Comment by echelon
I agree - they're winning big and booking big revenue.
If you discount R&D and "sales and marketing", they've got a net loss of "only" $500 million.
They're trying to land grab as much surface area as they can. They're trying to magic themselves into a trillion dollar FAANG and kill their peers. At some point, you won't be able to train a model to compete with their core products, and they'll have a thousand times the distribution advantage.
ChatGPT is already a new default "pane of glass" for normal people.
Is this all really so unreasonable?
I certainly want exposure to their stock.
> If you discount R&D and "sales and marketing"
If you discount sales & marketing, they will start losing enterprise deals (like the US government). The lack of a free tier will impact consumer/prosumer uptake (free usage usually comes out of the sales & marketing budget).
If you discount R&D, there will be no point to the business in 12 months or so. Other foundation models will eclipse them and some open source models will likely reach parity.
Both of these costs are likely to increase rather than decrease over time.
> ChatGPT is already a new default "pane of glass" for normal people.
OpenAI should certainly hope this is not true, because then the only way to scale the business is to get all those "normal" people to spend a lot more.