Idk about the destroying their business model part. Running your models on your own is akin to running a media server, which most people seem to be fine paying $20/m for someone else to do.
Well, maybe if your own media server only showed videos in 1fps compared to hosted ones. You’re not comparing apples to apples here. You might argue that local inference speeds are acceptable, but that doesn’t make them comparably equivalent.
What? Local inference is slow so people will want to self-host? I don't understand your argument.
AI companies can’t sustain this bleeding of runway. Eventually they’re all going to have to significantly bump up pricing. I guarantee personal accounts will be between double and triple your current internet and phone plan combined
That is not going to convince people to self-host models. The AI companies effectively have infinite money. Even if they did somehow run out, the US government would bail them out just like they did the car manufacturers.
People have been saying this for three years. When will it happen?