I certainly agree with your counterargument. I’m surprised there’s so much skepticism that LLMs can lead to AGI. It’s like this group of futurists doesn’t think in terms of ongoing progress.
The author nonetheless predicts around 60% chance that LLMs lead to AGI, and 80% chance they do if they have scaffolding to turn them into language model cognitive architectures. My argument for how those likely lead quickly to AGI is here.
I certainly agree with your counterargument. I’m surprised there’s so much skepticism that LLMs can lead to AGI. It’s like this group of futurists doesn’t think in terms of ongoing progress.
The recent post LLM Generality is a Timeline Crux addresses this question and generated some good discussion, much of it skeptical.
The author nonetheless predicts around 60% chance that LLMs lead to AGI, and 80% chance they do if they have scaffolding to turn them into language model cognitive architectures. My argument for how those likely lead quickly to AGI is here.