Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
An Anthropic employee expressed confusion about their role, stating that on good days they feel automated work renders their efforts meaningless, while on bad days they cannot understand why systems break. The employee's…
A tech writer fine-tuned a local LLM to replicate the style of 1980s and 1990s software documentation, using a 37-million-word corpus of scanned Microsoft manuals from the Bitsavers archive. The experiment, conducted on …
A developer has built a "Talk Mode" extension for the Pi coding agent that prevents the AI from editing, creating, or modifying files. The extension, toggled with Shift+Tab, injects a system prompt that restricts the age…
Scientists are investigating whether honey bees and artificial intelligence systems like ChatGPT could possess consciousness, moving beyond behavioral tests to examine internal information processing mechanisms. Two new …
A new survey of open-source agent systems finds that task execution reliability depends more on the infrastructure layer wrapping large language models than on the models themselves. The research introduces the seven-lay…
Mira Murati, former OpenAI CTO and current CEO of Thinking Machines Lab, made her first major media appearance in 18 months on Thursday, using a Bloomberg interview to preview a new AI interface technology while carefull…
A developer built a fleet system to run swarms of 10-15 coding agents concurrently and documented lessons learned from scaling the system. Key findings include that CLAUDE.md files waste tokens by loading irrelevant inst…
Meta Platforms Inc. Chief AI Officer Alexandr Wang said health capabilities will be a key differentiator for future Meta AI models as the company scales them to billions of users. Wang, who joined Meta from Scale AI afte…
BoxAgnts has released a runtime architecture that separates the Model Context Protocol (MCP) layer from a dedicated execution runtime, arguing that protocol standardization alone is insufficient for safe AI tool executio…
A new study from researchers at Northwestern, Stanford, Cornell, and All Hands AI reveals that frontier LLM agents waste 28–64% of tokens on tasks they are doomed to fail. The paper, BAGEN: Are LLM Agents Budget-Aware?, …
Researchers at Tevatron released BrowseComp-Plus, a new benchmark designed to evaluate deep-research AI agents by isolating the effects of retrievers and large language models for fair and reproducible comparisons. The b…
OpenAI's ad partner Criteo cut its minimum ad spend requirement from $50,000 to $10,000 and will match client spending dollar-for-dollar to lower barriers for ChatGPT advertising. The incentives come as ChatGPT surpasses…
In the week of May 27 to June 3, 2026, the cost of running AI agent infrastructure emerged as its own distinct category of work, marked by three key signals. The compression tool Headroom gained 6,322 GitHub stars by pro…
Researchers have developed ReasoningFlow, a framework that maps the non-linear discourse structures of large reasoning model (LRM) traces into directed acyclic graphs (DAGs) to improve evaluation and monitoring. Through …
Researchers introduced a two-stage early failure alerting system for dialogs and LLM-agent trajectories that learns from sparse turn-level evidence rather than assigning terminal labels to every prefix. The attention-bas…
Researchers have developed a system that automatically discovers an executable schema from raw multi-source data, using it as a shared contract for knowledge graph construction and query-time retrieval. The system improv…
Researchers have developed TTT-RTL, the first per-design test-time training framework that uses reinforcement learning to optimize hardware designs generated by large language models. The system closes the loop between a…
Researchers at a major AI lab have identified a specific neural subgraph within a large language model (Qwen3-4B-Instruct-2507) that governs how the model trades off short-term gains against long-term consequences. The t…
Researchers have developed BUMP, a self-supervised framework that trains large language models to generate personalized user profiles without requiring labeled data from downstream tasks. The system uses a bidirectional …
Researchers introduced VideoKR, a large-scale training corpus of 315,000 video reasoning examples over 145,000 expert-domain videos designed to strengthen knowledge- and reasoning-intensive video understanding. The team …