Here's why I know that OpenAI is stuck in a hype cycle. For all of 2024, the cry from employees was "PhD level models are coming this year; just imagine what you can do when everyone has PhD level intelligence at their beck and call". And, indeed, PhD level models did arrive...if you consider GPQA to be a benchmark that is particularly meaningful in the real world. Why should I take this year's pronouncements seriously, given this?

OpenAI is what you get when you take Goodhart's Law to the extreme. They are so focused on benchmarks that they are completely blind to the rate of progress that actual matters (hint...it's not model capability in a vacuum).

Yann indeed does believe that AGI will arrive in a decade, but the important thing is that he is honest that this is an uncertain estimate and is based off of extrapolation.