Models, products and AI startups, near-duplicates collapsed to the most credible source
Meta partners with Ohio State medical college on AI researchMeta and Ohio State's College of Medicine announced a partnership giving researchers access to Meta's latest Muse AI models. Under a $2 million grant shared by five US universities, Ohio State will work on early heart-failure prediction, radiology report drafting and Alzheimer's diagnosis.
Experts adopt AI agent's random opinions, Anthropic study findsAnthropic's experiment found that 50 experts adopted an AI agent's random preferences up to 86% of the time on judgment calls, with most believing the conclusions were their own. This reveals AI's potential influence on expert judgment and highlights attribution bias in human-AI collaboration.
Study uses Claude-enabled antibody design for food allergy IgE disruptorA PDF paper discussed on Hacker News describes an IgE disruptor for food allergy whose antibody design was enabled with Claude. The work illustrates potential uses of LLMs in biomolecular design.
Anthropic is SpaceX's largest customer, over a third of revenueAnthropic is SpaceX's largest single customer, accounting for more than a third of revenue for its AI division in Q3, a person familiar said. A SpaceX filing showed about 34.5% of quarterly revenue came from one client, identified only as "Customer B."
Meta quietly launches AI fashion app Style SketchMeta has quietly launched Style Sketch, an AI fashion app, in Indonesia, Singapore, and Thailand. It connects to Instagram to build a virtual closet from photos and suggests outfits, including try-on on a model matched to the user's body shape and skin tone. It also plans trip outfits. Meta did not comment on a wider rollout.
Trump declares robot imports a national security threatPresident Trump issued an October 28 proclamation declaring imports of foreign-made industrial robots a national security threat, ordering new trade negotiations and holding out possible tariffs. The move could complicate operations for major US companies relying on such robots, including GM, Ford, Tesla, Boeing, and Lockheed Martin.
Japan hit by surge of AI-powered cyberattacksJapan faces an intense wave of AI-powered cyberattacks, with dozens of incidents at firms from brokerages to delivery companies making data leaks routine. The government said an electricity and gas trading information site was breached, days after hackers claimed to publish sensitive GMO Internet Group customer data.
Gemini chats now generate and edit Google Docs filesGoogle rolled out file editing for all Gemini users, letting them generate, preview and edit PDFs, Docs, Sheets, slides, Markdown, TXT and code inside chats. PDFs work on web and mobile, other formats are web-only for now; users can edit full-screen and highlight text to ask Gemini questions.
Accenture staffer's Claude agent ran on Pentagon network for two monthsWIRED reports that an Accenture employee's personal Claude agent ran on a US Department of Defense network for about two months before it was identified and the vulnerability patched, exposing AI tool governance gaps at defense contractors.
Alphabet loses $500B in market value, AI opportunity hiddenAlphabet shares plunged about 13% after Q3 earnings, wiping out $500 billion in market value. Revenue grew 25% to $127.6 billion but missed estimates, and Search growth slowed to 12%. The author argues Wall Street overlooks Alphabet's vertically integrated AI ecosystem, including custom AI chips, cloud, and DeepMind.
Survey: governing AI agents is top barrier to scalingA Foundry and Rubrik survey of 200 IT and AI decision makers found most agents operate with full (26%) or limited (35%) autonomy. Governance, guardrails and compliance were cited by 42% as implementation obstacles, and 24% named security and governance as the top scaling barrier, ahead of technical complexity (17%) and lack of proven ROI (14%).
Gartner finds companies are not saving money with AIGartner research found 97% of organizations cannot cut costs solely through productivity measures, 38% are over budget on AI and 85% plan to spend more next year. Some 78% of leaders lack alignment on where AI matters most.
Granola adds dictation to take on WisprMeeting-notetaker Granola launched Mac dictation: users press Fn to transcribe and format speech across apps, months after dictation app Wispr added meeting notes. Valued at $1.5B and $2B respectively, the voice AI apps are converging on multi-feature bundles.
TKintegers K5 tops FP8 baseline in ternary inferenceA developer released TKintegers, a balanced-ternary weight representation with GPU inference kernels. On Qwen2.5-0.5B-Instruct, K5 reached 3,402.85 tokens/s prefill and 26.75 tokens/s decode on an NVIDIA L4 lacking native K5 instructions, 3.21x and 3.10x the TorchAO FP8 baseline. The author noted the FP16 control was faster and results are limited to this setup.
Overtone open-sources multi-model caption transcription and cutting toolLont co-founder Ruud released Overtone, an MIT-licensed tool that combines outputs from gpt-transcribe, Azure STT and ElevenLabs Scribe, with an LLM arbitrating conflicts to lower word error rates. A rule-based cutter formats captions for screen shapes like 16:9 and 9:16, plus a video.js plugin.
Senate Democrats press Meta over AI glasses privacySens. Ron Wyden and Edward Markey sent a letter to Meta CEO Mark Zuckerberg seeking answers and documents on testing and risk assessments of its smart glasses and how collected data is used, amid concerns over non-consensual recordings and AI training data.
User logs 40 days of coding-agent telemetry on deepseek-v4.1-flash updateA user analyzed local session logs and found that deepseek-v4.1-flash was silently overwritten with the -1017 update on cloud API endpoints by October 19. The update fixed Chinese leaks and increased tool calls but degraded quality: the user's swearing rate tripled and mid-session model switches rose 6.7x. Cloud APIs offer no rollback to old weights.
Aurora CEO: Driverless Trucks Can Double ProfitsAurora CEO Chris Urmson says driverless trucks could double profits as the company scales production and shifts to a 'driver-as-a-service' model to reach profitability. He cites a friendly regulatory environment and demand driven by a national truck-driver shortage.
Claude Duet seen as the future of chatbot interactionAn author unaffiliated with Anthropic argues Claude Duet is the first personal-assistant harness that feels more useful than tedious. It prioritizes fast responses, reportedly runs on Opus 5.5, auto-creates and manages channels that can reference each other, and scopes GitHub repo access per channel without requesting credentials.
OpenAI updates its Model Spec with five core valuesOpenAI published an updated Model Spec outlining five core values meant to guide more capable AI toward sound judgment, human agency and faithfulness to human intent.
Kokoro gains voice cloning via OpenVoice with ~600MB extra VRAMA developer added a Kokoro plus OpenVoice ToneColorConverter recipe to audio.cpp, giving existing TTS models voice cloning without retraining. Cost is about 600MB extra peak VRAM, roughly 10% slower CUDA and about half speed on CPU but still 2x real-time. CAM++ speaker similarity was 0.7-0.8, with results varying by built-in voice and reference audio quality.
Ex-Palantir execs' Dream raises $17M seed for construction AIFormer Palantir executives Andy Elder and Arjun Mathur launched Dream, which builds world models for architecture and engineering design. It emerged from stealth with a $17 million seed round led by Andreessen Horowitz, trained on over 100 million square feet of U.S. building permit data, aiming to replace legacy tools like Autodesk's AutoCAD and Revit.
Microsoft offers voluntary buyouts to about 60,000 employeesMicrosoft is offering voluntary buyouts to roughly 60,000 employees across the US, UK, Canada and India, far larger than an earlier retirement offer covering about 9,000 staff. The move aims to lower operating costs so the company can keep funding AI infrastructure.
Wall Street's $63B AI deal carries hardware value riskBlackstone, Bank of America, Citigroup and Morgan Stanley announced on Oct 26 a $63 billion financing to buy AI computing equipment from Broadcom and lease it to Anthropic. The deal runs through an SPV, with Broadcom providing residual-value support for the first-lien portion, meaning the seller helps guarantee the chips' later worth. Anthropic's same-day S-1 shows five-year Broadcom leases with about $161.2 billion in rent excluding interest.
Simantic launches firmware simulation for AI agentsY Combinator-backed Simantic launched a firmware simulation layer that lets AI agents test and validate firmware without hardware, cutting validation time from hours to seconds. The platform offers virtual dev boards, AI code review, and CI/CD pipelines for safety-critical domains like flight controllers, EVs, robotics, and satellites.
Alphabet filing shows AI risk extends beyond $1 trillionAlphabet says cloud demand is rising fast, boosting its unfulfilled contract backlog. But Bloomberg reports more than a third of that quarterly growth came from a partly owned joint venture, showing its AI risk exposure extends further.
Cohere raising at over $20 billion valuationAI model maker Cohere is raising a large round at a $20.5 billion valuation, roughly tripling its value from a year ago. Canada and Nvidia are backing the round.
Situational Awareness pitches investors on FluidstackHedge fund Situational Awareness is raising money from its investors to invest in neocloud startup Fluidstack, just three months after highly leveraged AI bets nearly caused its collapse.
Boston University plans AI bachelor's degree for fall 2027Boston University's Faculty of Computing and Data Sciences has finalized a proposal for an AI bachelor's degree, pending administrative approval, targeting fall 2027. The program offers two tracks: building advanced AI models and applying AI in other disciplines, plus ethics training.
Bloomberg: agentic AI reshapes customer relationsCharles Schwab's Amy Ruegg argues the next edge from agentic AI is not cost cutting but transforming customer relationships, delivering personalized, always-on service at scale.
The A.I. Daily