9mo ago

Why OpenAI Is A Terrible Business

Why OpenAI Is A Terrible Business - Better Offline

As of last week, OpenAI is now worth $157 billion - yet below the hood is a far darker story. In this episode, Ed Zitron explains the cold, hard truth - that OpenAI is a terrible business that burns billions of dollars, and its failure to scale its cloud business tells a dark tale about the wider generative AI industry.

5 comments

One of the points I find concerningis that businesses built on the current price point for the different models are going to get wiped out when they have to raise prices to keep the lights on.
- none of them can say they haven't been warned. It's common knowledge that they're haemhorraging money, it's only a matter of time before the investors start wanting some sort of return
  
  I just want to know what's going to happen to all the data centers being built to handle all of this needless LLM shit that almost no one wants.
- I think smaller businesses and the true believers are victims of OpenAI and its ilk, and don't necessarily deserve the fallout.
Current neural networks do really fancy statistics. To make the model better, you need to make the statistics more precise. Leading to marginal improvements of accuracy requiring exponentially growing marginal amounts of training data. This leads to exponentially decaying marginal utility coupled with exponentially growing marginal expense. Which quickly becomes unsustainable. Edit: On the plus side, this likely means you won't have to give up much utility when the market adjusts.