Preferences

Revenue isn't profit

Countless examples of a lack of profit not spelling the doom of a company (uber, amazon, tesla).

Countless examples of companies that strive for profit too early, only to die

Amazon was profitable, they just prioritized scaling up so they reinvested all profits.
OpenAI is also very profitable if they deprioritize... scaling up.
A distinction that matters if the unit economics are bad, which nobody has full visibility over.
No matter the unit economics, commoditization and advancement of open models and small startups means that there is at almost a year or two to exploit competitive advantage. If scaling stops, the window to make a profit is extremely narrow.
Open models serve as a safeguard, not a competitor, to closed models.

They still cost billions to pre-train

Training costs will come down, at a tremendous pace.

They cost billions to train right now because people are willing to throw billions away to get to the goal first. Given more time, cheaper and more clever training methods will be found.

Which is an issue with unit economics.
Google made $35 billion in profit last quarter.

This item has no comments currently.