2026
11,340 articles · 139 days · 10 categories · avg 82 per day
OpenAI GPT-OSS: Running the Open-Weight Models in Colab with MXFP4 and Advanced Inference
A new guide shows how to run openai/gpt-oss-20b in Google Colab: install Transformers dependencies, check the GPU, enable MXFP4 quantization and build a working…
Habr AI publishes a beginner's guide to ChatGPT, Claude, and MCP
Habr AI skips the theory to explain local vs. cloud models, why subscriptions are worth paying for, how agents work, and why MCP is becoming a standard.
Three OpenAI executives leave as Sora shuts down and side projects wind down
In a single day, OpenAI's former chief product officer, the head of Sora and the CTO of enterprise applications left, as the company cuts non-core initiatives and…
Sony AI's Ace robot beats elite table tennis players under official rules
US MATCH Act aims to curb China's access to chipmaking equipment
Intel stock hits record high after strong forecast amid AI boom
Oakland court to hear Elon Musk's suit against OpenAI and Sam Altman over the company's mission
OpenAI takes GPT-Rosalind into biology: what the new model can do and where its limits lie
OpenAI has launched GPT-Rosalind for life sciences and paired it with a Codex research module, promising faster hypotheses and experiments without stitching…
Anthropic locks down Claude Mythos over global cybersecurity risk
Anthropic is not releasing Claude Mythos publicly: the model can find and potentially exploit thousands of vulnerabilities, and reports of unauthorized access have…
Cursor security audit uncovers four weaknesses in the code editor's defenses but fails to break authorization
Anthropic unveils Claude Mythos: a zero-day-hunting AI shifts the balance of power online
Manitoba Prepares Strict Ban on Social Media and AI Chatbots for Minors
Anthropic ships Opus 4.7 as OpenAI turns Codex into a computer-using agent
Anthropic Predicts Engineering Jobs Will Fade, While Hiring 429 Developers
Anthropic's chief says AI is gradually taking over programming, yet the company has opened hundreds of engineering roles paying up to $405,000.
Anthropic releases Claude Opus 4.7 for agentic coding, vision and autonomous tasks
Anthropic has unveiled Claude Opus 4.7, an update to its flagship model with notable gains in agentic coding, high-detail image analysis and long autonomous tasks.
PrismML Bonsai: How to Run a 1-Bit Model on CUDA with GGUF, JSON and RAG
A new PrismML Bonsai tutorial shows how to run Bonsai-1.7B on a GPU via CUDA and GGUF, measure speed, set up chat, strict JSON output and a basic RAG scenario.
OpenKB and OpenRouter show how to build a local AI knowledge base with Llama-powered search
How Moscow Credit Bank Traces the Evolution of Bank Training, from Clerks to AI
OpenAI brings Sky technology into Codex for Mac and strengthens background app control
SABIC Strike Hits the AI Supply Chain: PCB Resin Costs More, Lead Times Grow
Rio Tinto and the race for copper: how the AI boom is forcing the US to scramble for more output
Data center growth and strain on power grids are turning copper into a strategic resource. The US is trying to open new mines but runs into imports, red tape and…
Google in talks with Marvell on AI inference chips, cutting reliance on Broadcom
Google is negotiating with Marvell over two specialized AI chips for inference and memory, widening its silicon partner pool beyond Broadcom and MediaTek.
Habr AI explains how data-driven solution search reshapes client-server architecture
Seedance 2.0 draws attention with hyperrealistic generation of people and scenes
Anthropic and the White House discuss Mythos AI after it finds thousands of vulnerabilities
Copenhagen's Atech raises pre-seed from Lovable, Sequoia Scout and a16z Scout
MIT researchers teach AI to honestly say "I'm not sure" and hallucinate less
MIT researchers have a training method that helps language models judge their own confidence and give fewer confident but wrong answers.
Lyubomir Gorbatko Presents Sessa, an Alternative to Transformers and Mamba
A Habr article breaks down Sessa, an architecture that tries to combine the strengths of Transformer and Mamba to handle long context without losing controllable…
MIT unveils EnergAIzer, a fast way to estimate AI energy use in data centers
Researchers at MIT and the MIT-IBM Watson AI Lab built EnergAIzer, a tool that estimates the energy use of AI workloads in seconds and helps data centers cut energy…
MIT releases MathNet, the world's largest dataset of math olympiad problems
Harvard Business Review: how heavy AI use leads to "brain fry"
Flag Soft: The "Dali Challenge" Benchmark Helped Pick an LLM on Quality, Speed and Price
