Content Dated and sourced

News

Commentary on AI news as it happens. Each piece is dated and links to its sources. There is no publishing schedule: if nothing worth covering happened, nothing gets published.

All Funding Product Research Regulation Tools

ResearchMicrosoft publishes a 38-page draft rulebook for its own AI models and opens it for public comment2026-09-14 · The Code of Conduct sets absolute limits on what Microsoft's in-house MAI models can do, including a ban on resisting shutdown. It arrives a day after Satya Nadella said the company "welcomes" a slower pace on frontier AI.FundingOpenAI buys smartphone-camera AI startup Glass Imaging for over $300 million2026-09-14 · The Wall Street Journal reports OpenAI has acquired Glass Imaging, a computational-photography startup founded by former Apple engineers, adding another piece to its Jony Ive-led hardware push.FundingTandem Health raises $100M to grow from AI scribe into a full clinic operating system2026-09-14 · The Stockholm startup's Series B, led by the EU-backed Scaleup Europe Fund, is aimed at expanding beyond AI-generated medical notes into scheduling, triage and patient communications for European clinics.RegulationTrump dismisses the AI-pacing chorus: "the only guardrail AI needs" is him2026-09-14 · Two days after Amodei, Altman, Musk and Hassabis all backed some version of slowing frontier AI development, Trump used Truth Social to reject regulation and call the safety concerns a "SICK conspiracy."RegulationUK parliamentary committee says no country's AI law is "fit for purpose" -- calls for a dedicated AI Bill2026-09-14 · The Joint Committee on Human Rights wants a single statutory AI regulator with the power to block high-risk systems before release and pay individuals directly when AI causes harm.RegulationAnthropic, OpenAI and Google have been meeting privately since July on a joint AI standards body2026-09-13 · The Information reports working-level teams from the three labs have met regularly for two months on a self-regulatory testing body. The talks predate, and go further than, Dario Amodei's public "pace the frontier" essay.ProductBox brings enterprise files into ChatGPT -- with its own permissions layer intact2026-09-13 · Box expanded its MCP-based integration beyond Claude to ChatGPT, Microsoft 365 Copilot and Glean, letting employees search, preview and act on company content from inside whichever AI assistant they already use.RegulationCalifornia bans toys with AI companion chatbots for four years2026-09-13 · SB 867, signed September 10, prohibits manufacturing or selling any toy built around a companion chatbot through January 1, 2031 -- with no compliance path through disclosures, filters, or parental controls.FundingFortaegis raises $50M to root AI infrastructure security in silicon itself2026-09-13 · The Amsterdam startup's oversubscribed Series A, led by Serendipity Capital, will fund commercial manufacturing of chips that generate encryption keys from a device's own physical properties instead of storing them.RegulationSpeaker Johnson rejects an AI moratorium, says Congress won't lead on AI safety rules2026-09-13 · Days after Anthropic, OpenAI and xAI's CEOs called for the industry to slow down, House Speaker Mike Johnson told CNN safety is the industry's job to solve and proposed a White House sit-down.RegulationNYT, Authors Guild, OpenAI and Microsoft all move for summary judgment in consolidated copyright case2026-09-13 · Cross-motions filed September 4-5 ask Judge Sidney Stein to decide the core fair-use question in the New York-based multidistrict litigation without a trial. Opposition briefs are due in October.ProductServiceTitan doubled AI-product locations in a quarter -- and trades customers are paying for it2026-09-13 · Fiscal Q2 earnings show Max, ServiceTitan's AI layer for HVAC, plumbing and electrical contractors, on pace for over 700 enrolled locations by fiscal year-end, alongside 21% revenue growth.ToolsSourcegraph's Amp drops its subscription price to zero for bring-your-own-key users2026-09-13 · Amp now runs free on a developer's own compute and model subscriptions, with no BYOK token fees for anyone but Enterprise, plus new free Hobby and no-extra-charge Teams tiers.RegulationFive tech trade groups ask Congress to strip AI chip export bills from the NDAA2026-09-13 · ITI, CCIA, and three other industry associations are lobbying against folding the AI Overwatch Act, the Chip Security Act, and the Match Act into the FY2027 defense authorization bill.ProductYelp puts OpenAI's newest voice model to work answering the phone for local businesses2026-09-13 · Yelp Host and Hatch, which together have handled over a million calls for restaurants and service pros, are adopting GPT-Live-1 to let callers interrupt and talk over the AI mid-sentence.FundingChina's Z.AI raises $5 billion in its second multibillion-dollar raise in two months2026-09-13 · The Hong Kong-listed lab, formerly Zhipu AI, combined a discounted share placement with zero-coupon convertible bonds -- and says 60% of the proceeds go straight to model development and compute.ResearchAnthropic's CEO calls for slowing AI development, and OpenAI and xAI CEOs back it within hours2026-09-12 · Dario Amodei's essay "We Must Pace the Frontier" commits Anthropic to giving outside evaluators employee-level access to its systems; Sam Altman and Elon Musk both endorsed the idea publicly the same day.ToolsA new benchmark tests coding agents on real, private codebases -- and none crack 40%2026-09-12 · Specific Labs' Real-SWE evaluates frontier coding agents against licensed production code instead of public repositories, and even the leading combination, Claude Fable 5.1 running in Claude Code, resolves only 38.8% of tasks.FundingTAR raises $120M Series A at $1B valuation to build off-grid power for AI data centers2026-09-12 · The Austin startup builds modular, self-contained power systems so AI data centers can add generation capacity without competing with local communities for grid capacity.ResearchResearchers show a booby-trapped repo could escape Claude Code's sandbox with no prompt2026-09-11 · Accomplish disclosed "Beltdown," a chain of git config quirks that let an untrusted repository run commands outside Claude Code's macOS sandbox, and says Anthropic took roughly 50 days and 30 releases to fully fix it.ToolsClaude Code adds a formal testing framework for its own plugins and skills2026-09-11 · The new claude plugin eval command scores a plugin against realistic prompts with six grader types, compares the result to a no-plugin baseline, and can gate CI on the score.ToolsCognition pairs a frontier model with a cheap sidekick to cut Devin's coding costs2026-09-11 · Fusion runs a "lead" model alongside a lower-cost "sidekick" with separate persistent contexts, which Cognition says gets up to 39% more efficient results on coding benchmarks than running one model alone.FundingCohere in advanced talks to raise $2-3B at a $20B valuation2026-09-11 · The Canadian sovereign-AI lab is reportedly courting the Canadian and German governments alongside existing backers, positioning itself as a non-U.S., non-Chinese alternative for enterprise and government buyers.ToolsGitHub Copilot's code review now closes its own comment threads and grades harder2026-09-11 · Comments now auto-resolve when a follow-up commit addresses them, and Copilot's lower-effort review tier switched to multiple agents voting on one review -- a change GitHub says raised how often high-severity findings get fixed.