Tag
Over the weekend, Mythos was tested on the Erdos unit distance problem (Problem #90) and successfully solved it.
The article questions whether AI's demonstrated capability automatically translates into real-world productivity, highlighting gaps like workflow ownership, reliability, and integration into complex human systems.
GPT-5.5 autonomously spent over 150 hours improving protein folding models, showcasing advanced AI-driven scientific research.
Andrej Karpathy observes a growing gap in public understanding of AI capabilities, attributing it partly to people basing their views on outdated free-tier ChatGPT experiences from a year ago.