A Really Big Computer
geohot.github.io
A Really Big Computer
1–10 of 38 posts
Re: A Really Big Computer
#2Wyoming seems like a good place to build such a thing.
Re: A Really Big Computer
#3Maybe I'm overly critical, but if fixing the Twitter search box is too hard, maybe buying solar (and batteries), fabbing chips, and building a giant data center to replicate the current generation of what another company already built is foolhardy.
Re: A Really Big Computer
#4Maybe I'm overly critical, but if fixing the Twitter search box is too hard, maybe buying solar (and batteries), fabbing chips, and building a giant data center to replicate the current generation of what another company already built is foolhardy.
I armchair-diagnose geohot with bipolar disorder.
Only a few weeks ago, he was writing about how he was slated to lose because of his low tolerance for bullshit.
Re: A Really Big Computer
#5He doesn't mention yield, which I hear is currently pretty bad on 3nm. And "just route around the bad silicon" might be easier said than done; widespread defects might create too many roadblocks to efficiently move data around your wafer-chip.
If that's not factored into his 3450 wafer estimate, it could be double that.
Re: A Really Big Computer
#6Maybe I'm overly critical, but if fixing the Twitter search box is too hard, maybe buying solar (and batteries), fabbing chips, and building a giant data center to replicate the current generation of what another company already built is foolhardy.
I armchair-diagnose geohot with bipolar disorder. Only a few weeks ago, he was writing about how he was slated to lose because of his low tolerance for bullshit.
I am not going to try to come up with an underlying reason but he's really poisoned his reputation in my book. How can you so emphatically jump into a problem space that (in hindsight) you're not an expert at and then manage to distance yourself so vigorously from your failure like a month later? What am I supposed to think when you're faced with a real world hundreds-of-millions-of-dollars project spanning tons of large, complex industries? How am I supposed to believe you'll have better success with this than a few thousand lines of JavaScript?
Re: A Really Big Computer
#7I will refer to this as the moment when LLM hype jumped the shark.
Re: A Really Big Computer
#8I will refer to this as the moment when LLM hype jumped the shark.
AI is not just LLMs.
As I keep mentioning on HN.
AI is real results + hype as opposed to crypto of only hype.
A computer of this size will be immensely useful for training Vision models very quickly, as well as other kinds of models.
Why only AI? Any branch of Science (all of them) that benefit from fast, parallel compute, will benefit from this.
It is extremely narrow view to view AI only as LLMs.
When I say "real use" I mean solving real world problems that existed before AI. And I don't even count generative AI of any shape or form.
Re: A Really Big Computer
#9He doesn't mention yield, which I hear is currently pretty bad on 3nm. And "just route around the bad silicon" might be easier said than done; widespread defects might create too many roadblocks to efficiently move data around your wafer-chip. If that's not factored into his 3450 wafer estimate, it could be double that.
This was my first thought reading this article. I am not an expert and my knowledge is limited.
When he mentions the prices of the wafers, does it come with the bad yields or without?
Re: A Really Big Computer
#10How does he cool those wafers?