Hacker News new | ask | show | jobs
by bardfinn 1106 days ago
The AI “gold rush” has already left the station, yyeeeaarrsss ago. The people who are going to monetize it are the people for whom the threshold of entry to the data sets is an insignificant threshold even at tens of thousands of dollars a year - and they’re the same entities already owning rights to datasets. And don’t need Reddit’s.

They could Elsevier a bunch of researchers who get institutional grants, but that’s not going to be a predictable, dependable, forecastable revenue stream.

For anyone with experience in pricing data services for partners, and anyone with experience with auditable, publicly owned businesses (like Reddit is trying to become, through IPO)

the API TOS & the “premium” firehose pricing is about having all the costs and cost centres of their business operations either justified by a revenue stream, or having a potential revenue stream assigned to those cost centres, or having a business case justifying the existence of that cost and cost centre, OR having a pricing model on that aspect of business which justifies it only being operated as a distinctly separate business model from the main business.

If that happens, then they’d likely spin it off as a subsidiary business, or a licensor, and operate it in a way that locks down their liability from privacy violation cases or GDPR violation cases etc.

In short: Reddit isn’t doing this for an “AI Gold Rush”. Reddit is doing this because they have to structure their business operations in a way that limits liabilities, minimizes or justifies cost centres, and maximizes profit streams.

To get people to buy the stock. To show that the company is profitable and is leveraging all their assets.