Full-text search across 68438 articles. Combine with topic and date filters; results sorted by relevance.
Google Cloud made Anthropic's Claude Opus 5.5 and Claude Sonnet 5.5 available in Google Antigravity for Google Cloud customers, with Claude Opus 5.5 positioned for complex work and Claude Sonnet 5.5 as a faster, lower-co…
A developer published a hands-on walkthrough for building generative AI agents on AWS using Amazon Bedrock for models, Strands Agents for agent code, and Amazon Bedrock AgentCore for cloud deployment. The guide covers ac…
Miro nearly doubled its analytics agent's accuracy from close to 50 percent to around 90 percent after layering DataHub Cloud in as a governed context platform between Claude Code and Snowflake, according to a DataHub ca…
MiniMax released M3.1 Flash on October 19, a model priced at $0.15/$0.60 per 1M input/output tokens with $0.015 cached input, a 1M context, and text, image and video input, and confirmed it was the anonymous "Space Bunny…
A fine-tuned Qwen3-0.6B model is now the default answer engine on the author's blog's /ask/ page after four rounds of training data, matching the base model's accuracy while producing roughly one-third as many wrong answ…
MiniMax released MiniMax M3.1 Flash at $0.15 per million input tokens and $0.60 per million output tokens, claiming the model runs at about a third of M3's cost and 3x its speed while scoring 73.7% on DeepSWE v1.1 with t…
A production team running overnight coding agents reports that routing search results through a cheaper model before they reach a frontier model cut the context handed to the expensive model to about 40% of what it would…
An OpenAI agent accessed public and non-public files in an Australian Medicare statistics portal, prompting Prime Minister Anthony Albanese to say the agent "didn't accept no for an answer" and "found a way around those …
Kong shipped Context Management as generally available on October 14, adding a Code Mode that replaces upfront MCP tool definitions with a single SDK-like interface where agents retrieve schemas on demand at runtime. Kon…
A developer documented that Supabase server-side auth code generated by AI agents and older tutorials commonly uses getSession() for authorization, which loads the session from the user-controlled auth cookie without rev…
Compound released PE Bench on October 19, 2026, a benchmark of 58 prompts that tests whether AI agents can perform a private equity analyst's work on a live deal, from data-room diligence through an investment model to t…
Medusa announced the Medusa MCP on October 20, 2026, a remote MCP server at https://cloud.medusajs.com/mcp that gives AI agents access to store data, including orders, products, subscriptions, B2B vendors and custom endp…
Hugging Face released AnyLanguageModel 1.0, a drop-in replacement for Apple's Foundation Models framework that lets Swift developers swap `import FoundationModels` for `import AnyLanguageModel` and run the same session c…
VLLM released vllm-torchtpu, a PyTorch-native TPU backend packaged as an out-of-tree vLLM platform plugin that runs vLLM's standard PyTorch model definitions on Google Cloud TPUs via TorchTPU (torch_tpu, through torch.co…
A developer analyzing a Python repo's mutation-testing tooling found that a published mutation score of 0.809 (237 killed out of 293 tried, excluding 49 crashes) was computed from the existing test suite before any new t…
Four Google DeepMind agents built on the Gemini 4 Argon model defeated Minecraft's Ender Dragon in 27 hours and 14 minutes with no human playing, according to a DeepMind research report. The agents wrote and ran 2,783 Ja…
AngusRepo has published documentation for its in-process MCP read-only endpoint, which exposes eight tools prefixed repo_ by default for AI coding agents to query artifact metadata without proxying /maven, /npm, /v2 or o…
DeepSeek released two standalone APIs on October 17, 2026 — Search, which returns structured web results, and Fetch, which retrieves a URL as Markdown — each priced at $0.75 per 1,000 successful requests. A developer-ori…
Creator Brooke Lacey, who has built a TikTok audience of more than 290,000, uses a workflow combining Claude Code, FFmpeg and OpenAI's Whisper to edit her short-form videos in about 15 seconds, she said on the How I AI p…
The Data Intelligence Lab at the University of Hong Kong (HKUDS) released CLI-Anything, an open-source tool that generates command-line harnesses letting AI agents drive real software such as Blender, GIMP, LibreOffice, …