Introducing Gemini 3.8 Live and 3.8 Live Extended Thinking
Google DeepMind says Gemini 3.8 Live adds real‑time reasoning and visual grounding for scalable voice agents.
Daily edition
Google launches Gemini 3.8 Live, Apple revamps Siri, and Salesforce unveils Koa model
Daily edition
93 of 95 feeds responded
393 distinct stories
88 previously issued, held back
25 worth your time
Collection needs attention
The separate collector was overdue when this page was built.
357 feed entries found this sweep; 393 candidate events; 59 shortlisted stories; 46 with article text available; 25 selected. Counts describe different stages; clustering, validation and refills can change the shortlist.
7,446 article records retained; 5,265 feed entries observed by collection this sweep; 70 retained candidates recovered for this ranking. Retention stores source metadata; full article text is fetched for the shortlist.
Collection snapshot: 2026-09-15T05:00:03.060846+00:00. This snapshot updates with the published briefing.
Active filters: none. Sort: Ranked.
All 25 stories, in ranked orderMore filters
Building AI to accelerate science and improve lives
Google AI reports its tools now support over 300 languages for 7 billion people, plus a 50% more accurate WeatherNext 3 precipitation model.
Judge Pitts sets scheduling order for Bartz-class opt outs in Cambronne v. Anthropic. Trial in 2028.
Judge Pitts set a scheduling order, pushing the Bartz‑class opt‑out trial to 2028 with motions due May 13 2027 and Aug 26 2027.
OpenAI Said to Buy Startup Glass Imaging for $300 Million
OpenAI says the $300 million deal brings higher‑quality smartphone camera tech into its portfolio.
Also covered byTechCrunch AIApple brings a fully revamped Siri built on Google's Gemini, but not to the EU
Apple says its rebuilt Siri AI runs on Google’s Gemini, combining on‑device and private‑cloud compute, but it won’t launch in the EU.
AI for everyone in every language
Google AI highlights that its technologies now operate in more than 300 languages, reaching 86% of the global population.
China issues first AI-plus-brain-computer interface medical-device standard
China’s NMPA approved YY/T 2029‑2026, setting EEG dataset and AI algorithm quality standards for brain‑computer‑interface devices.
Introducing System One Models and Jev
TypeSafe AI claims its System One model enables fast, structured decisions for software automation, built on a new stack.
OpenAI has hundreds of contract workers reading your ChatGPT conversations
404 Media reports that contract reviewers read real ChatGPT conversations, risking exposure of personal data even with anonymization.
China elevates HBM and HBF in five-year plan, opening a new front in AI memory
China's MIIT and NDRC included HBM and HBF upgrades in the 15th Five‑Year Plan, signaling a new front in AI memory development.
Judge Martinez-Olguín approves Bartz Class Counsel plan to distribute payments to class by Nov. 15
Judge Martinez‑Olguín approved the Bartz class counsel plan to issue first payments of about $2,203.56 by Nov 15.
Amodei's AI slowdown plan never says open weights. It doesn't have to.
Anthropic proposes mandatory embedded evaluators, a narrow antitrust waiver, and a future treaty to pace AI development.
MLCommons Agent Reliability Profile Named a Finalist in Global Agentic Regulator Hackathon
MLCommons reports its Agent Reliability Profile, addressing the authorization and supervision gap, was selected as a finalist in the Agentic Regulator hackathon.
AEF-1 standard emerges for Third Party Evaluators, as Xai, OpenAI, and Anthropic all cosign
OpenAI, Anthropic and XAI co‑signed the AEF‑1 standard, committing embedded third‑party evaluators to verify safety practices.
What execs and politicians are saying about slowing down AI development
Highlights growing policy and industry pressure to pace AI progress, influencing strategic planning.
Salesforce and Nvidia’s new reasoning model is everything the AI labs should fear
TechCrunch says Salesforce’s Koa, built on Nvidia’s open‑weight Nemotron, is trained for sales, marketing and support tasks.
OpenAI’s ‘Top Priority’ for AI Agents is Automating AI Research, Says Noam Brown
OpenAI says GPT‑6 Astra’s main goal is to automate AI research, boosting its ability to design and test new models.
A single firm is behind OpenAI, Anthropic, and Meta hacking scandals
Raises concerns about supply‑chain security of models from leading labs, prompting risk assessments.
What's the best AI model for graphic design, video ads, lip sync and more? OpenArt's new Arena offers leaderboards for different media jobs
VentureBeat reports OpenArt’s Arena benchmark ranks models per media job, using blind pairwise evaluations from creators.
What’s at stake in AI’s trillion-dollar gamble
MIT’s Wachter estimates hyperscaler AI spending will hit $1.1 trillion by 2027, requiring a 2.7× productivity boost to break even by 2030.
OpenAI stuck fighting Musk antitrust suit after Apple finds a way out
Musk’s court filing confirms he settled the antitrust suit against Apple over ChatGPT integration, while his fight with OpenAI continues.
Exclusive: Paying for frontier AI models buys 4-month head start at 5x the cost
Ars Technica cites Mozilla’s report that open‑weight Chinese models now trail US frontier models by only 4.4 months, costing 30% as much.
DeepSeek Open-Sources Harness Agent Runtime With Everything-Is-a-Plugin Design
Pandaily reports DeepSeek released the MIT‑licensed Harness runtime, letting models, tools and UI function as npm plugins.
Tell HN: iOS 27 does not allow Apple Intelligence to be disabled
Removes a user control, forcing developers to handle always‑on AI features or workarounds.
With iOS 27, I’m actually using Siri again
Apple’s iOS 27 revamps Siri, making it useful for complex requests and boosting everyday AI assistant interaction.
No stories match. Active filters.
Follow this briefing: Daily digest RSS.