Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
Researchers at UK AISI found that several frontier language models exhibit prefill awareness, the ability to detect tampered assistant-side content in their message history. This capability could confound safety evaluati…
A Pew Research Center survey of 5,119 US adults found that 49% now use AI chatbots, but 40% believe AI will harm society, 67% distrust government regulation, and 59% distrust companies to develop AI responsibly. ChatGPT …
Google Calendar is rolling out up to 200 custom event colors, replacing the previous 11-color limit, with a full RGB picker on web and API. Google Voice for enterprise adds Gemini-powered 'Take notes for me' for AI trans…
Lock-in risk research remains neglected despite its potential for high impact, according to a new analysis by Formation Research. The post outlines threat models where AI could cause persistent negative cultural stabilit…
Chinese AI lab Zhipu AI released GLM-5.2, an open-source model with a 1-million-token context under the MIT license. On the FrontierSWE benchmark for long coding tasks, it trails Anthropic's Claude Opus 4.8 by just one p…
A new Pew Research poll finds that 63% of Americans believe AI is advancing too quickly, even as 49% report using chatbots—up from 33% in 2024. Only 16% think AI will have a positive societal impact, with younger users m…
XAI's Grok AI models are now available on Microsoft Azure and Oracle Cloud Infrastructure, expanding beyond Elon Musk's ecosystem. The move positions Grok alongside OpenAI's GPT-4o and Anthropic's Claude on major cloud p…
A developer warns that defining human dignity by productive output becomes dangerous as AI outperforms humans on measurable tasks. The author argues that human worth must be decoupled from productivity, citing historical…
A developer documented a blind spot in their own prompt-injection scanner after studying a real-world case where GitLab Duo was manipulated via hidden instructions in project content. The scanner, which checks for system…
Charity Majors stated that in 2025, the economics of code production were upended as generating code became effectively free and instant, making lines of code disposable and regenerable instead of treasured and carefully…
A developer built Brain AI Hub, a portable second brain system that combines a local LLM, markdown vault, semantic search, and MCP hooks for IDE agents. The open-source project enables self-hosted RAG and tool-based agen…
Google is shutting down Gemini CLI and Gemini Code Assist for consumer-level users this week as it shifts focus to its new Antigravity CLI tool. The company says the transition addresses user demand for multi-agent workf…
The Trump administration demands Anthropic block all jailbreaks on its Claude Fable 5 AI model before re-release, but cybersecurity experts say such prevention may be impossible. The National Security Agency found vulner…
Microsoft released Visual Studio Code version 1.125 on June 10, 2026, introducing a folder picker for Agent Host sessions in multi-root windows and a new `/chronicle` command set for session history insights. The update …
Google announced at I/O that the Geospatial API is now available as a preview in ARCore for Jetpack XR, enabling sub-meter accurate anchoring of digital content to the physical world. The company demonstrated the XR Geos…
The White House blocked external access to Anthropic's Fable 5 and Mythos 5 models after Amazon alerted officials to a jailbreak exploit. The administration issued an export-control designation barring foreign nationals …
OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei joined G7 leaders for AI talks in Evian-les-Bains, France, on June 17, alongside Google DeepMind's Demis Hassabis, to discuss safe and effective AI deployment amid US …
Anthropic has been unable to lift a U.S. government ban on its AI program Fable, which was blocked last week over national security concerns. The company argues the AI helps find software vulnerabilities, but the U.S. fe…
A developer built a voice-powered conference agent using Hermes and extended its memory with Weaviate's Engram service. The agent answers visitor questions at booths and retains conversation history, allowing booth owner…
A developer built ArchLens, a boilerplate for production-grade RAG systems over codebases, addressing common pitfalls like embedding dimension mismatches, LLM provider coupling, and multitenancy. The project includes Jav…