4 ms·

Forget about the market reaction for a minutes, which is always difficult to interpret. More holistically, aren't we underestimating their dev ecosystems (CUDA

by gz5 2y ago

Forget about the market reaction for a minutes, which is always difficult to interpret. More holistically, aren't we underestimating their dev ecosystems (CUDA etc.)? Think of the dev ecosystem effects in all the various permutations, e.g. from Windows or Apple to AWS to more granular like PyTorch. Those advantages aren't easily eliminated just by more competitors or even 'better' competitors.

TheAlchemist 2y ago
Reactions to earnings are always a bit random. But holicistally, if you were to buy the entire Nvidia company today, even if those advantages continue untouched for the next 10 years, and competition doesn't catch up, and the AI boom continues ... you're still not getting your money back. That's how high the valuation is.
- p1esk 2y ago
  Do you assume the current growth rate?
jauntywundrkind 2y ago
Jim Keller calls CUDA NVidia's swamp (not their moat) and I think that's so so so accurate. Nvidia's fief has a lot of folks in it, who've built sloppy stuff all over the place because they can. Most of that won't get replaced. But the modern tooling like PyTorch and its underlying JAX doesn't need or care about this legacy. The future can build far more intimately & with deliberation to folks who haven't such large swamps around their ecosystems. We can tune systems more efficiently & carefully when our vendors don't keep us so so far away from the core hardware. Nvidia has to keep maintaining this huge footprint of software, keep it going themselves, supporting a vast legacy, while everyone else gets to innovate forward, with much better access to the hardware.
- Mc91 2y ago
  > Nvidia has to keep maintaining this huge footprint of software, keep it going themselves, supporting a vast legacy, while everyone else gets to innovate forward, with much better access to the hardware. When I hit an input box on my phone, a keyboard pops up on the screen, with the first row saying QWERTYUIOP, just like the Sholes and Glidden typewriter of 1878. I'm waiting for innovation on this first, then I'll look for innovation beyond CUDA.
  - nine_k 2y ago
    Wat. The keyboard you pop up this way will autocomplete your words, autocorrect your typos, and allow you to input whole words by swiping, twirling gestures. It's very much not your legacy typewriter keyboard, even though the order of the letters is the same. Maintaining the order of a couple dozen keys is not expensive or hard.
    - Mc91 2y ago
      > It's very much not your legacy typewriter keyboard, even though the order of the letters is the same. That random order of letters from 1878 still appearing on my screen is the definition of legacy.
- mFixman 2y ago
  I thought the opposite: if CUDA were easy to replace, wouldn't it have been replaced already? PyTorch supports both ROCm and Ibex, but it's support is spotty and incomplete to the point where it becomes a pain to use despite AMD spending billions on improving its software layer. NVIDIA is reaping a decade of design decisions on CUDA compatibility made less sloppily than its competitors.
Jlagreen 2y ago
Listen to earanings call, SW runtime is $2b for 2024 double of the $1b of last year. Nvidia Enterprise AI is spreading. People totally underestimate that and still think of Nvidia as a Semi company which only makes chips lol.

by gz5 2y ago

TheAlchemist 2y ago

Reactions to earnings are always a bit random. But holicistally, if you were to buy the entire Nvidia company today, even if those advantages continue untouched for the next 10 years, and competition doesn't catch up, and the AI boom continues ... you're still not getting your money back. That's how high the valuation is.

p1esk 2y ago
Do you assume the current growth rate?

jauntywundrkind 2y ago

Jim Keller calls CUDA NVidia's swamp (not their moat) and I think that's so so so accurate. Nvidia's fief has a lot of folks in it, who've built sloppy stuff all over the place because they can. Most of that won't get replaced. But the modern tooling like PyTorch and its underlying JAX doesn't need or care about this legacy. The future can build far more intimately & with deliberation to folks who haven't such large swamps around their ecosystems. We can tune systems more efficiently & carefully when our vendors don't keep us so so far away from the core hardware. Nvidia has to keep maintaining this huge footprint of software, keep it going themselves, supporting a vast legacy, while everyone else gets to innovate forward, with much better access to the hardware.

Mc91 2y ago
> Nvidia has to keep maintaining this huge footprint of software, keep it going themselves, supporting a vast legacy, while everyone else gets to innovate forward, with much better access to the hardware. When I hit an input box on my phone, a keyboard pops up on the screen, with the first row saying QWERTYUIOP, just like the Sholes and Glidden typewriter of 1878. I'm waiting for innovation on this first, then I'll look for innovation beyond CUDA.
- nine_k 2y ago
  Wat. The keyboard you pop up this way will autocomplete your words, autocorrect your typos, and allow you to input whole words by swiping, twirling gestures. It's very much not your legacy typewriter keyboard, even though the order of the letters is the same. Maintaining the order of a couple dozen keys is not expensive or hard.
  - Mc91 2y ago
    > It's very much not your legacy typewriter keyboard, even though the order of the letters is the same. That random order of letters from 1878 still appearing on my screen is the definition of legacy.
mFixman 2y ago
I thought the opposite: if CUDA were easy to replace, wouldn't it have been replaced already? PyTorch supports both ROCm and Ibex, but it's support is spotty and incomplete to the point where it becomes a pain to use despite AMD spending billions on improving its software layer. NVIDIA is reaping a decade of design decisions on CUDA compatibility made less sloppily than its competitors.

Jlagreen 2y ago

Listen to earanings call, SW runtime is $2b for 2024 double of the $1b of last year. Nvidia Enterprise AI is spreading. People totally underestimate that and still think of Nvidia as a Semi company which only makes chips lol.