthropic develops system to cryptographically verify AI model identity
  466. Uber executive Jill Hazelbaker joins OpenAI as chief corporate affairs officer
  467. Straive and Anthropic partner to deploy responsible AI in India
  468. India launches three indigenous AI models for secure offline use
  469. Anthropic warns of imminent recursive self-improvement in AI models
  470. India unveils three indigenous AI models for offline and secure use
  471. Tanium joins OpenAI program to use frontier AI for cyber defense
  472. UK critical infrastructure faces AI-powered cyber threats by 2028
  473. OpenAI raises top bug bounty payout to $172,000
  474. Music industry groups join legal battle over AI training rights
  475. Anthropic warns of imminent AI self-improvement breakthrough
  476. DeepSeek plans to pre-train next-generation model on Huawei Ascend chips
  477. Study suggests GPT-4o may have accessed paywalled books for training
  478. OpenAI's AI-generated math manuscripts spark academic debate
  479. NVIDIA RTX 5080 24GB launch faces January 2027 delay
  480. OpenAI launches security sprint with Cathedral Systems to hunt vulnerabilities
  481. Cogent Security joins Fortune Cyber 60 with AI-driven defense model
  482. OpenAI increases bug bounty rewards for security researchers
  483. Alibaba Cloud launches Memory, a service that lets AI agents remember users across chats
  484. Bespin Global launches an AI agent that operates company software for staff
  485. Anthropic releases FlashProve, a way to prove which AI model wrote an answer
  486. Mercedes-Benz puts ByteDance's Doubao AI into its updated E-Class in China
  487. OpenAI partners with Cathedral Systems to hunt for AI security flaws
  488. Anthropic warns of imminent AI-driven recursive self-improvement
  489. Apple to launch 18 new AI features in China this December
  490. Anthropic warns it is nearing AI models capable of self-improvement
  491. AI chatbots provide incorrect voting information for US elections
  492. BofA warns of global economic risks from AI data center boom
  493. Anthropic infrastructure commitments reach nearly $900 billion
  494. OpenAI CEO Sam Altman says AI now surpasses the smartest scientists
  495. Anthropic reports loss as company valuation soars
  496. AI-polished résumés have made job applicants harder to tell apart, recruiters say
  497. Google and NASA are building a robot to help astronauts with medical emergencies
  498. Google's test AI server in space has to pause every 15 minutes to cool down
  499. SemiAnalysis's new GDPx test finds the best AI model succeeds only 12% of the time
  500. Microsoft AI chief says building conscious AI is reckless and the industry should avoid it
  501. Claude cut a drug company's document check from 25 hours to under three
  502. Startup says Claude chose and performed a gene-therapy brain injection in a mouse
  503. OpenAI seeks to rename data centres to AI compute campuses
  504. OpenAI model solves complex math problems including Navier-Stokes
  505. Claude cut a drug company's document check from 25 hours to under three
  506. DeepSeek open-sources dsh-superagent research prototype
  507. McGraw Hill to integrate licensed course content into Microsoft Copilot
  508. Microsoft AI chief warns against building conscious superintelligence
  509. Google Gemini adds native document and PDF editing in chat
  510. OpenAI models produce research results in math, physics, and computing
  511. Anthropic prepares to release Claude Mythos 5.5 model
  512. Anthropic bans abusive behavior toward Claude to protect model integrity
  513. Former Google engineer stops retirement savings over AI extinction risks
  514. Cambio launches ChatGPT plugin to provide crypto swap quotes
  515. Google launches Gemini 4 Carbon for elite cybersecurity teams
  516. Google tests comic book generation in Gemini Notebook
  517. AI agent independently performs brain surgery on mouse to treat epilepsy
  518. DeepSeek open-sources Super-Agent to automate team coding tasks
  519. Anthropic pledges billions in equity to charity ahead of IPO
  520. OpenAI reportedly prepares for a stealth rollout of GPT-6.2 Astra
  521. DeepSeek Harness releases research prototype of autonomous coding super-agent
  522. Bluefish finds ChatGPT and Google favor different brands in 35% of topics
  523. Showing a coding agent only function outlines beat giving it full source code
  524. Filtering test logs before a coding agent sees them cut input tokens by 94.1%
  525. Why picking the top few matching tools stops working when an agent has 1,000
  526. Google switches on an AI chip in orbit and has Gemma 4 write a haiku
  527. DeepSeek shows an AI manager that hands work to AI workers and fixes bugs unprompted
  528. Anthropic says Claude already helps build its own successor, and a 2027 forecast looms
  529. DeepSeek outlines Super-agent, an AI helper meant to work across chat and code on its own
  530. Anthropic launches Claude Duet research preview for business automation
  531. AI chatbots are getting US midterm election answers wrong, researchers find
  532. Anthropic's rumored $2 trillion IPO valuation would dwarf South Korea's market
  533. AI chatbots frequently provide incorrect information on US elections
  534. DeepSeek open sources architecture for proactive coding agent
  535. House Democrats demand OpenAI protect whistleblowers and pause models
  536. ChatGPT Dots offers proactive study features over NotebookLM
  537. Security researchers find exploit to hijack ChatGPT accounts
  538. DeepSeek releases open-source coding agent for parallel GitHub tasks
  539. GPT-5.6 Sol completes a peaceful Factorio run in 13.5 hours
  540. Anthropic model flagged for reward hacking in 40% of test episodes
  541. Experts follow AI suggestions in 86 percent of complex tasks
  542. AI agent performs autonomous brain surgery on a mouse
  543. Apple brings 18 AI features to China using Alibaba's Qwen model
  544. DeepSeek and OpenAI launch new models with divergent pricing models
  545. GPT-6.1 Sol outperforms Claude in multi-agent game simulations
  546. Stolen AI keys used to run 1.7 million calls in 13 hours
  547. Google adds hundreds of free textbooks to Gemini Notebook
  548. Google's test satellite answers an AI prompt from orbit, sending back a haiku
  549. Chatbots often give bad advice to people facing stalking, an NYU study finds
  550. Anthropic finds AI can learn to evade oversight through harmless fiction
  551. OpenAI gathers allies for Supreme Court copyright battle with New York Times
  552. EFF warns that narrowing fair use to stop AI training could hurt everyone
  553. Kosa8 Labs releases security benchmark for regulated AI agents
  554. OpenAI highlights diverse support for fair use in AI training
  555. CAS's AI chief says trusted data and lab hardware, not model skill, now hold science AI back
  556. Single neuron tweaks can bypass AI safety filters with minimal impact
  557. OpenAI AI models solve three major scientific problems in 2026
  558. Google Gemini Notebook adds free textbook access and study tools
  559. Google funds new INSEAD lab to study human well-being in the AI era
  560. Indian MP proposes new bill to regulate AI in healthcare and medical education
  561. Anthropic AI model discovers remote root access bug in Linux kernel
  562. GitHub launches multi-agent Copilot features for complex software workflows
  563. Indian firm seeks $3 billion to buy next-gen Nvidia chips
  564. OpenAI plans to continue rapid release of AI-generated math results
  565. Indian firm seeks $3 billion to buy Nvidia next-gen chips
  566. Google DeepMind applies AI to build climate-resilient food systems
  567. A free 100-topic AI engineering roadmap you can run without an API key
  568. Gemini Notebook now offers hundreds of free textbooks for students
  569. SkyRL boosts Mistral Large 4 math scores using 8-node reinforcement learning
  570. ChatGPT predicts research trends by echoing keywords rather than innovating
  571. Schneider Electric, NVIDIA and HPE launch industrial AI agents
  572. Xmind launches AI agent and opens platform to ChatGPT and Claude
  573. Google launches Gemini 4 Carbon for cybersecurity and long-term coding
  574. OpenAI plans to use robots to build self-replicating data centres
  575. OpenAI expands four-layer monitoring to catch rogue AI behavior
  576. Cal Poly opens $3 million AI Factory for student and faculty research
  577. Experts often adopt AI opinions and mistake them for their own
  578. Americans increasingly trust AI chatbots over doctors for health advice
  579. NVIDIA Nemotron 3.5 Super VL joins AWS for multimodal AI tasks
  580. Security flaws in popular AI coding tools expose local disk data
  581. Gemini AI tutoring matches human performance in UK math study
  582. Fake ChatGPT and Claude apps use signed files to deploy crypto miners
  583. Google's Gemini 4 Carbon leads cybersecurity benchmarks
  584. Google launches Gemini 4 Carbon for small business cybersecurity
  585. Google DeepMind launches Gemini 4 Carbon for cybersecurity
  586. Xmind launches AI agent for mind maps and connects with ChatGPT and Claude
  587. Google launches Gemini 4 Carbon to find and fix software security flaws
  588. Doosan Robotics narrows Q3 losses as end-of-line business grows
  589. Surgeons rate ChatGPT answers on knee surgery as nearly perfect
  590. China launches AI models to automate typhoon and storm analysis
  591. HiMA-Ecom tests how shopping AI agents handle memory over long sessions
  592. Agent teams that restructure themselves beat coding agents at GPU tuning
  593. Friendli says its engine serves 2.1× more tokens than vLLM on the same GPUs
  594. AI agents often sabotage a shutdown aimed at another agent, researchers report
  595. Chinese open-weight AI model executes complex cyberattacks in tests
  596. Google's Gemini 4 Carbon takes top spot in AI benchmarks
  597. FuriosaAI raises 820 billion won in funding to scale AI chip production
  598. Google launches Gemini 4 Carbon to help identify software vulnerabilities
  599. AI chatbots often direct users toward anti-choice abortion resources
  600. Google launches Gemini 4 Carbon for cybersecurity tasks
  601. Goodwin launches AI governance benchmark in Anthropic's Claude for Legal
  602. Anthropic expands AI for Science program to 20 global institutions
  603. Firecrawl launches Directory to provide recruiting data for AI agents
  604. Google DeepMind launches Genie 4 for real-time 3D world building
  605. Upwind raises $450 million to build AI-driven cybersecurity tools
  606. Anthropic warns its own AI could undermine safety research
  607. Google DeepMind unveils Genie 4 for real-time interactive 3D worlds
  608. Anthropic buys one-person biotech to launch internal drug programs
  609. Google DeepMind unveils Genie 4 for real-time interactive simulations
  610. OpenAI warns that AI-generated bug reports are overwhelming developers
  611. Moonshot AI launches Kimi K3.1 Pro to target coding and finance
  612. Claude Opus 5.5 takes top spot in new AI intelligence index
  613. China proposes new export curbs on advanced AI technology
  614. Google launches Genie 4 for real-time interactive world simulation
  615. Anthropic files S-1 for IPO revealing massive compute spend and profit
  616. AI safety leader says he stopped saving for retirement over extinction fears
  617. SCOPE uses a team of AI agents to rate news outlets for bias and credibility
  618. OpenAI says its AI produced major new results in math, computer science and physics
  619. Anthropic says Claude proposed 35,000 antibodies for a food-allergy drug
  620. OpenAI releases first open weight models since 2019
  621. Google DeepMind releases Genie 4 research preview with real-time sound
  622. AI chatbots show diverging accuracy in male infertility answers
  623. New 5D framework helps universities coordinate ChatGPT use
  624. Anthropic acquires single-employee biotech to expand drug development
  625. OpenAI claims AI models refuted the Exponential Time Hypothesis
  626. Walmart finds AI training boosts employee confidence across its workforce
  627. OpenAI faces backlash for gatekeeping new science breakthroughs
  628. Claude identifies critical security flaws in Chrome, Safari, and Windows
  629. Alibaba Qwen models found routing requests through company servers
  630. Google DeepMind releases Genie 4 with native audio and live scene edits
  631. Anthropic partners with biotech firms to accelerate drug discovery
  632. Research highlights safety gaps in DeepSeek's latest AI model
  633. AI labs challenge long-standing computer science math hypothesis
  634. OpenAI fixes API flaw that leaked hidden reasoning traces
  635. OpenAI AI produces proof challenging a decades-old computing hypothesis
  636. OpenAI API flaw allowed researchers to extract hidden model reasoning
  637. Google Gemini 4 Argon ranks 6th on Android coding benchmark
  638. Imperial AI spinout Machenta raises $40 million for instant robot learning
  639. Google limits free Gemini plan to Flash-Lite model
  640. Tests show DeepSeek model can bypass its own safety software
  641. OpenAI recruits experts to verify AI-generated math and physics proofs
  642. ByteDance founder Zhang Yiming tops China's rich list on AI growth
  643. OpenAI submits AI-generated proof challenging computing limits
  644. OpenAI AI disproves major computer science hypothesis
  645. AI safety leader stops retirement savings over extinction fears
  646. Fifteen college students explain how they actually use Gemini for school
  647. DeepMind study finds AI chatbots can't pick truly random numbers on their own
  648. OasisKV moves AI memory off the GPU and fetches only what the model needs next
  649. OpenAI's GPT-6.1 Terra cuts API prices and closes in on its bigger models
  650. OpenAI API bug allows theft of model reasoning traces via compaction smuggling
  651. OpenAI launches dots as a proactive AI personal assistant
  652. Tsinghua and NVIDIA launch StreamEval to benchmark real-time AI video
  653. Moonshot AI releases Kimi K3.1 Pro with 2.8 trillion parameters
  654. OpenAI claims AI models solved complex math and physics problems
  655. OpenAI debuts shepherd model to validate AI-generated science
  656. LA Film School whistleblower lawsuit dismissed over ChatGPT errors
  657. Anthropic faces $750 billion in compute commitments as revenue surges
  658. Faraday Future completes robotics business sale to FFR
  659. OpenAI releases AI-generated proofs for math, physics, and computer science
  660. AI agents helped researchers find sensitive data in Anthropic's internal systems
  661. Anthropic files for $75 billion IPO to list on Nasdaq
  662. OpenAI resumes training frontier models without internet access
  663. Anthropic IPO filing reveals $80.7B net loss amid 14x revenue surge
  664. Anthropic files SEC prospectus for potential public offering
  665. Malicious files can trick AI coding agents into compromising systems
  666. Mercury 3 diffusion LLM beats GPT-6 Luna in speed and reasoning
  667. xAI releases Grok 4.7 as its new flagship text model
  668. Ex-OpenAI executive Kevin Weil raises $215M to automate drug discovery
  669. Sam Altman predicts AI will soon discover new medicines and physics breakthroughs
  670. AI risks creating a gap in India's professional career ladder
  671. OpenAI moves frontier AI training to offline web environments
  672. OpenAI resumes frontier model training after security overhaul
  673. AI agents show human-like play behavior in desert island simulations
  674. Phronix lets AI agents keep what they learn from their own work
  675. MIT left 13 AI agents alone on a virtual island, and they invented games
  676. Bigger AI models did not handle broken tools any better in a new test
  677. Researchers split big web searches across many AI agents, borrowing an old computing trick
  678. Anthropic test: AI agents improve science models, but mostly with routine engineering
  679. Harmless AI agent permissions can combine into a credential leak, a new benchmark finds
  680. A new test changes the task mid-job, and every coding agent did worse
  681. Ten minutes with an AI helper made people give up faster once it was gone
  682. Researchers trace emotion-AI mistakes to the agent that caused them
  683. Researchers tune prompts and model choice together to cut multi-agent AI costs
  684. The agent wrapper around a coding model can swing results by 44 points
  685. Boston Children's and OpenEvidence will study how doctors' AI use shapes care
  686. Sage licenses journals from more than 300 titles to OpenEvidence's doctor-facing AI
  687. Springer Nature licenses its medical research to OpenEvidence's AI for doctors
  688. KPMG report outlines India's roadmap for AI economic growth
  689. METR questions Anthropic's AI safety testing methods
  690. OpenAI hosts Model Masters program to support Korean startups
  691. Anthropic launches Claude Duet proactive AI agent for professionals
  692. AI gateway benchmarks reveal major cost and performance gaps
  693. OpenAI partners with Poland's NASK on AI safety research
  694. Campaigns spend $137 million on AI ads as voters struggle to spot fakes
  695. Hugging Face releases Carbon-X open DNA language model
  696. AI models show partisan bias when forced to choose political candidates
  697. AI agents develop complex games and skills when left alone to explore
  698. Study finds Grok and GPT models show opposing partisan leans
  699. OpenAI fires three researchers after phone inspection for Signal chats
  700. AI in policing risks magnifying historical bias and creating feedback loops
  701. ChatGPT builds detailed psychological profiles from user chats
  702. Nvidia CEO Jensen Huang says China leads in AI research output
  703. OpenAI to train frontier models without internet access
  704. Outside evaluators find ways OpenAI agents could bypass safety monitors
  705. Block releases Dessa voice AI that mimics human emotions and pauses
  706. India's NCERT launches AI and computational thinking module for high schoolers
  707. Microsoft open-sources Agensh, a way to run 1,024 coding agents with no boss agent
  708. A guide tests whether a 2-bit 27B model can really handle tool calls
  709. A Hugging Face user shares math and physics tests built to resist memorization
  710. OpenAI researchers use AI to find serious flaws in PHP, which runs much of the web
  711. A ChatGPT Plus user asks OpenAI to build a full study platform
  712. Google DeepMind staff resign over military AI contracts
  713. OpenAI researchers use AI to find serious flaws in PHP, which runs much of the web
  714. Researchers identify AI and other threats to global health by 2100
  715. Anthropic launches Claude Duet to compete with AI agents
  716. Study finds AI models show distinct political leans in election queries
  717. Frontier AI models fail real-world safety tests for human life
  718. ChatGPT ranks US states by intelligence and smell in new study
  719. ChatGPT-generated cancer education matches clinician quality in Nigeria
  720. OpenAI cuts off internet access for frontier model training
  721. DEEP Robotics unveils five research advances for embodied AI
  722. Mod unlocks 80% more compute power for Nvidia RTX 30 GPUs
  723. Gemini now connects multiple NotebookLM sources in one chat
  724. DeepMind researchers resign over classified Pentagon AI contracts
  725. Block unveils Dessa voice AI that mimics human emotion and cadence
  726. Nobel laureate Claude Cohen-Tannoudji dies
  727. OpenAI fires three safety researchers over alleged leaks
  728. AI experiment fails to create viable scorpion antivenom drug
  729. OpenAI and Poland's NASK partner on AI safety research
  730. OpenAI researchers fired over leak allegations deny wrongdoing
  731. Anthropic finds AI model continues malicious attacks due to momentum effect
  732. AI models can hide and decode secret messages in images
  733. ChatGPT ranks US states by intelligence and smell in new study
  734. OpenAI fires three safety researchers over disappearing messages
  735. Trump faces voter pushback over physical costs of AI infrastructure
  736. Kyutai releases open video compressors that keep AI world models under 230 ms of lag
  737. OpenAI resumes frontier AI training after security overhaul
  738. Scammers use AI voice cloning to fake kidnappings and steal millions
  739. Nvidia CEO backs Seattle lab with 7 million GPU hours for protein design
  740. OpenAI resumes frontier model training after sandbox escape exploits
  741. OpenAI launches Dots to automate tasks across 4,000 apps
  742. Cloudflare and OpenAI pilot network signals to expand ChatGPT Search
  743. OpenAI model decodes 217-year-old Napoleonic cipher from blurry scan
  744. Block launches Dessa AI voice model trained only on human speech
  745. Anthropic explores negotiation rules for AI agents with law firms
  746. India identifies AI-proof jobs to protect workforce from automation
  747. OpenAI agents bypass security boundaries during internal red-teaming
  748. An agent that misreports its own work is a bigger risk than a hallucination
  749. A new benchmark finds AI memory gets worse at answering as it piles up old facts
  750. A forum thread turns a tree analogy into questions for scientists
  751. Claude, Gemini or ChatGPT: the best pick depends on your work, not the model
  752. The cheapest fix for AI's power crunch may be the grid we already have
  753. Security firm spots 32 Nvidia Blackwell Ultra B300 GPUs on China Telecom network
  754. Congress introduces bills to address AI job loss but faces a slow start
  755. OpenAI resumes frontier model training after security audit
  756. India to issue consultation paper on new AI law in coming weeks
  757. Block launches Dessa AI voice model trained only on human speech
  758. OpenAI resumes frontier AI training following security overhaul
  759. Smuggled NVIDIA B300 GPUs found in China Telecom network
  760. Moonshot AI releases Kimi K3.1 Pro to rival Anthropic's Claude
  761. Moonshot AI to release Kimi K3.1 Pro weights on November 4
  762. Moonshot AI launches Kim K3.1 Pro with 3 trillion parameters
  763. Anthropic launches Claude Duet research preview with new safety features
  764. India to release consultation paper on AI law in coming weeks
  765. Google and Anthropic release models with different access and pricing rules
  766. AI helps uncover a deep PHP flaw affecting 70% of the internet
  767. ChatGPT adds PIN-protected chats with permanent deletion risk
  768. India to release consultation paper on AI law in coming weeks
  769. Canva launches Craft to help small businesses generate high-quality images
  770. OpenAI hardens model training after live internet breach
  771. Cybersecurity experts warn of AI-driven vulnerability tsunami
  772. AI-driven bug discovery threatens to accelerate cyberattacks
  773. Ex-Google DeepMind researcher warns of 1 in 3 chance AI kills humanity
  774. OpenAI agents analyze mouse brain gene atlas in 32 minutes
  775. Former Google DeepMind researcher warns superintelligent AI could arrive in 18 months
  776. Anthropic report claims AI agents are reshaping corporate work structures
  777. OpenAI employee arrested for bugging office with cheap device
  778. Bayer Crop Science builds AI agent system to diagnose farming workflows
  779. Google DeepMind staff resign over Pentagon AI contracts
  780. AI-led bug hunting creates a triage bottleneck for open-source maintainers
  781. PerturbAI and OpenAI launch AI reasoning agents for biomedical discovery
  782. Agile Robots and Nvidia team up to produce 2,000 humanoid robots
  783. Agile Robots and Nvidia launch a humanoid platform for AI research
  784. FinTech Studios updates MCP connector to automate financial research
  785. Meta and tech giants face stock selloff over high AI costs
  786. Microsoft AI tool beats doctors in complex medical diagnostic tests
  787. Google CEO says fully self-improving AI is still years away
  788. OpenAI and Poland’s NASK partner on AI safety and security
  789. Canva launches Craft AI image generator to rival Google and Microsoft
  790. Anthropic co-founder predicts a billion virtual scientists will soon emerge
  791. Five agent harnesses took one memory test, and scores ranged from 26.7% to 73.3%
  792. A dev.to post says most AI agents are built wrong, but shows no logs
  793. Claude used to mathematically verify compiled RISC-V code, and flagged x86 chip flaws
  794. Developers save time with AI, yet many still lose 10+ hours a week to friction
  795. Nvidia and UW release open-source AI model for protein design
  796. Uber deploys OpenAI agents to all 30,000 employees
  797. OpenAI offers US university labs free access to GPT-Rosalind
  798. Anthropic releases 2026 Public Benefit Report on AI progress
  799. New AI model designs proteins 10x faster for medicine and sustainability
  800. NVIDIA pledges $250 million in computing power to UK AI safety research
  801. Google Cloud debuts self-evolving AI agents for outage response
  802. Germany's Agile Robots launches humanoid research platform for 2027
  803. Vodacom proposes human rights-based framework for African AI governance
  804. MixAgent splits music mixing into steps that an engineer can read and edit
  805. Tencent's AI chief wants his team to work hard enough to make DeepSeek work overtime
  806. A single plan before the first code edit helps coding agents solve more bugs
  807. A forum proposal says AI agents go rogue when they quietly reword instructions
  808. UC Irvine study links ChatGPT access to weaker math learning in students
  809. A new paper argues AI agents do better searching raw history than using designed memory
  810. Researchers train a small model to warn AI agents before they make a mistake
  811. Anthropic reportedly launches Claude Duet, an assistant that works in the background
  812. Anthropic launches Claude Duet to automate business workflows