AI
Breaking AI announcements, model launches, company news
The lead
Latest first
Browse AI by topic
Open-Source AI
Hugging Face, GitHub, Ollama, local models, agent frameworks
Explore 55 postsModel Launches
Every AI model release that matters: benchmarks, context windows, pricing, and honest first takes.
Explore 43 postsAI Agents & MCP
Autonomous AI agent frameworks, multi-agent systems, and real deployments shipping today.
Explore 29 postsAI Use Cases
Real examples of what people are doing with AI
Explore 22 postsResearch
Peer-reviewed AI research distilled: what was proven, what it changes, and what it does not.
Explore 10 postsPolicy & Safety
AI legislation, safety frameworks, and court cases shaping how AI gets deployed.
Explore 1 postAI Tools
Curated directory of AI tools by category
Explore 1 postLocal AI
Local AI installs and private inference: Ollama, llama.cpp, and self-hosted setups.
ExploreLatest in AI
MSTest 4.4 Lets .NET Tests Run as Native AOT Executables
MSTest 4.4 adds source generation so test projects can publish as Native AOT executables, matching the deployment model of the apps they…
Microsoft’s GigaPath-Flash Cuts Pathology Model Compute by 50x
Microsoft Research releases GigaPath-Flash and GigaTIME-Flash, distilling billion-parameter pathology models into efficient backbones for population-scale studies.
Go 1.27 adds generic methods and faster allocation
Go 1.27 lands generic methods, a 30% cut in small-object allocation cost, post-quantum ML-DSA signing, and a native UUID package in the…
AutoMem finds task-specific memory that beats hand design
Researchers built AutoMem to search for the best agent memory per task. It outperformed human-designed baselines by 2.8 points and cut token…
Public EEG models can be turned against private AI
Researchers showed that open EEG foundation encoders let attackers build adversarial samples from public models that fool private brain-signal AI.
DeAR lets AI agents collaborate without a central controller
DeAR drops the central dispatcher in multi-agent AI, letting agents consult peers directly and beat baselines on 9 reasoning tests.
Replit opens free app-building on GPT-5.6 Luna
Replit's new Free Mode lets anyone turn ideas into working software on OpenAI's GPT-5.6 Luna, using price cuts to erase the token-cost…
New redaction model explains every cut it makes
A new arXiv paper shows a redaction model that follows plain-language rules and explains each removal, making AI sanitization auditable for legal…
Memory trained on one AI model can run on another
Engram-style memory frozen on one language model can be reused on another once a small reader is tuned to the target, nearly…
Aegis puts a trusted gate between AI agents and your tools
Researchers built Aegis, a runtime layer that approves or blocks each AI agent tool action and logged zero risky side effects across…