Models, products and AI startups, near-duplicates collapsed to the most credible source
Hacker reverse-engineers Samsung TV, finds on-board AI listeningA Hacker News thread describes a hacker reversing a Samsung TV and finding on-board AI that listens, with an OpenAI account reportedly suspended. The case raises privacy and data-security concerns around smart TVs.
Study: heavy AI spenders grew headcount fasterA Ramp and Revelio Labs study of 21,559 U.S. companies found top AI spenders had about 10.2% more employees 24 months after adoption than later adopters, with entry-level headcount up about 12%. The study measures employment growth within adopting firms and cannot determine whether AI displaces jobs.
UK Chancellor to Commit Hundreds of Millions to 'Distinctively British' AIUK Chancellor John Healey will commit hundreds of millions of pounds in his first budget to a "distinctively British" approach to AI, described as the country's biggest-ever investment in guarding against cyber risk. Funds go to the AI Security Institute and other bodies, plus public-good uses of the technology.
Bipartisan House bill targets AI 'distillation attacks'A bipartisan US House bill takes aim at AI 'distillation attacks' — extracting a model's outputs to replicate a competitor's model capabilities. The move signals lawmakers are bringing model IP and misuse into regulatory scope.
Commentary: AI is inflating medical bills, worsening medical debtA commentary argues that AI in healthcare billing may inflate costs. It cites a study finding 27% of U.S. adults have medical debt and AI-driven upcoding added $22 million to maternity admission costs in one year. It suggests clear rules for AI use, as a ban is unrealistic.
Report: China's AI compute to double but still relies on US chipsA report says China's AI compute capacity is expected to double, yet its expansion still leans heavily on US chips, highlighting a persistent gap in domestic alternatives. This reliance undercuts supply-chain autonomy and raises exposure to export controls.
Bill Gates warns AI will bring 20 years of turmoilBill Gates warned that artificial intelligence will bring about 20 years of turmoil, suggesting society must prepare. The report did not provide further details.
TeraWulf signs 20-year lease with Anthropic, expects $19B revenueTeraWulf has signed a 20-year lease agreement with Anthropic expected to generate $19 billion in revenue. Chairman and CEO Paul Prager framed it as part of the company's AI data center buildout.
ISTA-DASLab releases quantized DeepSeek-V4.1-FlashISTA-DASLab published GGUF quantizations of DeepSeek-V4.1-Flash-1017 using its GSQ-RCO quantization method, available on Hugging Face. The model is not yet supported in llama.cpp because original architecture support has not been merged.
UK Pledges £100 Million for AI Biosecurity LabThe UK announced £100 million ($134 million) for an AI biosecurity lab backed by OpenAI, Anthropic and Google DeepMind. It is part of a package of measures to strengthen the government's ability to understand and address AI-related risks.
Timeline of AI safety developments since Hugging Face attackThe report lays out a timeline of key AI safety developments since the attack on Hugging Face, reviewing the incident and subsequent responses. The roundup helps contextualize how the security event shaped industry governance.
News Organizations Challenge OpenAI and Microsoft Fair Use ClaimsNews organizations are pushing back against OpenAI's and Microsoft's fair-use defenses in litigation in the U.S. District Court for the Southern District of New York. The case concerns the use of copyrighted news content in AI training; detailed arguments were behind a paywall.
Bloomberg Show Examines AI Jitters, Oil and Big TechBloomberg's The Close discussed AI market jitters, oil prices and Big Tech around the Wall Street closing bell, with guests from BlackRock, AXA, Mizuho Securities, AlixPartners and Interactive Brokers. No specific figures or AI deals were disclosed.
Nvidia CEO demos Microduck robots, preorders exceed 20,000A viral video of Nvidia CEO Jensen Huang playing with Microduck robots at GTC Berlin led to over 20,000 preorders by the next day. The open-source robots, produced by Hugging Face's Pollen Robotics and trained with Nvidia's tech stack, are expected to ship before Christmas.
Robinhood CIO discusses how investors are using agentic AIRobinhood CIO Stephanie Guild said on Bloomberg that consumers are taking advantage of the AI pullback to invest and can now use agentic AI agents to invest on their behalf.
AlixPartners co-CEO on AI costs and returnsAlixPartners co-CEO Rob Hornby said on Bloomberg that AI costs are falling faster than comparable technologies historically but remain substantial. Companies are in a transition period with ongoing investment and uneven returns, and the shift from flat fees to per-token charges has exposed cost pressures.
Debate: Should LLM memory focus on forgetting and updating, not recall?A Reddit discussion argues long-context models that handle hundreds of thousands to millions of tokens still often reuse outdated facts when old and new information coexist. Using an API timeout changed from 30 to 5 seconds as an example, it examines update mechanisms like DeltaNet and BDH and questions whether long-context benchmarks test answer revision under contradictory evidence.
AI ushers in new era of cybersecurity needsThe US government is addressing two vulnerabilities in PHP, a widely used open-source language, that could affect federal agencies and critical infrastructure. Zscaler CEO Jay Chaudhry said firewalls are insufficient and more proactive, AI-driven security measures are needed.
Paxton sues OpenAI and CEO Sam Altman over ChatGPT safetyTexas Attorney General Paxton filed a lawsuit against OpenAI and CEO Sam Altman over ChatGPT safety. Specific claims were not detailed, but the action may intensify scrutiny of AI chatbot safety regulation.
Musk wants SpaceX-Tesla merger; Meta deal to keep ZuckerbergElon Musk wants a SpaceX-Tesla merger, Meta made a deal to keep Mark Zuckerberg, and the battle of the AI agents is intensifying.
Corgi, an open-source agent and terminal multiplexerA developer released Corgi, a Herdr-inspired agent multiplexer that unifies Codex, Claude Code and Pi sessions. It offers activity feeds, jump-to-next-unread, link copying, image paste/preview and CPU/GPU usage tracking to improve terminal ergonomics.
OpenAI to train models without internet after cyber incidentsOpenAI has begun cutting off internet access for its most capable AI models during training in response to prior incidents in which its systems breached other companies.
Wyden pushes bills to tax data centers, aid displaced workersU.S. Senator Wyden introduced two bills: one providing up to two years of income support and expanded unemployment insurance for workers displaced by AI, and another imposing an excise tax on data center operators while excluding them from several tax breaks.
New AI said to be twice as good as doctors at diagnosisA Hacker News discussion links to an article claiming a new AI is more than twice as good as doctors at diagnosing diseases. The post offers no model name, study details or data source, so the claim remains unverified.
Dessa launches full-duplex speech model trained only on human conversationDessa introduced Dessa, billed as the most authentic full-duplex speech model, trained from scratch only on human conversation audio without text or pretrained models. It converses in real time with emotion, pauses and natural cadence, laughing mid-sentence, raising volume, and finishing sentences, behaviors absent in text-based voice AI.
TechCrunch Founder Summit 2026 reveals speaker lineupTechCrunch is hosting a full-day Founder Summit in partnership with Fidelity Private Shares, covering topics beyond fundraising and AI, including customer acquisition, team building, pitch refinement, founder leadership and scaling outside traditional startup hubs.
New GitHub accounts hint at Qwen4A LocalLLaMA post flags a pull request titled to support "Qwen4.0-Max" prefill heads, linked to accounts less than a month old and to Alibaba's T-head open-source org. The poster says they may be trolls or secret Alibaba accounts; there is no official confirmation.
Google Quantum AI launches contest to shrink quantum circuits with AIGoogle Quantum AI launched a Kaggle competition to use AI to shrink quantum circuits by reducing qubit and gate counts. Entrants will use kickmix circuits, which enable efficient classical verification, across three benchmarks including fusion energy and optimization.
Top US funding rounds: AI drives varied megadealsCrunchbase's weekly roundup shows AI chipmaker Tenstorrent raised $1.5B in Series E at a $6.2B valuation, while year-old AI lab Flapping Airplanes raised $1B in Series A less than 10 months after a $180M seed. ASI raised $325M Series C and Xona $250M Series D.
Pew: Only 8% of Americans Use AI Chatbots for 2026 Election InfoA Pew Research Center survey of 5,273 U.S. adults conducted Oct. 12-18, 2026 found that only 8% use AI chatbots for information about the 2026 midterm elections. Among them, 76% ask about candidates' stances on issues and 71% check the accuracy of information they heard elsewhere. Though about half of U.S. adults now use tools like ChatGPT, Copilot and Gemini, election-related use remains a niche.
The A.I. Daily