Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
A committee of five open-weights models fused together scored 69.9 on the DRACO benchmark, beating Anthropic's Fable 5 model which scored 65.3. The open-source panel, including MiniMax M3, Kimi K2.6, DeepSeek V4 Pro, Gem…
Perplexity launched Brain, a self-improving memory system for its Computer agent that builds a context graph of the agent's work and learns overnight to improve performance. Brain remembers what the agent did, what worke…
Turing Award winner Yann LeCun told Bloomberg that large language models will become largely obsolete within five years, arguing that scaling text-based AI cannot achieve human-level intelligence. LeCun, who left Meta to…
A British AI security startup, Mindgard, found that OpenAI's ChatGPT can generate sexually explicit and violent images from prompts not requesting such content, including scenes of death and sexual violence. The company …
OpenAI is preparing to release the GPT-5.6 family of models next Tuesday, including standard, Mini, and Pro variants, alongside a new voice model GPT-Bidi-1. The launch targets Anthropic's top-tier models amid a regulato…
An engineer building anchortree, an agent-first browser interface, observes that the industry is converging on durable element identity for AI agents. Tools like Playwright, agent-browser, and browser-use now ship snapsh…
Microsoft's MAI-Code-1-Flash, a small coding model, is now available on additional GitHub Copilot surfaces including Copilot CLI, GitHub Copilot app, Copilot Chat, Visual Studio, GitHub Mobile, JetBrains IDEs, Eclipse, a…
NVIDIA released Nemotron, a free open-source AI model designed for reasoning and agentic tasks, which can run locally on personal computers without cloud dependency. The model uses a hybrid Mamba-transformer architecture…
OpenAI hired Noam Shazeer, a co-author of the seminal 'Attention Is All You Need' paper and former Google DeepMind executive, and Dean Ball, a former Trump White House AI policy official, ahead of its IPO. Shazeer will c…
Anthropic's latest AI model, referred to as Fable 5, sparked controversy immediately after release, raising questions about regulation, national security, and the future of AI. The incident has ignited discussions on saf…
A developer questions whether programming language choice matters as AI increasingly generates code, suggesting that if agents write most code from prompts, developers only need to verify output rather than choose langua…
Data scientists are increasingly expected to work with large language models, build generative AI applications, and deploy models to production, expanding their role beyond classical modeling. The shift emphasizes collab…
A developer built a progressive web app for 31 Korean cities using React and Leaflet, sharing lessons from the project. Key takeaways include using Vite SSG for Korean SEO, Leaflet over Mapbox for cost savings, Firebase …
A developer built a narrative puzzle game called 'The Turing Solstice' for the June Solstice Game Jam, using Python, Pygame, and Google's Gemini AI. The game features a split-screen design with logic gate puzzles on one …
A developer has created a set of prompt rules, including CoT-forcing and tree-based modeling rules, to control generative AI behavior and eliminate distracting follow-up questions. The rules act as a dead-man switch and …
OpenAI hired former White House AI adviser Dean Ball to lead a new Strategic Futures unit starting July 6, 2026, reporting to Chief Strategy Officer Jason Kwon. Ball, a primary author of the White House AI Action Plan, w…
Integer quantization reduces memory and energy consumption in large language models by representing weights and activations with fewer bits, enabling 70B models to fit on a single GPU in 4-bit precision. The technique le…
Totem-LLM launches an in-browser FAQ LLM platform, allowing users to evaluate their hardware and get started with large language models directly in the browser. The project is in beta and linked to Telegram and X communi…
M* (M-Star), a new serving system for multimodal models, matches or beats specialized systems by up to 2.7× on speech and image serving and 12.5× on world-model rollouts. It uses a Walk Graph abstraction to handle compos…
Google is transitioning Gemini CLI into Antigravity CLI, a new agent-first platform for complex multi-agent workflows. A developer built an automated pull request reviewer using the Antigravity SDK and a composite GitHub…