I think what makes this worse is that their deficits are in the 10s of billions compared to most other startups being in the millions, even if the revenue to expenses ratio is better.

If I start a tech company and I spend 100 million dollars but only pull in 10 million in revenue, that's not great having a 10:1 expense to revenue ratio and 90 million in debt. In the leaked openAI financials from earlier this year, it was something like 13 billion revenue to 33 billion expenses, or about 3:1 expenses to revenue, which certainly seems better. However, the magnitude in actual costs is so much bigger at that point and the risk so much higher for investors. A billionaire can stand to lose on a 90 mill dollar investment, but it hurts a lot more on a 10 billion dollar investment.

And also, what's their path to profitablity? The premiere models are already expensive and getting more expensive. They'd have to charge significantly more to start breaking even, and people just won't pay for that. The cost to continue building and maintaining new DCs is t going to lessen the expenses either. I just think at some point investors are going to realize they aren't going to get an ROI on their investment and either cut back or drop out all together.

I definitely agree LLMs generate a lot of bloated code.

Look, LLMs are pattern identifiers. They identity patterns. They don't guess. Guessing is what humans do. They 'latch' onto a pattern. They don't learn patterns, they recognise them. Yes, is what studies have shown. LLMs don't learn gradually like Humans. They "latch" onto patterns. It's not sentient. It's just a machine learning algorithm, and if you think any of the AI that came before LLM is intelligent, then even LLMs are intelligent. If you think deep learning algorithms are just non linear matrix multiplications, they're exactly that.

LLMs coming for Coding jobs isn't because they're superhuman intelligent, its cause what most coders do is mostly same things. LLMs cannot do novel architecture, but if it's similar to 50 different softwares who's code it has trained on, yes, LLMs can so that.

LLMs are marvelous machines. They're not sentient, and only intelligent if you're willing to call every other ML algorithm intelligent. But the underlying statistics algorithm can and will keep getting better and/or more efficient, that was the point I was trying to make earlier.

Edit: actually you make some more good points. If LLM gets free reign over a project, they do make code unreadable. And Idk if that problem will be solved. It's just too much code they write, and efficiency is not something their training data gave them.