I’m about as biased as possible here, but I had the same thought a few years ago while I was at a large hedge fund and recently founded a company building this.

Our goal is to create an “AI Brain” that is customized to each investor to help generate ideas, ramp up to speed, and make portfolio allocation decisions. The example you provided around buying down 20% is squarely on our roadmap.

Would love to chat sometime and hear more of your thoughts of how you’d like to use an AI in your process.


Expand full comment

sure, seems interesting. send me an email

Expand full comment

You're exactly right. I've mainly used ChatGPT to expand my circle of competence. I was researching the horse racing industry in Alberta for one of the stocks I'm researching, CNTY. I got on the learning curve quickly when it came to learning about the different breeds, lingo, and life spans of different horses, the ages at which each horse reaches its peak performance, and using this info to infer whether the incentive programs for breeders would have an immediate effect on the Alberta horse racing industry. I could imagine that using Google to find the same answers would take 3-4 times as long. 

I think what you're talking about is probably quickly approaching us.

Just like how Mohnish Pabrai has an extensive investment checklist. We could feed our checklist into an AI, and whenever we make an investment decision, the AI could be our personal assistant and keep us accountable to our checklist. It's definitely not a stretch for AI to analyze our winning and losing trades and try to find patterns that we aren't aware of. i.e. doubling down on losers too often and winners less often

Expand full comment

Would love to see a follow-up article on whatever you find. I would like to incorporate AI into my process but do not really know where to start

Expand full comment

What do folks here about BloombergGPT?

Expand full comment

Andrew, I’m always surprised when I hear people say they don’t see the potential. I say this as someone who has a finance education and work experience and then went into tech.

A good principe to consider is that most human activity can be “programmed” - stock picking for the most part is that.

Thought experiment: if you took all the historical data for a special stock (company A) and predicted it’s price trends and then fed real-time data about news and events about that company to the ai - at some point the ai will start to predict the news. This will work for day trading. Value investing is a different story.

Expand full comment