Startups and Category Theory

I miss the big data days.

For a few years I received an onslaught of pitches from startups in the space – sometimes “big” wasn’t even big enough and we’d get into superlatives to describe just how enormous it really was.

Then, quite suddenly, big data had its own K-T extinction event and the ML/AI mammals crept out of hiding to rule my inbox.

But if we step back there’s an interesting connection between the two: the build-out of big data primed the ability to do interesting things with that data. I think of this dynamic as one part of a slow march to a world that will be understood through category theory.

Category theory (CT) is sort of the Agamemnon of mathematics, a king-of-kings in the field. With CT it is possible to find patterns among whole realms of mathematics. There are many beautiful things you can do with CT – like proving a topological theorem such as Brouwer’s Fixed Point Theorem by porting it into algebra. Tai-Danae Bradley explains this much better than I could here.

My theory, very non-precisely and unmathed, is that the big data/AI dynamic can be thought of as focused on morphisms within a category and most companies are only concentrated on a limited slice of that. The next few years could be interesting if startups broaden their framework, and start to address IT with things like functors, which define the relationships between categories. (I say this with many apologies to real mathematicians for my inexact descriptions and worse).

A few other people have been interested in this topic – I know Conexus was evaluating early-stage companies in the CT space, but I’m not sure if they ended up making investments. I would also argue that Golden fits here with their self-constructing knowledge database.

If you are a startup or investor doing something interesting related to category theory I would love to chat!