AI news for builders and product teamsUpdated Nov 17, 2026, 19:01 UTC
AI news, straight from the source.
Releases, research, and ideas for developers and product teams. Updated every hour.
Across the feeds
Newest first
Google Research and the Stanford Accelerator for Learning released "Polar Peaks," the third AI Quest, which teaches students ages 11 to 14 to map brain neurons with AI while practicing data cleaning and model testing. New teacher training modules are available in the Google AI Educator Series, with the Quest #3 module due next month.

Microsoft is offering voluntary buyouts to up to 60,000 employees in the US, UK, Canada and India, spanning teams including Copilot, CoreAI, Windows + Devices, Xbox, finance and HR, as it looks to reduce staff without another layoff round.

Apple has acknowledged that a small number of iPhone Duo foldables have inner-display failures, with white lines appearing within hours of unboxing, and is offering AppleCare replacements. Teardown specialists suspect the hinge, but the failure rate and cause are unconfirmed.

NVIDIA Mission Control 2.5 extends its autonomous recovery engine beyond Slurm to Kubernetes workloads managed by NVIDIA Run:ai, letting faulty nodes be isolated and jobs restarted without waiting for root-cause diagnosis. The release also adds beta AI network anomaly detection, central Keycloak authentication, hardening updates and early access for Vera Rubin NVL72.

InfoQ is running three online certification cohorts in November 2026: two AI engineering programs (led by Adi Polak on November 12 PST and Lan Chu on November 13 CET) and a High-Performing Teams program with Olimpiu Pop on November 17.

Anthropic filed for an IPO under the ticker "ANT," reporting $16.3 billion in first-half 2026 revenue against a $693 million operating loss, with a founder-controlled structure holding 50.1% of voting power.

Tesla says 1.7 million people now pay for its supervised Full Self-Driving software, up 67% year on year, as it shifts focus from vehicle sales toward recurring software revenue while exploring a merger with SpaceX.

Kevin Roose, author of the new book The AGI Chronicles, has left The New York Times with Casey Newton to launch Machine Gods, a video-first tech podcast produced in partnership with NPR that debuted last week.

Databricks argues that media and advertising are shifting from an attention-based model to a prediction economy.

Anthropic has publicly filed its S-1 with the SEC to list on Nasdaq under the ticker "ANT," disclosing an $80.7 billion net loss on $16.3 billion in revenue for the first half of the year.

Google's Gemini 4 Argon model card and safety card state that the unmitigated model cannot be ruled out from approaching several frontier capability thresholds, including Cyber Uplift Level 1 and CBRN Uplift Level 1, while hallucination rates fell by more than two thirds versus earlier Gemini models.

Claude Opus 5.5 now scores 18.5% resolved on ProgramBench, a benchmark where every model scored 0% when it launched in May. Separately, startup Mirendil reported one researcher pretrained a 100B–500B parameter coding model in 3.5 months for under $10m in compute.

Cloudflare released Auto Router in public beta through AI Gateway, where a classifier scores each request on complexity, ambiguity, stakes and context dependence and picks a model. Its own benchmark shows ten points lower success than Claude Opus 5.5 at roughly a third of the cost.

OpenAI published a Signals analysis of ChatGPT work usage covering 165 work activities across nearly 100 countries, examining how people use the tool for work and how that usage shifts over time.

High-frequency trading firms including Jane Street, Hudson River Trading and XTX Markets are spending billions on data centers, cloud contracts and chip startups to secure AI compute, with one executive putting spending at ten to twenty times pre-ChatGPT levels.

Toyota plans to deploy 400,000 robots as part of a $6.4 billion annual investment aimed at turning the automaker into a global robotics company, according to the Financial Times.

A new MIT Technology Review opinion piece argues that regulating only catastrophic AI harms lets companies meet narrow safety requirements while still harming young people through addictive feeds, AI companions and chatbot-assisted learning.

Fanatics is using OpenAI technology for a feature called Shop the Look, which helps fans complete game-day outfits; testing across roughly 30,000 MLB products showed 10% higher engagement.

OpenAI engineers use an internal tool called Codex Ultrafast to investigate security incidents in minutes rather than hours, according to an OpenAI post. The post also describes the tool being used for code refactoring and everyday workflows.

Grok released a seven-minute AI-generated video, "Episode 1 of our Odyssey series," made with Grok Imagine in collaboration with Wonder Studios and director Dev Walker, roughly three months after Elon Musk promised a full-length film version of The Odyssey.