#LLM
Curated coverage on «LLM»: releases, research, benchmarks, open-source.
AI is forcing universities to rethink coursework, but cheating predates ChatGPT
Dr. Nafisa Baba-Ahmed argues AI did not create the crisis in university coursework. It only scaled an old habit of moving thinking outside the student.
LLM experiment shows how a model's "personality" takes shape in latent space
An experiment with a modular LLM suggests a separate latent vector can store not just text style but stable behavioral traits, resembling the scaffolding of a…
IBM releases Granite 4.0 1B Speech, a compact multilingual speech model for edge AI
IBM has introduced Granite 4.0 1B Speech, a lightweight model for speech recognition and translation built for fast deployment in resource-constrained enterprise…
Google AI Ultra: Turning a Subscription into a Pool of Parallel Agents and Model Consensus
Britannica and Merriam-Webster sue OpenAI over nearly 100,000 articles
Gateway Global AI builds a voice platform to speed up AI adoption in business
OpenAI releases GPT-5.4 Pro: new records on ARC-AGI-2, FrontierMath and logic tasks
Anthropic and Claude at the Center of a Debate: Could a Chatbot Go Against Big Tech's Logic?
A column on Claude shows how politeness toward a chatbot and talk of possible AI consciousness raise an unexpected question: could such an assistant one day defy…
OpenAI's ChatGPT 5.4 Beats Claude Opus 4.6 and Gemini 3.1 Pro in a Habr Comparison
A Habr comparison of three flagship models on routine tasks ranks ChatGPT 5.4 first on total score. Gemini 3.1 Pro is the cheapest, and Claude Opus 4.6 the most…
Anthropic explains how to build skills for Claude Code and why teams need their own marketplace
Claude Opus 4.6 spots hidden trap in PDF, shows new rules of hiring
ChatGPT Search and Google AI Overviews Are Changing SEO: Habr AI Publishes a GEO Guide
OTUS: Why Language Models Hallucinate and Which Measures Reduce the Risk of Errors
MIT proposes a metric that catches confident LLM errors and hallucinations
MIT researchers built a system that scores not just an LLM's confidence in its answer but also how far it diverges from other models, making hallucinations easier…
Reson8 raises €5 million to build European voice AI for accurate speech recognition
Amsterdam startup Reson8 has raised €5 million in pre-seed funding to develop customizable voice AI for European languages and industry terminology, challenging US…