Models, products and AI startups, near-duplicates collapsed to the most credible source
NYT: How AI Blew Up a Key Tenet of Computer ScienceThe New York Times published a story on how AI overturned a key tenet of computer science, but the provided text contains only the headline and no details.
Florida researcher develops agriculture AI agent GaiaUniversity of Florida assistant professor Nikolaos Tziolas built Gaia Bot, a conversational agricultural AI model trained on Florida academic datasets and satellite maps, helping small growers query crop damage, irrigation and pesticide use.
Politico: How the AI Doomers Won 2026Politico published a piece on how 'AI doomers' prevailed in 2026, but the provided text contains only the headline with no details to summarize.
Tau Robotics Launches Humanoid Robot Cleaning Service in San FranciscoStartup Tau Robotics offers humanoid robot cleaning in San Francisco, with robots currently teleoperated by humans who use head- and wrist-mounted cameras to collect training data for future autonomy. An hour-long session cost $30 initially; pricing is now $60 weekly with two robots or $100 per one-time visit. The robots are capable but slow.
Building visibility via open source contributionsAn MSc student in 3D computer vision asks how to build credibility in AI/ML by contributing to established open source projects rather than standalone GitHub repos, seeking advice on challenges for new contributors and how to get meaningful work noticed.
French charity warns teens not to rely on AI chatbots for bullying supportFrench charity e-Enfance found one in three young people have experienced bullying, including cyberbullying, and a quarter of them turned to AI chatbots for advice. Experts warn chatbots offer only temporary relief and cannot replace adult support, citing a 2025 case where a 16-year-old US teen died by suicide after months of chatting with ChatGPT.
Study of 207,000 Claude sessions shows how scientists use AIResearchers analyzed 207,000 Claude and Claude Code research sessions and found heaviest use in formal sciences like computer science and math, less in health and lab-based physical sciences. Data analysis accounts for about a third of Claude sessions and nearly half of Claude Code sessions, while idea generation and data gathering—rated most important by scientists—are used less. Tasks AI performs well get two to four times the usage predicted by researcher time spent.
Reddit Users Find Qwen2 0.5B Model's Answers AmusingA user on the Reddit LocalLLaMA community shared humorous exchanges with the tiny Qwen2 0.5B model (Q4_K_M quantization), calling its answers amusingly dumb. The post is casual community banter with no substantive news.
HSBC strategist says market fixated on Anthropic IPOHSBC chief multi-asset strategist Max Kettner told Bloomberg Television that mounting uncertainty over the AI trade has triggered a global stock selloff, with his team in full P&L protection mode. He said market stabilization could hinge on the Anthropic IPO and whether it goes through in the next two weeks.
Rightmove, Scout24 raise outlook on AI productsUK property platform Rightmove raised its 2026 EPS growth guidance to at least 7% from 5%, crediting AI tools such as call transcription and property performance summaries for estate agents with boosting revenue and productivity targets. German peer Scout24 also lifted its outlook, with both citing AI products as new revenue streams amid subdued housing markets.
OpenAI Partners with Hacktron to Find VulnerabilitiesOpenAI announced a partnership with Hacktron to leverage frontier AI to find vulnerabilities in its systems, test connections with third-party services, and strengthen defenses. The move aims to enhance OpenAI's own security.
App Store 2026: 9.8% of apps mention AI, 20.3% unupdated for a yearAn analysis of 178,637 charting apps across 173 countries found 9.8% mention AI in their listings, rising to 43% in Graphics & Design, and apps with in-app purchases are three times as likely to cite AI as free apps. Meanwhile, 20.3% have gone a year or more without an update, and independent developers account for 67.3% of charting apps with ratings matching big publishers.
ING's Treon Says AI Fundamentals Remain IntactAnneka Treon, ING's global head of private banking, wealth management and investments, says firms across tech and other sectors still report tangible AI value and want to invest more, signaling strong demand. She argues recent news does not derail the AI thesis.
AI Unwind Turns Europe Into a Go-To TradeAs global investors move to shed artificial intelligence exposure, Europe's diversified equity market has suddenly become one of the most attractive trades, signaling a shift away from concentrated AI bets toward regional diversification.
AI selloff deepens as Japan, South Korea markets tumbleGlobal stock selling intensified in Asia on concern the US will tighten regulation of the AI industry after Democrats' decisive election win, with Japan's benchmark index posting its biggest drop in over a year. AI-linked shares came under pressure as market sentiment turned cautious.
Palo Alto Networks in talks to buy AI agent security startup Clutch for $500MPalo Alto Networks is in talks to acquire Clutch, a three-year-old startup securing AI agents' access to enterprise systems, for $500M in cash plus milestone payments. Clutch was valued at $85M in its $20M Series A in January 2025 and is nearing $10M in ARR, implying a roughly 50x ARR price. The deal would be Palo Alto Networks' seventh acquisition this year amid rising AI-driven cyberattacks.
Hugging Face CEO says his P(Doom) is 'close to zero'Hugging Face's CEO told WIRED that his estimate of the probability that AI leads to human extinction is 'close to zero', contrasting with warnings from some AI-safety advocates. The remarks sparked discussion on Hacker News.
Did a data center revolt shape the US midterms?An AP-led voter poll found 56% of respondents called data centers a crucial or major factor in their vote, and 76% would oppose a new data center in their community. BBC notes the issue ranked well below corruption and cost of living.
Asia hedge funds, wealthy investors hit by AI selloffA sharp overnight US equity selloff reignited fears that AI bets are still unwinding, leaving Asian hedge fund managers and brokers fielding panicked client calls. Wealthy clients from Taiwan to Hong Kong faced margin calls on leveraged bets, and some in Hong Kong braced for losses on structured notes after being forced to take delivery of shares.
SoftBank leads Asia bond losses as AI rout spreads to creditAs the rout in AI stocks deepened and spread into credit markets, SoftBank Group led declines in Asian dollar bonds.
Researcher Questions Whether AI R&D Is Truly IngeniousIn a blog post, Arjun Panickssery cites side projects including a Scrabble engine, an Ultimate Tic-Tac-Toe bot and philosophy proof formalization, saying models show basic judgment flaws such as missing key papers or designing uninformative experiments, casting doubt on autonomous AI R&D ingenuity.
How accurately did science fiction predict AI?A retrospective essay assessed how well science fiction predicted AI, arguing sci-fi describes rather than predicts. The author broke down how AI was actually built, used and deployed, then scored stories; Ted Chiang's 'The Lifecycle of Software Objects' ranked first, with Richard Powers' 'Galatea 2.2' at 8.4/10. No pre-2017 work clearly described Transformer attention or the path from chatbots to agents.
CR Longdation and GCSA partner to launch AI and cybersecurity research centreCR Longdation Aethereal Cybersecurity and the Global Cybersecurity Alliance (GCSA) announced a partnership to set up an "AI and Cybersecurity Research Centre" in Hong Kong. The centre will combine both sides' resources in AI, cybersecurity and industrial applications, focusing on AI anti-fraud, threat detection, digital finance security and security automation for the Hong Kong government, public bodies and enterprises.
Plato's Allen says AI selloff stems from conflated return thesisDavid Allen, head of long/short strategies at Plato Investment Management, said Thursday's selloff came from investors conflating AI's transformational potential with the investment return thesis. The Nasdaq 100 fell 6.4%, wiping out $2.3 trillion in value, on speculation congressional Democrats will tighten AI industry rules.
US political shift could reshape AI regulationUniversity of Queensland AI risk specialist Michael Noetel told Ticker News that a changing US political landscape could open the door to stricter AI oversight. He said AI is growing unpopular with parts of the US public, and opposition to energy-hungry data centers even in Republican areas could foster bipartisan risk legislation. He favors safeguards over halting AI, and said US changes may influence Australia's governance approach.
Inkling 2 Small launches as 594B-parameter MoE modelInkling 2 Small, the second-generation model, is a Mixture-of-Experts system with 594B total and 17B active parameters, image inputs and variable effort. It scores 31 on the Artificial Analysis Intelligence Index, beating prior models. It debuts via Tinker with closed weights; full weights are planned under Apache 2.0 next week after external safety testing and red-teaming.
Dan Niles: open-source competition slows AI revenue growthVeteran tech investor Dan Niles says open-source competition is pushing down token prices and slowing revenue growth at leading AI firms. With higher financing costs and delayed IPOs, valuations must adjust, so investors should be selective instead of treating every AI-linked asset as an opportunity.
Australian investor Sahota avoids data centers, AI picksRevolution Asset Management CIO Bob Sahota said his firm takes a conservative, cash-flow-based approach to private credit and is not betting on data centers or picking AI winners. He spoke on the sidelines of the Sohn Hearts & Minds conference in Queenstown, New Zealand.
Nvidia launches NVKE to run managed Kubernetes on own GPUsNvidia released a preview of NVIDIA Kubernetes Engine (NVKE), a management plane that lets platform teams create and run Kubernetes clusters on their own GPU infrastructure, with each workload cluster getting its own control plane, etcd and node pools. Nvidia warns APIs may change and does not recommend NVKE for production workloads yet.
Report: closed models dominate price-performance frontier as open share hits lowVillage's LLM Atlas shows open-weight models are the best buy at only about 3% of budgets, the lowest measured, versus at least 6% at each quarter-end since 2024. The share hit 41% in June on DeepSeek V4 Flash, before OpenAI's GPT-6 Luna, GPT-6.1 Terra and Anthropic's Claude Haiku 5.5 undercut those budgets.
The A.I. Daily