AI news for builders and product teamsUpdated Nov 18, 2026, 11:00 UTC
AI news, straight from the source.
Releases, research, and ideas for developers and product teams. Updated every hour.
Across the feeds
Newest first
Researchers with the Institute for Law & AI propose "radical optionality" for AI governance, urging governments to build institutions and legal authorities now while avoiding overregulation. Separately, a Meta and KAIST paper explores neural computers, and economists model how automating AI research could produce explosive economic growth.

OpenAI and Microsoft resolved legal issues tied to a $50 billion Amazon deal, while Elon Musk's lawsuit against Sam Altman began its first week. DeepSeek also previewed a new AI model it says narrows the gap with frontier models.

Import AI editor Jack Clark writes that there is a 60%+ chance that no-human-involved AI R&D, where a system could autonomously build its own successor, happens by the end of 2028, citing benchmark trends in coding, reproducibility, ML engineering, and kernel design.

Last Week in AI's episode 243, recorded April 29, 2026, covers OpenAI's GPT-5.5 release, xAI's Grok Voice Think Fast 1.0, DeepSeek's open-sourced V4, and business, legal, and safety updates including Google's planned investment in Anthropic and China blocking Meta's Manus acquisition.

Last Week in AI's 242nd podcast episode, recorded April 22, 2026, covers OpenAI's new ChatGPT Images 2.0 model, Alibaba's API-only Qwen 3.6 Max Preview, open releases from Moonshot AI and MiniMax, and business news including a reported SpaceX-Cursor deal and Cerebras' IPO filing.

Mistral AI has released Workflows in public preview, an orchestration layer for enterprise AI built on Temporal's durable execution engine. It offers durable execution, observability, human-in-the-loop approvals, and role-based access control within Mistral's Studio, with a Python SDK v3.0 available.

Ethan Mollick reports early access to OpenAI's GPT-5.5 family, including a GPT-5.5 Pro tier available only on the website, which solved a 3D harbor-town coding task in 20 minutes versus 33 for GPT-5.4 Pro.

Anthropic's Claude Cowork, launched in January, now includes Dispatch, which lets users message Claude from a phone while it works on their desktop; the author argues such specialized interfaces matter more than raw model capability.

Mistral AI published an engineering blog post describing Spaces, its internal CLI for scaffolding projects, running dev environments, and deploying to staging. The post explains how the tool was adapted for use by coding agents, through flags, explicit state, introspectable plugins, and generated context.json and AGENTS.md files.

Ollama has released a preview build powered by Apple's MLX framework on Apple Silicon, citing large speedups on all Apple Silicon devices and support for NVIDIA's NVFP4 format. The preview, Ollama 0.19, accelerates the Qwen3.5-35B-A3B model and requires a Mac with more than 32GB of unified memory.

Mistral AI released Voxtral TTS, a 4B-parameter open-weights text-to-speech model supporting 9 languages with low latency and voice adaptation. It is available via API at $0.016 per 1k characters and in Mistral Studio.

Mistral AI announced Forge, a system for enterprises to build AI models trained on their proprietary knowledge. The company said it has partnered with organizations including ASML, DSO National Laboratories Singapore, Ericsson, the European Space Agency, HTX Singapore and Reply.

Mistral AI announced Mistral Small 4, an Apache 2.0 licensed model that unifies its Magistral reasoning, Pixtral multimodal, and Devstral agentic coding capabilities into a single model. It has 119B total parameters with 6B active per token, a 256k context window, and configurable reasoning effort.

Mistral AI announced it is a founding member of the NVIDIA Nemotron Coalition and plans to co-develop frontier open-source AI models with NVIDIA. Mistral also released Mistral Small 4 and said the coalition's first initiative is a base model trained on NVIDIA DGX Cloud for the upcoming Nemotron 4 family.

Mistral AI released Leanstral, an open-source code agent for the Lean 4 proof assistant, with weights under an Apache 2.0 license, a free API endpoint, and integration into Mistral Vibe. Mistral reports Leanstral-120B-A6B outperforms much larger open-source models on its new FLTEval benchmark at lower cost.

Mistral AI's Applied AI Proto team built an autonomous agent on top of Vibe, Mistral's open-source coding assistant, that generates and improves RSpec tests for Rails codebases and runs in CI/CD without human intervention. In an experiment across 275 source files, the agent reached 100% test pass rate and 100% average line coverage, with an LLM-as-a-judge score of 0.74.

Ollama 0.17 can install and configure OpenClaw, a personal AI assistant, with the single command ollama launch openclaw --model kimi-k2.5:cloud. Setup requires Ollama 0.17 or later, Node.js, and a Mac or Linux system, with Windows via WSL.

An essay by The Gradient argues that rational humans and rational AIs should not be goal-directed, proposing 'eudaimonic rationality' based on practices instead. It claims this framework could help align AI with safety properties like transparency, helpfulness, harmlessness, and corrigibility.

Ollama announced support for subagents and built-in web search in Claude Code, requiring no MCP servers or API keys. Subagents run tasks in parallel, and web search is integrated into Ollama's Anthropic compatibility layer, working with any model on Ollama's cloud.

Mistral AI released Voxtral Transcribe 2, two speech-to-text models: Voxtral Mini Transcribe V2 for batch transcription and Voxtral Realtime for live applications, which ships under the Apache 2.0 license. An audio playground in Mistral Studio was also launched.