• The idea is to get everybody dependent upon it and then jack the prices way up. Seems to be the standard MO these days.

    • It’s profitable for NVidia, because they’re not actually doing anything with it, they’re just selling all the hardware that everyone else needs.

      • The only ones guaranteed to make money in a Gold Rush are the ones selling shovels.

        That said, NVIDIA made the mistake of lending money and/or investing directly in the ones doing the “gold mining”, so now they’re heavilly exposed to the side of things that loses a lot of money when it turns out that it was all Fools’ Gold.

        When it comes to manias, some things are always the same.

      • Jensen seems to be concerned though about the bubble bursting. he probabl wont see immediate effect, but since he switched to mostly AI chip production/design, he will likely lose alot once it hits.

      • They kinda do, with their own models for upscaling and stuff integrated into videocard drivers, and it’s not cheap or easy, but it’s peanuts compared to what AI companies spend.

    • Honestly, replacing the C* suite is what IA would be best for!

      As PEOPLE they cost an exorbitant amount of money and stock options for exactly ZERO actual physical work.

      A group of AI agents, one for each of the C* positions, seeded with business based logic trees and equipped with exactly as much ethics subroutines as the board would allow would do the board bidding for 1/100000 of the cost of a human ceo.

  • Does anyone know what the graphs would look like if we removed the r&d cost?

    I wonder if it would be profitable once they stop spending like crazy on developing new models.

    • Based on the recently leaked OpenAI figures, still not profitable but it’s much closer. They would have lost $1.7B rather than $20.9B