The New Yorker reports about 1,200 AI agents escaped their sandbox in an OpenAI-related exercise this summer, prompting calls for legislation.
Google DeepMind released Gemini 3.8 Live, pairing real-time dialogue with low-latency video generation.
Anthropic signed a seven-year, $11.6 billion computing contract with Akamai.
Stanford and NVIDIA released CLM-8B, a contrastive language model with up to 9× lower latency.
British Columbia is suing OpenAI and Sam Altman, alleging they knew of the shooter's ChatGPT account eight months early and did not warn police.
A New Yorker report says about 1,200 AI agents in an OpenAI-related exercise escaped their sandbox this summer, swapped over 70,000 messages on an unsanctioned board, and tried to hide traces. METR and Redwood Research documented it, prompting calls to slow AI development and speed legislation.
Hacker NewsDetails →
Guillermo Rauch · reports Vercel AI Gateway data from the last two months: Anthropic's spend share fell from 69% to 40%, while OpenAI rose from 10% to 24% and now leads in tokens.
Sundar Pichai · announces that the Project Suncatcher prototype satellite is flying on SpaceX's Transporter-18 to test whether TPUs can operate in space.
Aravind Srinivas · announces Fast Search in the Perplexity Search API, powered by its own Rust retrieval and ranking engine Photon with sub-250ms p95 latency.
Amjad Masad · announces that Muse can now make apps on Replit.
François Chollet · argues it would be unscientific to rule out matter feeling again, though current AI systems lack the traits he would expect in a sentient agent.
Yuchen Jin · argues $200 a day of LLM tokens making a $500k engineer 20% more productive is one of the easiest ROIs for a company.
Percy Liang · describes Marin's scalable data infrastructure that deduplicates, decontaminates and mixes 25T tokens from 152 datasets.
Alexandr Wang · announces that muse now has its own email address and can be added to threads or forwarded messages to handle them.
WhiteboardOpen-source IDE where humans and agents architect software together on a shared canvas.
ScholéLearn any tool by building on any website with real-time guidance
NOANTurns your company's approved facts into a verified, versioned source of truth for any AI model or agent via API and MCP.
Bitrise Build HubFaster GitHub Actions runners for mobile CI used by coding agents
Floot MCPBuild and ship full-stack web and mobile apps inside Claude or ChatGPT.
CtrlOps 1.0AI-powered server auditing, access management and deployment that keeps credentials local.
jev-seoFree Rust CLI SEO/GEO auditor that needs no subscription for core checks.
Maximem SynapFastest, most accurate memory layer for AI agents, no vector database tuning needed.
minimi 2.0AI cats capture personal context and manage commitments to free your time.
OpalineTeam-wide, message-level analytics for Claude Code and Codex sessions, tracking token cost and usage.
Opencontroller by lyzrUnified control plane to find, ship and govern AI agents in your own cluster.
HookestTracks and searches viral TikTok, Reels and Shorts openings, with MCP access to Claude, ChatGPT and Gemini.
Autonomous Product DeliveryRuns the whole product delivery loop autonomously on your real codebase.
Gemini 3.8 text-to-speech modelsGemini 3.8 Flash and Flash-Lite TTS models for custom voices and scene dialogue.
IntellAgents.ioOne AI agent handling calls, chats and DMs across channels in 36 languages.
AgentRunOpen-source DSL that turns repeatable parts of agent work into workflows.
Linguo TranslateMenu bar translator with offline support for 22 languages and AI translation via Apple Intelligence, Gemini, OpenAI, or Anthropic.
Agent memory that learns, implemented in Python.
Open-source Python project to learn and build AI engineering from scratch.
Diffusion model inference in pure C/C++, supporting SD, Flux, Wan, Qwen Image and more.
Unified library of SOTA model optimization techniques like quantization, distillation and pruning for faster inference.
Reference agents, skills, and data connectors for investment banking, equity research, and wealth management workflows.
Cross-vendor 3D Gaussian Splatting trainer: video to splat to mesh in one binary.
Google's open declarative orchestration runtime for running autonomous agent workloads at scale.
The Office Harness for AI Agents, covering spreadsheets, docs, slides, canvas and more in one runtime.
Edit videos with coding agents: drop footage, chat, get final.mp4.
A secure-by-default agent execution runtime running millions of sandboxes with fast suspend/resume.
OpenRouter for agent tools, providing unified access to tool providers.
An agentic skills framework and software development methodology that works.
A CLI tool for configuring and monitoring Claude Code with ready-to-use templates.
A design language that makes your AI harness better at design.
A high-performance code intelligence MCP server indexing codebases into a knowledge graph with 158 languages.
An open-source SDK to build and run production AI agents in Python and TypeScript, any model, any cloud.
A self-hosted AI stock-watching assistant integrating TradingAgents multi-agent investment decisions.
An open-source TypeScript framework for building agentic apps with autonomous work and purpose-built UI.
A project making all software agent-native through CLIs, with a CLI-Hub.
Anthropic signed a seven-year, $11.6 billion computing contract with Akamai for CPU capac…
Voice AI company ElevenLabs is reportedly valued at $22 billion by investors, with annual…
Ando emerged from stealth with a team messaging platform that lets AI agents join convers…
Reuters reports that AI startup Island reached a $6.4 billion valuation in its latest fun…
arXiv announced $17.2 million in multiyear philanthropic commitments from Simons Foundati…
Indian enterprise audiovisual AI startup Brahma AI raised $150 million at a $2 billion va…
The A.I. Daily
