Open-Source AI
Open models — Llama, Mistral, DeepSeek, Qwen and others — are catching up to closed labs and let you run AI on your own hardware, vendor-free. That shifts the balance of power, the safety debate and the economics of the field. Here is everything on open-source AI: open-weight releases, licenses, local deployment and the race against closed models.
Computing Power Over Checks: Why Sequoia Is Distributing Hardware to AI Developers
Sequoia partner Alfred Lin gifted 200 custom Mac Minis to attendees of a closed event. We explore how venture capital is shifting strategy when hardware becomes…
Why DeepSeek is Delaying V4 Release: Forced Transition to Chinese Chips
The main intrigue of China's AI market revealed: the delay in the anticipated DeepSeek V4 model release is driven by a large-scale migration to domestic hardware.
How OpenAI Radically Speeds Up AI Agents Through WebSockets
OpenAI updated the Responses API by implementing WebSockets and session caching. This radically reduces latency and opens the path to creating truly fast and…
The 12 best LLMs in 2026: comparing Claude, ChatGPT, Gemini, DeepSeek, and Grok
Moonshot AI releases Kimi K2.6: an agentic model with a swarm of 300 sub-agents
Bezos's AI lab seeks to raise $10 billion and challenge xAI
Mistral AI shifts to full stack: focus on custom enterprise models
Raft’s AI КОМП-АС framework: how to avoid mistakes when choosing the architecture of an AI solution
The Raft team breaks down section ‘A’ of the AI КОМП-АС framework — an architecture and product design methodology that reduces the risk of choosing the wrong…
Qwen 3.6-35B-A3B in practice: multimodality, MoE, and RAG in a single pipeline
A detailed breakdown of implementing Qwen 3.6-35B-A3B, from model loading to RAG, tool calling, and session persistence in practical workflows.
OpenAI launched Chronicle in Codex — a Windows Recall equivalent for developers
SpeShu.AI launches feature crowdsourcing: Habr users decide what goes into the next patch
Claude Code barred to minors, Qwen Code goes paid, and Opus 4.7 uses 30% more tokens
Amazon SageMaker and DVC: end-to-end traceability of ML models from data to prediction
Multiverse Computing launched a public API with compressed OpenAI, Meta and DeepSeek models
Spanish startup Multiverse Computing launched a public API with compressed versions of models from OpenAI, Meta, DeepSeek and Mistral AI for local deployment…
Cursor admits: its new coding model is built on Chinese Kimi from Moonshot AI
Popular AI code editor Cursor used the Chinese model Kimi as the foundation for its new coding engine — and that sparked a wave of questions.