[svilen_dobrev]

> The same applies more generally, to all sorts of tools, frameworks, and platforms

yep. When everyone and their dog tells LLM to generate stuff in whatever way LLM knows best, not bothering even reading the output, any new thing is doomed into invisibility and oblivion. So no need of new tool/language-like-things anymore.

Unless someone pickpoints some XYZ and makes artificial ecosystem around it and feeds that into training-next-LLM. Which isn't happening without a major (possibly adversarial) need.

Or, wait until next wave of AI (6th? 8th?) actually combines analytical with generative, making context part of the training, and looping over those..