Andy Halliday

Then the,

data labeling for AI training came along, and we've heard about the big company, Scale AI, that the CEO and founder of which now is heading up all of meta super intelligence thing, right?

So data, Scale AI, collecting data, being used as training for AI, data quality is really, really important.

Now, but in the development of agentic AI,

We need to understand what businesses and what different professions do and what the workflow processes are for those things.

And so Handshake AI was created as a separate business unit inside Handshake, the company that was doing internships.

Why?

Because Handshake, the parent business, had proven...

.edu credentials for thousands and thousands of interns where they can basically say, yeah, this person is really a PhD candidate at MIT doing this work.

And so I can tell you with certainty that if you hire that person, they can give you inside, you know, advanced information about this area that you want to train your AI on.

Okay.

So they pivoted the handshake company into AI data labeling in 2024.

And now it's revealed that their big client is open AI.

And so what they're doing is they are, they have about 4,000 professionals.

They're paying them depending on their area of expertise.

They're paying them between $50 an hour and $500 an hour to basically create

spittle their guts on what it is that they know about their jobs.

So they gave some examples, uh, domain experts, commercial pilots, like what is it like to be a commercial pilot?

And what are the things you have to know and learn and understand to give that to the agent agentic AI training so that if you want to build an, you know, a, uh,

Supporting AI assistant for a commercial pilot, you've got real professionals giving up their hard-earned experience and learning.