I believe that the May 2023 CNBC article used the 2021 valuation...

I believe that the May 2023 CNBC article used the 2021 valuation because there wasn't anything more recent. According to Pitchbook, Scale AI did a fund raise in April of 2023, but the details (including amount raised and valuation) were undisclosed. This often means that it was a down round.

Scale AI was considered special because they were the company that reaped, by far, the most benefit from the autonomous vehicle data push (a trend that Brayan missed completely). Their connections with the auto industry were unique (several advisors from auto) which allowed them to gain the lion's share of that business. They then tried to leverage that strength into other areas, with mixed success. And don't get to hung up on crowd size - in practice, there is little difference between 250K and 1M. Fewer than 30-40k workers are active at any time - and most of those positions were filled through recruiting, not the 'database'.

I wasn't trying to say that Appen wasn't doing GenAI work -- they definitely are. I was saying that several companies are doing GenAI now. There is a real risk that it could get commoditized -- as I've said before, the key differentiator of successful companies will be their recruiting capabilities, given the domain expertise required in GenAI.

I can see why, from an Operational standpoint, Deep Learning and GenAI would be together. The actual mechanics of the tasks are somewhat similar and can therefore leverage the same annotation tools, quality approaches, team management tools, etc. Appen needs to be very careful with their cash right now so leveraging operational synergies (this might be the most business-y sentence I have ever typed) makes sense to me.