Photo by Vlada Karpovich on Pexels

Meta has observed preliminary evidence suggesting that its AI models are beginning to demonstrate self-improvement. This development has prompted the tech giant to re-evaluate its open-source strategy for AI, with a more cautious approach now under consideration. The initial discussion surrounding this can be found on Reddit: https://old.reddit.com/r/artificial/comments/1mdw2id/meta_sees_early_signs_of_selfimproving_ai_signals/