When AI models will not get better and will be stuck at one place?
retained text
My idea is that AI models learn from data from the internet, so when the data from the internet and people's responses from AI conversations run out, when will there be a point when AI will no longer improve and its progress will stop? At that point, AI will only improve by tiny bits that we won't even notice that the model has improved. Is my knowledge about model training wrong? Do you think so too?
document fetched 2026-10-04 00:35 · canonical https://reddit.com/r/ClaudeCode/comments/1wwhxns/when_ai_models_will_not_get_better_and_will_be