I am an economist, founder, passionate sailor and coach. I'm fascinated by the opportunities the agentic era presents for economists, and I fundamentally believe economists are better equipped than anyone to define what this era looks like — particularly with respect to how agents will interact in a marketplace. For the past year I have been researching the applications of machine learning and reinforcement learning in economics, and crucially the applications of economics in ML and RL.
The most interesting intersection, I believe, lies in defining what a marketplace looks like in agentic commerce. Is it time to rethink what it means for something to have a price? When the barriers to negotiation have fallen to almost zero (agents negotiate almost instantly), listing fixed prices is no longer an efficient mechanism. To what extent can we utilise the perfect rationality of agents to design favourable mechanisms previously unattainable to humans with bounded rationality? Do agents behave perfectly rationally? Do different sessions of the same underlying model / agent behave as if it's a one-shot interaction or a repeated interaction? If millions of people utilise the same underlying ‘buying agent’, will they coordinate and give consumers market power?
I am applying this research to the real world as a founder. We're starting with the problem of “How do you make an agent choose your product?” — and we already have a good idea of the answer.
Do read my research reports, or for a more consumable format, consult the blog summaries.