OpenAI is currently the best in some major criterias, but not all. Gemini has much larger context window, for example.
And behemoths like Google/Meta/Apple can remain irrational with their AI spending longer than OpenAI can remain solvent. Unless Microsoft steps in, which they already did, kind of.
I guess the big problem with AI/ML of the LLM type is that the data and compute is the gold, or fuel if you will. How does one commercialize a product in the long term, when each iteration becomes magnitudes more expensive than the previous, and the previous products become almost worthless within 12 months.
Oh yeah, keep growing your number of subscribers. But that also has a natural limit.
Maybe at some point, these AI models will be so integrated and omnipresent in everyone's daily like that some kind of oligopoly or monopoly arises, and they can leverage that.
This is most Hollywood productions in a nutshell
Doesn't seem too outrageous
It's replacing a lot of my Google searches. I also see OpenAI used by people around me in the office.
So they definitely have first mover advantage and if they can keep up with competitors they can take a big chunk of the market.
To make a 10x return, OpenAI would need to be worth 1.6T.
So investors are thinking that there is a about ~10% chance of AI growing to about ~10% the size of the service industry (10% of service industries 16T/year is 1.6T, 10% chance => 160B).
While I'm not sure I agree, this math does actually make more sense than you'd think given the sticker-shock of 160B for a young company.
What I'm more curious about is the (as far as I know) non-public guesses as to which industries openai believes it will take bites out of. Software? Service? An entirely new industry space? Assuming its going to just grow productivity by itself, and not crush other industries, the valuation seems off. For $160B, the expectation is more likely "this will result in no longer needing hundreds of thousands of employees" - but are we replacing their salaries wholesale with OpenAI contracts? Doubtful.
Not to say that there aren't ways around it. But you'd also expect crazy valuations to be going up the supply chain. And TSMC is wildly undervalued if you believe in OAI's ideal scenario. Market is inconsistent here.
You can’t predict breakthroughs and key members of the OG engineering team left.