Hacker News new | past | comments | ask | show | jobs | submit login

Exactly. NVDA wasn't clearly a winner until the past 2 years or so. From Jan 1 2016 to Jan 1 2020, it grew by 8x -- certainly impressive (70% year-over-year growth). From there until the ChatGPT announcement in Nov 2022, it maybe doubled once more (peaked from the crypto-induced GPU shortage, then was falling). But from there on out, in the course of 20 months, it's skyrocketed 8x (an absurd 250% year-over-year growth).

So sure, if you correctly guessed that ChatGPT was going to spur a ton of interest in Nvidia hardware, then you could have made lots of money in not very much time. Meanwhile, to me at the time, this seemed like an incremental release on top of OpenAI's prior GPT models, none of which were earth-shattering paradigm shifts. I certainly did not anticipate the surge of all these AI startups that wanted to build on top of it, or the industry shift to try to use GenAI to solve all of the world's problems.

--

If I got the math wrong anywhere, it was the magnitude of investing that $1k in BTC back in 2011 -- I'd have $30-35 million to my name, minus taxes for long term capital gains. Even then, it wouldn't have been clear to me at what point I should sell -- mid 2017, when that investment would have grown to $1 million? After it peaked in December 2017 at $20k, it lost 80% of its value -- what reason would I have to expect that it'd grow to more than 3x its previous peak, just a couple years later?




The bet I described was that GPUs would be more and more relevant in modern computation, and that Nvidia sells the best GPUs and the best toolkit for writing general purpose code on GPUs. GPT is just the latest on a large chain of applications.




Consider applying for YC's W25 batch! Applications are open till Nov 12.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: