Breakthroughs and research · ChatGPT

OpenAI publishes 372 families of mathematical results

OpenAI has posted 722 manuscripts divided into 372 result families online. The results come from an internal AI model and have not yet been assessed by other researchers.

Read the full article

Breakthroughs and research

OpenAI let new model take on 4,000 unsolved math problems

OpenAI has published 722 mathematical manuscripts from an internal, as yet unreleased model that tackled around 4,000 open problems, but the results have not been peer reviewed.

Breakthroughs and research

AI assistants disagree on which brands they mention

A new analysis from the company LLM Scout shows that ChatGPT, Claude, Gemini and Perplexity disagree on mentioning a particular brand in 27.2 percent of cases. This means that the answer to a question about, for example, buying a product may depend on which AI you ask.

Breakthroughs and research

AI creates 3D scenes from a single photo, but cannot see its own mistakes

A new method, LEGO-Anything, can turn an ordinary photo into code that builds a 3D scene in the program Blender. But the AI agents cannot assess for themselves whether the result is accurate.

More news

Get the week's AI news in your inbox

Choose your level, topics and length. One email a week, unsubscribe at any time.

Subscribe to Promptly Newsletter