The problem with a "sure thing" (AI will improve the world) isn't that it's not going to happen; it's just not going to happen as you expect it to. Case in point, venture investments. Let's review:

They said: Amazon owns eCommerce. Amazon is made up of sellers. Let's buy the sellers. (Alternatively: eCommerce is on a tear. Old brands are tired. D2C is the future.)

What happened: Most of the proceeds went to people building custom technology that was available off the shelf. The rest went to Amazon, Google, Meta's advertising coffers, proving yet again that retail is hard and margins are slim.

Artificial Intelligence is out to change all that however - right? Right? After the crypto-boondoggle (remember when investors thought that was the next big thing?) investors have universally zeroed in on AI.

Now they say: The world is inefficient. AI will triple our revenue and dramatically reduce our costs.

What's going to happen: AI will most definitely incrementally shave costs. A lot of the revenue gains will not materialize. Many of the models being built are not proven, and need to be constantly recomputed. As a result, a lot of money will flow to OpenAI, Microsoft, Amazon, Snowflake, etc.

Good businesses will be created, but there does not need to be so many investment firms chasing AI. Let's be clear:

* AI is not a disruptive investment thesis. It's a capability.

* A capability that the average startup has no key person advantage in. They are using the same tools everyone else has.

* You still need to create a profitable business that knows how to market and sell products and services to customers.

* Which is still tremendously difficult.

In short, getting to product-market fit is still the most expensive parts. And if VCs get involved before product-market fit in too many AI firms will fail unnecessarily because they "brute force" their way through.

* Look for businesses where the amount of model recompute is low because they aren't still searching for product/market fit.

* Look for businesses that already have more advanced sales and marketing motions. Who is their customer and who will it benefit?

* Look for businesses where the founders have an advantage not only in AI, but also in the industry they are building into. Beware imposters.

AI is great. Businesses should use it, and it will improve things. But that's very different from the "AI makeover" being the venture investable thing in every sector on the planet.