Zack Kass

That company was called Lilt.

And I stayed there long enough to graduate to OpenAI.

And so I joked that I just, at a time where everyone was making money hand over fist in things like cloud storage or e-commerce or sales optimization software, I just stuck around in AI long enough for it to pay off.

And at OpenAI, for a long time, if you emailed sales or support out, you arrived at my inbox.

And that started to change as we started to build API products, it became more popular.

Of course, GPT 3.5.

Most people don't remember, but it was state-of-the-art and really impressive in many ways.

And then of course, ChatGBT, November 30th, 2022.

And what was particularly impressive about that is that I had felt like OpenAI should have arrived much sooner.

My case to anyone who would listen was that GBT 3.5 was a commercially viable or economically viable model that people should have been adopting much faster and weren't.

And it was publicly available in the API since June 2022.

So you had the stretch of about five months, four months, where people could have built ChatGPT and didn't.

And ChatGPT was, as I remind people, not a research breakthrough, it was an application breakthrough.

And it serves...

For me, this is an incredible reminder that the application layer matters so much.

You have to build things that people can simply use, otherwise you cannot change consumer or even enterprise behavior in a material way.

And that moment is indelible in certainly a lot of technological history, but also my personal history, because a lot of us went home that night

Not sort of thinking anything of it.

I mean, the purpose of it really was to create an experience so that a CEO could come to the website, use the product, and then tell their CTO that they should build something with the API.

And then of course what happened was hundreds of thousands and millions of downloads and then billions of downloads later.