the reason why people think LLM progress is slowing is because the industry overhyped it with promises like: > AGI by 2025 > full job replacement by 2025 expectations are too high, so real gains look dull. we now have agentic coders, stronger audio models, and better reasoning models -- plus internal models winning gold in IMO/ICPC contests progress hasn't slowed; it's just steady, not explosive