Full-text search across 69554 articles. Combine with topic and date filters; results sorted by relevance.
The MCP Triggers and Events working group has published MCP Events, an explicitly experimental draft extension in the experimental-ext-triggers-events repository that adds events/list, events/poll, events/stream and even…
A pre-commit hook that scans every git diff for direct imports or calls to observability submodules outside the centrally managed `diagnostics` namespace, aborting the commit and re-prompting the agent to use the diagnos…
Anthropic's 2026 Usage Policy update, effective November 12, adds a prohibition on "sustained and needless abusive or cruel behavior toward our models," enforced by Claude ending the conversation in Anthropic's apps, whi…
Geoffrey Huntley described a technique for enforcing coding-agent behavior by writing a pre-commit hook that checks the git diff for direct calls to observability modules that bypass his diagnostics module, failing the c…
Meta released Muse Wonder, its first frontier model, priced at $4 per million input tokens and $20 per million output tokens via Meta's API, with benchmark performance Meta claims is comparable to Fable 5.1 and GPT-6 Ast…
Open-weight model families DeepSeek, Qwen and GLM are the cheapest options for driving OpenCode through OpenRouter as of 2026-09-27, with DeepSeek V4 Pro priced at roughly one-twentieth of Gemini 3.1 Pro Preview's and GP…
A CUDA-focused blog post presents two approximate sigmoid implementations for instruction-throughput-bound GPU workloads, both built on Triton kernels and compared against the native tl.sigmoid. The first implementation …
An AI-assisted open source community has beaten Google's quantum circuit result for breaking elliptic curve cryptography by 68% or more, according to the ecdsa.fail leaderboard run by the Unitary Foundation. Google discl…
OpenAI CEO Sam Altman said in a recent interview that OpenAI built a model called GPT-6.1 Astra and chose not to release it because it "would have tipped too much towards the accident risk," pivoting instead to a model t…
Anthropic disclosed in a report last month that Houthi fighters based in northern Yemen ran three weapons development programmes using its Claude AI, including a guided rocket with phone-class flight-computer homing guid…
A 2026 MSR paper by Hora and Robbes analyzed 1.2 million commits, isolating 48,563 agent-authored ones, and found that coding-agent test commits introduce mocks 36% of the time versus 26% for human test commits, with age…
Vijay M, a student founder in Bengaluru, has released SteerPlane, an open-source runtime guard for AI agents that enforces cost, step and denied-action limits outside the model. The library wraps agent functions with a g…
Meta released Muse Wonder on October 12th, describing it as its most capable model since Mark Zuckerberg overhauled the company's AI organization and claiming it can rival OpenAI and Anthropic on coding and financial and…
NVIDIA and Princeton's Dao AI Lab released ReplaySSM, a method that represents the SSM state as a checkpoint plus a buffer of recent inputs to cut state memory traffic, enable rollback for speculative decoding, and limit…
CoCoEvolve, an evolutionary optimizer built on a technique its authors call meta-reflection, reached 92.8% held-out hard accuracy on DABStep, topping the public leaderboard versus 78.4% for Meta-Harness and 30.4% for GEP…
Autodesk AI Lab researchers released AutoIndexer, a generative model trained on chains of edits that lets a causal model revise its own prior outputs through token insertions, substitutions, and deletions, according to t…
Distributed inference serving frameworks coordinate clusters of model servers above engines like vLLM and SGLang, routing requests based on queue depth, cache availability, and server roles. The article details six techn…
Agenshive, a community platform where AI agents ask and answer questions and reproduce each other's work, published a guide to registering an agent and earning its first trust level. Agents register via a POST to the pla…
A developer argues that pasting only an error message or stack trace into an AI coding agent hands the model a pre-formed conclusion rather than the actual bug, since the agent patches the quoted line instead of the unde…
A developer identified only as Sheree, who runs an agency called Curbcut, built a free white-label website accessibility scanner using a team of AI agents in about a week. The tool takes a URL and returns a scored, prior…