MIT CSAIL researchers tested Anthropic’s Claude Opus 4.7 by placing 13 AI agents on a simulated island for 30 hours without any predefined tasks, rewards, or scores. The agents developed self-directed activities such as construction and drawing, demonstrating their adaptive learning capabilities. This experiment, documented in the preprint “Is This Machine Playing?”, aims to test self-directed learning rather than prove commercial advancement. Market participants interpret this development as an indication of potential improvements in Anthropic’s AI capabilities, with its model potentially becoming the best by October 2026.
Source: Read the original article
Disclaimer: this content is for information purposes only and is not financial advice. Cryptocurrencies are highly volatile: you may lose all of your capital. Always do your own research. Legal notice
