We’re in the good ole days

“I wish there was a way to know you’re in the good old days before you’ve actually left them.”

Andy Bernard, The Office

There’s a bunch of different places where I can take this catchy opening.

Let’s start with AI valuations.

The beauty of right now is that there’s uncertainty. And so every company gets to pitch a story to investors. Let’s talk about every layer of the stack, from the build-out to the application layer.

At the model layer, the story is that Anthropic and OpenAI are going to win, and they’ll start to compete away every single industry. This then justifies 2 trillion dollar IPO prospects because their entire pitch is that they will “eat” the entire economy.

At the application layer, companies like Rogo, Harvey, and Basis will say, no, look, people want to be model agnostic. They don’t want to be locked into one provider. Therefore, value accrues with us because we are providing the harness, the workflow, and the platform for you to own your intelligence. So Rogo gets to raise, or Harvey gets to raise, at ten times next year’s revenue. (By the way, these revenues are negative gross margin.)

At the inference layer, providers claim that they are going to need so much demand and compute, and therefore they need to be front-running this demand and raising mass amounts of debt and equity.

Bottom line is, it is a free-for-all out there. With one fact that we can all be sure about, nobody knows what the future holds. It is the golden age of venture valuations because everyone with a story or vision of the future can go and raise against that story.

II

While this is the golden age of venture-backed valuations and euphoria for the build-out, it also happens to be the best time to be building with these tools.

I was inspired by Steven Strogatz’s interview on The New York Times’ The Daily. He described being brought to tears as he grappled with what AI could mean for mathematics. While fighting back tears, Strogatz explains that part of the joy of math was the possibility of being the first person to understand something and explain it to the world. With AI, he feared losing that sense of purpose. He wonders out loud, what happens when you spend your life working toward a discovery, and a machine gets there first?

In a separate interview with WIRED, he asked:

“Is that the future of mathematics, that it’s a beautiful game that we are second rate at, but we still enjoy it?”

We are so lucky that we get to be building alongside these tools and explaining the breakthroughs. Humans, I know this next part sounds silly, are still in charge of the majority of the breakthroughs we see every day. How amazing is that!

And we very well might stay in this equilibrium where it’s exciting for you or me to go and really use these tools and hammer out the workflow and build something special that we wouldn’t have been able to build before.

But, there’s a world in which 10 or 20 years from now, none of that is novel anymore. Instead, the models are already better than you at any of this stuff. There will be far less skill or learning needed. And human discovery will become almost extinct.

It becomes a much more depressing society, where people are going to seek the satisfaction of learning and discovery only to potentially realize that there is nothing left to discover (at least by a human).

III

We don’t know how long this moment lasts. Right now, everyone gets to tell a story about the future and we still get to help build it or dare I even say discover it.

Maybe we’ll look back and miss how open it all felt. How much there was to figure out and how much of it was still ours to do.

Point of all this is to just call out that we might get to recognize the good old days while we’re still in them.