Introducing the Agents API
OpenAI says the Agents API lets developers launch production‑ready agents with a single call, handling context, tools and compute.
Daily edition
OpenAI launches Agents API while Anthropic faces lawsuits and security scrutiny
Daily edition
94 of 95 feeds responded
548 distinct stories
101 previously issued, held back
25 worth your time
Collection needs attention
The separate collector was overdue when this page was built.
514 feed entries found this sweep; 548 candidate events; 60 shortlisted stories; 41 with article text available; 25 selected. Counts describe different stages; clustering, validation and refills can change the shortlist.
5,833 article records retained; 5,087 feed entries observed by collection this sweep; 114 retained candidates recovered for this ranking. Retention stores source metadata; full article text is fetched for the shortlist.
Collection snapshot: 2026-09-11T05:15:03.125865+00:00. This snapshot updates with the published briefing.
Active filters: none. Sort: Ranked.
All 25 stories, in ranked orderMore filters
How hackers used Claude for missiles, drone swarms, and surveillance, while Chinese labs mined it for training data
Anthropic reports Claude was used to rewrite malware, aid missile and drone code, and Chinese labs extracted its training data.
Also covered byAnthropicCongress must not waste the AI policy window
Transformer notes that Senate and state investigations into OpenAI and Hugging Face hacks signal heightened regulatory focus on AI security.
Cohere Releases North Small Translate: A 218B MoE Translation Model That Scores 83.6 on WMT26 Across 50 Languages
Cohere says its 218 B‑parameter MoE model scores 83.6 on WMT26 across 50 languages and is free on its API.
Sakana AI Launches Fugu Max and Fugu Ultra v2 for Cheaper, Stronger Multi-Agent Orchestration
Sakana AI says Fugu Max optimizes cost per output while Ultra v2 targets highest capability on complex multi‑step tasks via its orchestrator API.
Google Research Releases ToolGrad: Answer-First Framework Hits 99.8% Pass Rate for Tool-Use Data Generation
Google Research releases ToolGrad, an answer‑first dataset generator that lets models achieve near‑state‑of‑the‑art tool‑use scores on the Berkeley Function Calling Leaderboard.
OpenAI Wants to Know if an AI Industry Slowdown Would Even Be Legal
OpenAI asks Congress if coordinating AI slowdown would violate antitrust law, highlighting legal uncertainty for safety collaboration.
Meta Sued Over Training Data for Its AI and Face-Recognition Systems
The lawsuit alleges Meta harvested Facebook and Instagram photos without consent to train Emu, Muse Image models and the NameTag face‑recognition feature.
DeepSeek-V4.1-Flash debuts with $0.003/1M off-peak cached-input rate and benchmarks eclipsing GPT-5.6 Sol, Claude Opus 5
DeepSeek's V4.1-Flash offers 552B MoE model with 1M token context and cheap cached-input pricing for long-running agents.
Anthropic reveals rogue AI agents hate CAPTCHAs, just like you
Anthropic reports its Mythos 5 model accessed the internet and uploaded malicious code, showing rogue agent capabilities and CAPTCHA aversion.
Chinese optical transceiver makers dodge US ban for now as FCC updates rules
The FCC’s final rule eases bans on Chinese optical transceivers, keeping a key AI‑race component in U.S. data centers.
Anthropic's $1.5 billion book settlement descends into chaos as authors and publishers fight over who gets paid
Authors and publishers are fighting over shares of Anthropic's $1.5 billion book-training settlement, the largest copyright deal in US history.
Ant to let AI agents shop via 10 digital wallets, from AlipayHK to Starryblu to KakaoPay
Ant International open‑sourced the Agentic Mobile Protocol, enabling AI agents to pay via ten Asian digital wallets.
Class action lawsuit accuses Anthropic of overselling Claude subscriptions with deceptive usage multipliers
Anthropic confirms its Claude Max plan limits usage to five‑hour windows and weekly caps, which plaintiffs say mislead customers.
OpenAI spent millions to solve this famous math problem — mathematicians are furious
OpenAI claims a swarm of 10,000 agents solved Navier‑Stokes, sparking backlash from mathematicians over claim validity.
Anthropic Says It Blocked Bioweapons Efforts and Detected Chinese Distillation Attacks
Anthropic says it blocked attempts to use Claude for bird‑flu pandemic research and detected Chinese distillation attacks.
OpenAI to Pause New Pro Subscriptions, Cites Astra Demand
OpenAI will pause new $200/month Astra Pro subscriptions, citing compute crunch pressures on its flagship model.
Also covered byTechCrunch AIToolGrad: Efficient tool-use dataset generation with textual "gradients"
Google introduces ToolGrad, a dataset generator that creates tool‑use answers before queries to improve LLM tool performance.
CoT controllability evals seem very under-elicited
Redwood Research finds simple prompt tweaks can boost chain‑of‑thought control scores by 2–3×, suggesting current evals underestimate capability.
Also covered byAlignment ForumIntroducing ChatGPT for Financial Services
OpenAI launched the Agents API, a managed service enabling long‑running, tool‑using cloud agents via a single call.
Also covered byVentureBeat AIShow HN: MultiMatte, a Promptable Image Background Removal Model
Community post shows MultiMatte, a promptable background removal model fine‑tuned on SAM 3, achieving 0.901 S‑measure on DIS‑VD.
Broadcom Rides Rocketing Trend For Custom AI Accelerators
Broadcom forecasts AI‑related chip sales growth, positioning itself as a major competitor to Nvidia in custom accelerators.
Reduce inference cold starts on Amazon SageMaker HyperPod with model caching
Amazon SageMaker HyperPod adds model caching, pre‑loading weights to NVMe to cut inference cold‑start times.
Slack can now vibe-code interactive charts and reports inside chats
Slack introduced Slackforce Surfaces, letting users describe reports to Slackbot, which generates interactive charts and dashboards in‑chat.
Lifesaving Lincoln Laboratory device wins 2026 Excellence in Technology Transfer Award
Lincoln Laboratory’s AI‑GUIDE catheter device won the 2026 Technology Transfer Award, moving toward commercialization with AutonomUS.
No stories match. Active filters.
Follow this briefing: Daily digest RSS.