Understanding the LLM Bubble
americanaffairsjournal.org
americanaffairsjournal.org
Any time this quote pops up, credibility goes to zero. It’s virtually the most surface level understanding of what AI is and also what human intelligence is. Quoting outdated METR study (n=20, model=sonnet 3.x), claiming that no meaningful progress has happened in last couple of years (and then their citation footnote which I was pleased to see actually does not have anything to do with their statement).
Claiming Uber was basically a scam, running at a loss and investors just trying to cash out as soon as possible in an IPO (go look up the market cap of Uber today and tell me their business model was a scam).
But people eat this stuff up?
Anyone who has put a bit of thought into this knows that it’s outdated and that the study itself is flawed.
I think the author overcondifently makes claims and just cites things
To date, no firm has explicitly stated its gains from llm’s aside from ‘increasing proportion of LOC’.
If firms had incremental cash flows to show for it they would talk about it loudly and proudly. But they don’t.
The existing assets of meta, google et al pre-llm’s are taking them to new highs. The investments in llm’s to date are value destructive in terms of cash flows.
LLM investments “value destructive” makes no sense to me, can you clarify?
Read properly or don’t bother posting or admit to yourself that you don’t understand corporate finance+valuation and stop engaging in discussions around it.
* read properly
* don’t bother posting
* admit to yourself that you don’t understand
* stop engaging in discussions around it