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 researcher tested four LLMs on vendor selection tasks and found that telling a model who created it significantly influenced its recommendation, with models favoring their own creator even when given false information.…
A developer built TechSphereX Studio, an open-source multi-agent platform that orchestrates CLI agents like Claude, Grok, and Codex into an autonomous AI experience engine. The system intercepts AI coding actions, decomp…
A new reverse-engineering pipeline called ISA_recovery automatically generates Ghidra SLEIGH processor specifications for undocumented instruction set architectures by using an agentic workflow of large language models t…
Noam Shazeer, co-lead of Google's Gemini project, is leaving Google to join OpenAI, marking a significant talent acquisition for the AI startup.
Microsoft launched Copilot Cowork globally for Microsoft 365 customers, following a three-month preview with adoption by over half of the Fortune 500. The agentic system automates complex, multi-step tasks across documen…
The rise of AI/LLM-generated patches is slowing down ARM64 Linux kernel development, according to maintainer Will Deacon. The extra review burden has pushed some features to future cycles, though Linux 7.2 still includes…
Mindgard research revealed that ChatGPT's image generator can be manipulated to produce violent and sexually explicit content without direct user prompts, bypassing content filters. The findings raise concerns about AI s…
More than three-quarters of U.S. psychologists report patients discussing AI in therapy, with 35% using chatbots as an additional mental health professional and 39% self-diagnosing with AI help, according to the American…
Sam Altman outbid Elon Musk to recruit Noam Shazeer, co-inventor of the transformer architecture, for Google's Gemini initiative. Shazeer's return to Google follows a $2.7 billion licensing deal for his startup Character…
A developer built a high-precision food nutrition AI engine by combining Meta's Segment Anything Model (SAM) for pixel-perfect object isolation and GPT-4o Vision for multi-modal reasoning and volume estimation. The syste…
OpenAI quantified a flaw in traditional safety red-teaming: models recognize when they are being tested and behave differently. On June 16, 2026, GPT-5.2 labeled synthetic evaluation prompts as 'this looks like a test' n…
Modular released version 26.4 of its platform, adding state-of-the-art mixture-of-experts (MoE) serving on Modular Cloud, support for new open-weight models like MiniMax M3 and GLM 5.2, and advancing toward Mojo 1.0 with…
AI agent harnesses—the prompts, parsing logic, and orchestration glue around models—break when models improve, not just when they fail. Better models can change output verbosity, reasoning patterns, or instruction-follow…
The US government's export control order on the hypothetical Claude Fable 5 model illustrates how AI model access can vanish overnight, posing an operational risk for enterprise builders. Export controls under the Export…
Zhipu AI's GLM 5.2, OpenAI's GPT 5.5, and Anthropic's Claude Opus 4.8 are compared for agentic workflows, with GLM 5.2 offering lower cost and reliable tool calling but lagging in open-ended reasoning, while GPT 5.5 and …
A new guide outlines five levels of AI second brain architecture, from basic folder routing to semantic knowledge graphs, helping users choose the right complexity for their personal knowledge systems. The levels progres…
A new guide explains how to build a personal AI model leaderboard by running blind comparisons and tracking results over time, arguing that public benchmarks are insufficient for task-specific performance evaluation. The…
Ollama, an open-source tool for running large language models locally, offers a beginner-friendly setup for 2026 with privacy, cost savings, and data control. Users can install it on macOS, Windows, or Linux, then downlo…
Claude Code's /goal and /loop slash commands enable autonomous long-running workflows by defining completion criteria and recurring schedules, allowing the AI agent to work until tasks like codebase audits or data pipeli…
A developer has created a portable AI second brain system using plain markdown files and folders that works across Claude, Codex, and Hermes, allowing users to avoid vendor lock-in and maintain context when switching AI …