Uber was a literally life-changing product with an obvious value for anyone. LLMs have neither benefit.

Every cafe, airport, school I've been to has people using ChatGPT or its competitors. Its obviously valuable for almost anyone. Just like how people cant imagine life before smartphones, people wont be able to imagine life before LLMs became ubiquitous. Its everywhere.

True, but there isn’t really a moat for text LLMs. Llama is open source, Gemini is basically free.

define "moat", Open source search engines and competitive scraping strategies were built in the early 2000s. The Google search moat has always been that they were better for 5-6 years - and then they were the default.

Once they were the default option, they merely had to be the "same" as other options. If someone were to make a 1:1 clone of google with ~5% fewer ads - I do not believe they would break get a substantial market share of web search traffic (see bing).

Google has decades of people making it their habit as a pretty strong moat. They have inertia that chap bought simply do not yet. People have been using ChatGPT for a year or two now, it’s much easier for them to switch to another one.

I’m sorry, but what? How can you tell people at the cafe/airport/school are using LLMs?

You just walk around and see chatgpt interface on their screens?

Uber's spending was directly attributed to growth. They were launching in new countries, new cities, new markets every day, and that required burning through an immense amount of money. Of course that growth didn't need to last forever, and once the service was fairly established everywhere the spending stopped.

OpenAI on the other hand has to spend billions to train every new iteration of their model and still loses money on every query you make. They can't scale their way out of the problem – scaling will only make it worse. They are counting on (1) the price of GPUs to come down in the near term or (2) the development of AGI, and neither of these may realistically happen.