Reinforcement learning can fine-tune AI models by training them on how elite finance professionals work, teaching the system to build financial models and draft transaction advice at expert level. Speaker B says this approach will transform finance in the next 12-18 months, though it's largely overlooked on Wall Street today.
Chapters
What non obvious AI trends are you paying attention to that most people in finance are missing?
The big one coming out of that thread is the power of reinforcement learning in finance. Reinforcement learning is a, is a paradigm for building AI models that involves taking as an input the foundation models that come out of places like OpenAI and Anthropic and Google Gemini and so forth, and you know, in effect training them on what good looks like to elite human professionals in, for example, high finance, as it relates to, again, building financial models, delivering great advice around a transaction for, for a client, having fairness opinion materials, for example, that are incredibly high integrity and that can withstand the scrutiny of the public markets, you know, for every facet of what makes a phenomenal banker or investor, building a model that knows what good looks like and uses that to train that general finance, that general intelligence model, so that it becomes an elite thinker and doer as it relates to finance outputs. And the model learns what good looks like. And we think the power of that over the next 12, 18 months is going to be completely transformative above and beyond what the commercial AI that all of us are using is already doing. So that's a trend that I think is on top of mind for the AI community. A little under the radar right now in the world of Wall street and finance, but it's going to be top of mind for everyone, you know, within 12 to 18 months.