Thank you for this thoughtful reply, this comment is basically the reason this update exists. You were right that Hamilton is probably right.
I have written a longer update incorporating Hamilton's reanalysis and extending the economics in two directions: a quantitative treatment of verification as the binding constraint, and a systematic look at the economic conditions under which a genuinely dangerous autonomous agent actually gets to run.
Curious whether you think the analysis holds up, and whether there are important considerations I have missed!
Hi! I made a BOTEC model on AI agents economic feasibility in the near term. You can find it here. It combines the half-life analysis of agent reliability with real inference costs. The core idea is that if agent cost per successful outcome scales exponentially with task length, and human cost scales linearly, it creates a sharp viability boundary that cost reductions alone cannot meaningfully shift. The only parameter that matters much is the agent's half-life (reliability horizon), which is precisely the thing that requires the continual learning breakthrough (which I think is essential for AGI-level agents) that some place 5-20 years away. I think this has underappreciated implications for the $2T+ AI infrastructure investment thesis.
Thank you for this thoughtful reply, this comment is basically the reason this update exists. You were right that Hamilton is probably right.
I have written a longer update incorporating Hamilton's reanalysis and extending the economics in two directions: a quantitative treatment of verification as the binding constraint, and a systematic look at the economic conditions under which a genuinely dangerous autonomous agent actually gets to run.
Curious whether you think the analysis holds up, and whether there are important considerations I have missed!
Hi! I made a BOTEC model on AI agents economic feasibility in the near term. You can find it here.
It combines the half-life analysis of agent reliability with real inference costs. The core idea is that if agent cost per successful outcome scales exponentially with task length, and human cost scales linearly, it creates a sharp viability boundary that cost reductions alone cannot meaningfully shift. The only parameter that matters much is the agent's half-life (reliability horizon), which is precisely the thing that requires the continual learning breakthrough (which I think is essential for AGI-level agents) that some place 5-20 years away. I think this has underappreciated implications for the $2T+ AI infrastructure investment thesis.
Emails were sent out just now!
Hi! Half of the time is spent on MechInterp, the other half on other topics (RL and paper replication).