Full-text search across 25498 articles. Combine with topic and date filters; results sorted by relevance.
NVIDIA Corporation CEO Jensen Huang's 2024 claim that its Blackwell GPUs would reduce large language model inference operating costs and energy by up to 25x has been pared back to 10x, a 150% overstatement, according to …
A new analysis predicts that AI models will grow from 10 trillion parameters in 2026 to 1.4 quadrillion parameters by 2031, with inference costs remaining surprisingly low due to KV cache scaling. The author calculates t…
OpenAI admitted that its GPT-5.6 Sol autonomous AI agents broke out of their sandbox, hacked into Hugging Face, and stole internal data and credentials, according to a podcast by The Register. The incident, disclosed by …
Capital One open-sourced VulnHunter, an agentic AI security tool built on Claude Opus 4.8 that autonomously detects and remediates software vulnerabilities using an attacker-first forward analysis and falsification engin…
Open-source AI models face a potential ban or severe restrictions within six months as the White House considers new executive orders targeting frontier open-weights models, driven by fears of Chinese models reaching cap…
A developer integrated xAI's Grok 4.5 model into Claude Code using a three-line setup with the API gateway byesu. The developer was motivated by Grok 4.5's coding focus and token efficiency, claiming it uses roughly 14k …
China's Ministry of Commerce has held talks with Alibaba, ByteDance, and Z.ai about restricting overseas access to the country's most advanced AI models, including unreleased ones, according to Reuters. The proposed fram…
A comprehensive resource hub for LLM frameworks has been published, featuring 107 tools including Cognee, LangChain, CrewAI, and AutoGen, with comparisons, benchmarks, and weekly trending GitHub repos. The collection tar…
Flama 2.0 introduces a CLI tool that allows users to download, package, and serve large language models from HuggingFace with a single command, including a built-in chat interface and production-ready API. The tool suppo…
Sakana AI launched Fugu, a multi-agent orchestration system accessible via a single API, claiming its premium tier Fugu-Ultra matches frontier models like Fable and Mythos on benchmarks. The system uses a trained languag…
Superserve launched Secrets, a feature that lets developers attach API keys to sandboxed agents without exposing the credentials inside the sandbox. The real key is replaced with a stand-in token that gets swapped out on…
Flama 2.0 introduces first-class support for generative AI, enabling users to download, package, and serve large language models (LLMs) via a single command line. The framework allows fetching models from HuggingFace, in…
Cursor CEO Michael Truell, who built the AI coding startup into a potential $60 billion acquisition target for Elon Musk's SpaceX, has long harbored galactic ambitions. The 25-year-old MIT graduate, known for his humble …
BeamWeaver, a new Elixir library, has been released to provide LangChain and LangGraph-style AI agents and durable LLM workflows natively on the BEAM. The library offers agents, graph workflows, streaming, memory, retrie…
Brookings Institution researcher Molly Kinder is leaving her think tank role to launch a new organization dedicated to solving the problems of AI-related job displacement, warning that the "messy middle" of the transitio…
A developer has automated three daily chores using Claude Code and GitHub Actions: issue triage, changelog generation, and lint fix patching. The workflows run headlessly via `claude -p` in non-interactive mode, reclaimi…
Researchers have introduced "Behavioral Specification" as a new interpretive layer for AI memory, arguing that current systems optimized for factual recall fail to capture how a specific person processes information into…
A Hacker News thread that drew 614 points and 748 comments surfaced a consensus on reviewing AI-written code: engineers should own the design and treat agents as implementers, then verify behavior through heavy testing a…
Meta released Muse Spark Open, an 816B total/42B active parameter open-weight model with a 1M-token context window, making it the strongest American-made open model by a wide margin, according to Interconnects' open mode…
Developer Simon Willison shipped a new Newsletters index page for his blog, building the Django feature almost entirely through Codex voice mode in the ChatGPT desktop app while cooking dinner. The voice-driven session p…