The question we get most often asked at Preseen is if we’re good at forecasting the future why don’t we open up a hedge fund.

We believe that this question is provincial.

AI forecasting will be far larger than what any fund can manage. When people make a decision, they select an action that maximizes a certain outcome they care about. This conditional – expected outcome given an action – is a forecast. All decisions are forecasts.

Until now, decisions have been made on vibes, expensive consultants, analysts, and forecasters whose job is to decide on how to convert this forecast into an actionable decision.

We are now reaching a point where AI systems are achieving parity with the best humans at forecasting the future (checkout some of Preseen’s past track record). This means that soon AI systems will be better than humans at making decisions.

The problems with forecasting so far is that it’s been tough to productize (good article on this). We believe that this is because forecasts have been too focused on predicting one off events, without being personalized to the action space of the organization that cares about the one off event.

There are two sides to these forecasts that Preseen is building towards.

• What action should we take to maximize the chance of a certain outcome? If the USA wants to limit Iran’s enrichment what action should they take?

• What action should we take in response to a certain outcome in the world? If the Strait of Hormuz stays closed, how should we adjust our supply chain?

Forecasting is going to be the decision-making layer of AGI.