Andy Halliday

Pharmacists, plant scientists, HR specialists, they have 4,000 such domain experts.

Because those user profiles for all of their experts are tied directly to university registrars, which was a requirement of the process of taking .edu real candidate for interns over there, they can claim that they have validated sources.

And now AI is being trained at OpenAI by that company and those experts.

And why I say they're digging their own graves is that not for themselves individually, but for that class of profession, they're basically giving up the goods, which is, you know, I mean, I'm not resistant to that idea in general.

I think AIs won't completely and substantially replace human experts in the near term.

But over time, you can see how

Better and better training is happening with the human reinforcement based on real expertise validated by this process in the handshake company and open AI is leading the way in that.

I see that.

I see that as the optimistic view of the outcomes of, of supporting the training of AI and,

And I think there's really no reasonable way that you could effectively create resistance to that because there's somebody out there who has the same expertise that you have who's going to take $50 an hour to give up.

And what are they giving up?

What specifically the handshake platform does is it has the contractor, regardless of profession, simulate full professional workflows in that domain of expertise.

And it's basically the framework for that simulation is the context of

What are the requirements?

What are the goals?

So context, goals, references, what things do you need to be able to turn to in order to accomplish the workflows?

And then what are the deliverables?

What's the format and what's the output of this professional workflow?

And all of those things get organized.

And now you can imagine AI that's done this across thousands of professions.