Seriously, if they have to resort to ad revenue then no one is getting those billions and billions invested back. I am not even sure ad revenue would cover the cost of delivering the prompt response.
Its a stupid way to get more money. It will turn people off of AI when the long term plan is to make trillions of dollars with Agentic AI. Its like being a Wallstreet broker earning 10M a year and working at Trader Joe’s on the weekends for supplemental income.
Yeah. Basically OpenAI and the like have been convinced that the more GPU you add the better AI will be. They don’t factor in diminishing returns. Which means all these companies are spending a fraction of a trillion dollars on GPU and data warehouses. The level of expenditure and debt they have taken on means that when (if) they turn a profit by 2030ish they need to be making back trillions to pay back all the debt and become as profitable as they are valued.
The only way to do that is Agentic AI and replacing most of the workforce. This is a pipe dream, at least for now when “AI” is just different iterations of LLMs. Maybe it will happen in the future, but there is no way to tell if it really will need all these GPUs for it to happen.
The actions of a company on the brink of a gen ai breakthrough…
Seriously, if they have to resort to ad revenue then no one is getting those billions and billions invested back. I am not even sure ad revenue would cover the cost of delivering the prompt response.
Or maybe they simply want to get even more money.
Its a stupid way to get more money. It will turn people off of AI when the long term plan is to make trillions of dollars with Agentic AI. Its like being a Wallstreet broker earning 10M a year and working at Trader Joe’s on the weekends for supplemental income.
Isn’t the whole situation with AI is kinda stupid right now? I mean the management, not the scientific progress.
Yeah. Basically OpenAI and the like have been convinced that the more GPU you add the better AI will be. They don’t factor in diminishing returns. Which means all these companies are spending a fraction of a trillion dollars on GPU and data warehouses. The level of expenditure and debt they have taken on means that when (if) they turn a profit by 2030ish they need to be making back trillions to pay back all the debt and become as profitable as they are valued.
The only way to do that is Agentic AI and replacing most of the workforce. This is a pipe dream, at least for now when “AI” is just different iterations of LLMs. Maybe it will happen in the future, but there is no way to tell if it really will need all these GPUs for it to happen.