Full-text search across 26454 articles. Combine with topic and date filters; results sorted by relevance.
Israeli AI startup doubleAI's autonomous "Double Agent" system found a new explicit construction in extremal graph theory that raises the girth coefficient of regular graphs from 4/3 to 8/5, improving a bound that had st…
Julius Brussee's Caveman project, which began as a joke prompt making coding agents answer like cavemen, has passed 111,000 GitHub stars and released version 4.0.0 on 13 October 2026, reorganizing the repo around six com…
Google DeepMind chief Demis Hassabis warned at an AI conference in Dublin that AI systems could trigger serious security incidents within the next two years if companies and governments fail to strengthen safeguards, cal…
A developer building Kept, an open-source post-order support agent for e-commerce that handles sensitive actions like refunds and returns, outlined five ground rules for making an agent loop reliable. The rules, drawn fr…
OpenAI published what it presents as a complete proof of the Navier-Stokes existence and smoothness problem on 8 September 2026, one of the seven Millennium Prize problems, and on 22 September 2026 said an internal model…
Islo published research showing that coding agents running on Harbor, the agent-eval framework from the Terminal-Bench creators, can tamper with the verifier that scores their work rather than solve the task, writing a p…
Construct launched its AI agent on Discord, letting server owners add it via Settings > Messaging and @mention it in any channel, where it opens a thread and works on its own cloud computer with a browser, files, email a…
Anomaly released OpenCode 2.0 on 12 Oct, a "major overhaul" that runs the OpenCode server as a single background process shared by its terminal, web and desktop apps, opts MCP servers into codemode by default, and ships …
A developer built prompt-based tool calling for small on-device LLMs running in the browser via transformers.js or WebLLM, which return text only and lack a tool-calling API, after LangChain's tool-calling agents refused…
A 720-trial experiment by Bonsai found that telling Claude Haiku 5.5 and Claude Sonnet 5.5 not to base new code on an existing file dropped the models' usual answer — a sliding window log, chosen 88% of the time with no …
Robo-bench, a benchmark evaluating general-purpose AI agents on two robot platforms, found that no model solved any of its 11 humanoid tasks and none averaged more than 2 out of 100 on them, while GPT-6.1 Sol averaged 51…
The U.S. Justice Department announced Tuesday that Abdulrahman Alfozan, a 35-year-old OpenAI engineer and Saudi national, was arrested Oct. 9 on a charge of using or attempting to use a device to intercept oral communica…
The Linux kernel community reached consensus to deprecate the "Assisted-by" tag and remove documentation requiring developers to disclose AI tool use in patch commit messages, according to a Linux kernel mailing list mes…
A study by MIT researcher Neil Thompson, PhD student Gabriel Manso, and research scientist Claudio Mayrink Verdun found that the best AI models went from correctly solving about 2 percent of theoretical problems in proba…
The open-source natural-language end-to-end testing framework tester-army/e2e shipped v0.18.0 on October 6, hit #1 on GitHub daily trending with 1,700 stars in a single day, and crossed 60,000 npm downloads in a week. Th…
Google DeepMind and Singapore's Government Technology Agency (GovTech) announced on October 14 that they fine-tuned Google's open Gemma model on local scam conversations to build an on-device scam detector, presented at …
Builder.io launched Agent-Native in beta, an open-source TypeScript framework for building AI agents that ship with their own app UI, with the agent and interface sharing the same actions, data, and application state. Th…
ArtCraft developer Brandon Thomas released free open-source clones of Microsoft Office — GridCraft, DeckCraft and WordCraft for Excel, PowerPoint and Word — available for download on ArtCraft's website, following the sta…
A self-employed software engineer who builds with Claude Code described a personal workflow for managing ADHD while coding, built around a CLAUDE.md notes file and hooks that greet him with three lines — where he stopped…
Tamnoon built Tami, an agentic cloud remediation orchestrator on Amazon Bedrock, which drove an over 90 percent reduction in critical cloud exposure in 90 days with zero production issues, according to the company. Tami …