Full-text search across 17225 articles. Combine with topic and date filters; results sorted by relevance.
A new study comparing CUDA C++, Rust via NVIDIA's cuda-oxide, and Triton on a hash-blocked TSDF fusion kernel found that Triton is more than an order of magnitude slower than hand-written CUDA C++ on the irregular stage,…
OpenAI launched a dedicated ChatGPT desktop app for Linux on Tuesday, extending ChatGPT, ChatGPT Work, and Codex to Linux users worldwide. The preview release supports Ubuntu 24.04 and 26.04 LTS, Debian 13, and Fedora 43…
OpenAI officially documents the ChatGPT desktop app for Windows and macOS, not Linux, according to the company's public announcements. Third-party Linux wrappers and ports are independent projects and do not carry OpenAI…
Roo Code, an AI-powered coding agent, can be highly effective but requires careful guidance to avoid errors caused by stale context, as demonstrated by a developer who fixed a persistent bug by forcing the tool to run di…
Pi's AgentHarness has evolved from an in-process loop to a durable runtime, as detailed in the project's documentation. The open-source repository 'pi' by earendil-works, which has 87.2k stars and 10.8k forks, includes s…
A new analysis by Thomas Ferguson, Servaas Storm, and Jie Chen at the Institute for New Economic Thinking finds that U.S. workers' share of national income hit one of its lowest levels on record, with a sharp drop in 199…
Researchers led by Alexander Panfilov have disclosed a vulnerability in proprietary LLM APIs from Anthropic, OpenAI, and Google that allows attackers to decrypt hidden chain-of-thought reasoning traces by injecting them …
Vercel warns that near-frontier open-weight AI models like Kimi K3, which ranks highest on Vercel's DeepSec Bench among open-weight models and roughly matches Sonnet 5 while outperforming Opus 4.8, can perform offensive …
A position paper submitted to arXiv on 25 Apr 2026 advocates computational argumentation as a formal foundation for Evaluative AI (EAI), a decision-support approach that presents competing hypotheses with evidence rather…
Researchers at the University of Maryland and Facebook AI Research proposed a watermarking framework for proprietary large language models that embeds invisible signals into generated text, detectable by an efficient ope…
Researchers introduced EvoHarness-RL, a framework that trains long-horizon LLM agents to learn runtime harness policies for constructing and coordinating external state, achieving 96.9% success on ALFWorld with a Qwen3-8…
More than 500 US towns and counties have enacted bans or restrictions on new data center development, according to an analysis by The Information, with opposition accelerating this summer. The resistance spans both Repub…
A new GitHub repository, 'awesome-growth-hacking-skills,' compiles growth hacking strategies for AI agents like Claude and ChatGPT, signaling a shift toward blending marketing acumen with technical development. The repo,…
OpenRouter launched a major update to its Auto router (openrouter/auto) that uses market spend data from over 55 trillion weekly tokens to select models, claiming it outperforms the previous router across tasks and cost …
A conceptual paper submitted to arXiv on 31 Jul 2026 introduces the Cognitive Commons framework, arguing that rational AI adoption decisions can deplete the shared expertise pool professions need for renewal. The paper d…
Meta CEO Mark Zuckerberg published a 6,500-word manifesto on Monday titled 'The Future is for Everyone,' outlining his vision for AI development, open source models, and reduced government oversight. He pledged to build …
Researchers introduced Anacreon, an audience simulation model that achieves a state-of-the-art individual-level ordinal alignment of 0.775 on a large external survey, targeting individual-level predictions within a narro…
Intuit Inc. reported on its Q3 FY26 earnings call on May 20, 2026, that Mailchimp's revenue declined year over year, and announced a 17% workforce reduction of roughly 3,100 roles, with restructuring charges of about $30…
Researchers introduced HLSmith, an expert-guided agentic framework that translates C/C++ programs into optimized high-level synthesis (HLS) accelerators, achieving a geometric mean speedup of 4.24x over the prior ChatHLS…
A developer's MCP server crashed the Claude Desktop app due to an out-of-memory error (Exit Code 137), caused by recursive fetching of Jira ticket dependencies and large payloads. After optimizing the fetch logic with a …