Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Will I ever own a zettaflop?

geohot.github.io

Post reply on HN
Print view View original

51–60 of 98 posts

Re: Will I ever own a zettaflop?

#51

Post by bitwize »

mysecretaccount wrote: post #36

Fun post, but I find the industry's obsession with compute to be rather vapid, and this is a good example: > One million Claudes. To be able to search every book in history, solve math problems, write novels, read every comment, watch every reel, iterate over and over on a piece of code until it’s perfect – spend a human year in 10 minutes. 50,000 people working for you, all aligned with you, all answering as one. We…

Agentic development allows one Claude, multiplexed a few times, to vastly improve its output and tackle much bigger problems than just prompting the one instance. If you had a million Claudes in layered networks like we do with matmuls to form Claude, you'd be really cooking with gas.

(Maybe that's why they call it Gas Town?)

Re: Will I ever own a zettaflop?

#53

Post by danw1979 »

userbinator wrote: post #17
jmyeet wrote: post #8

I'm a big believer that humanity's future is in space in a Dyson Swarm. There are simply too many advantages. It's estimated that humanity currently uses ~10^11 Watts of power. About 10^16 Watts of solar energy hits the Earth but the Earth's cross-section is less than a billionth of the Sun's total energy output. A Dyson Swarm would give us access to ~10^25 Watts of power. With our current population that would give…

Dyson Swarm sounds like the name of an aggressive cleaning machine.

It's 2071. James Dyson, now 124 and in better health than ever, thanks to the AI-fuelled nanorobotic revolution, has just lost control of the last of the company's Dyson Swarms. What started as a fleet of cleaning-nanites, a dirt and dust-eating squadron-for-hire, has gone rogue; all of Earth's organic matter is now on the menu. People still haven't forgiven him about Brexit.

Re: Will I ever own a zettaflop?

#54

Post by throwaway198846 »

kubb wrote: post #44

Human appetite for more knows no bounds. Imagine what we’d have to do for everyone to have a zettaflop. We won’t have the resources for it. So guys like this one are in competition with normal people who just need a little bit of compute, so that he can feel powerful with a million Claudes. Sad.

I kind of agree with you there is only enough area for solar for 150 million people (even if we assume all land is solarable) but there is no reasons we couldn't eventually have everyone with a milion Claudes with fusion.

Edit: There is a problem with that - 10MW * 8 billion is something like half the solar power from the sun which implies wasted heat will heat earth considerably.

Re: Will I ever own a zettaflop?

#55

Post by ehnto »

eru wrote: post #39
nl wrote: post #10

Firstly, True Names is an awesome read, and the real origin of cyberpunk. I much prefer it to Neuromancer or Diamond Age. Secondly, I recently tried to work out what year on the Top500 list[1] I could reasonably be for around US$5000. It's surprisingly difficult to work out mostly because they use 64 bit flops and few other systems quote that number. [1] https://top500.org/lists/top500/2025/11/

Isn't the Diamond Age something like post-cyberpunk already? It came out three years after Snow Crash, which already ironically referenced "The sky above the port was the color of television, tuned to a dead channel". I agree that Neuromancer wasn't a great novel, though it obviously had vibes that resonated with many people. The novel being otherwise a bit of a dud actually speaks to how strong the vibes were to ove…

I feel that's a bit uncharitable, it wasn't just vibes, it was imaginative world building, with some truly interesting and novel concepts tied into a decent enough story to enjoy the world within.

As with much from this thread of cyberpunk writing, the cities and world are the most important characters, and the storyline is just an excuse to wander through their streets.

Re: Will I ever own a zettaflop?

#56

Post by mysecretaccount »

bitwize wrote: post #51
mysecretaccount wrote: post #36

Fun post, but I find the industry's obsession with compute to be rather vapid, and this is a good example: > One million Claudes. To be able to search every book in history, solve math problems, write novels, read every comment, watch every reel, iterate over and over on a piece of code until it’s perfect – spend a human year in 10 minutes. 50,000 people working for you, all aligned with you, all answering as one. We…

Agentic development allows one Claude, multiplexed a few times, to vastly improve its output and tackle much bigger problems than just prompting the one instance. If you had a million Claudes in layered networks like we do with matmuls to form Claude, you'd be really cooking with gas. (Maybe that's why they call it Gas Town?)

The output undoubtedly improves when looping LLM output back into the model at inference-time, but there is a limit to this and it is still bounded by the acumen of the underlying model. You cannot just recurse these models with tooling and compute to e.g. solve new physics.

Re: Will I ever own a zettaflop?

#57

Post by exe34 »

kubb wrote: post #44

Human appetite for more knows no bounds. Imagine what we’d have to do for everyone to have a zettaflop. We won’t have the resources for it. So guys like this one are in competition with normal people who just need a little bit of compute, so that he can feel powerful with a million Claudes. Sad.

Electron apps will expand to fill the available space!

Re: Will I ever own a zettaflop?

#58

Post by Nevermark »

throw0101d wrote: post #7

Somewhat related, why the creators of Zettabyte File System (ZFS) decided to make it 128 bits (writing in 2004): > Some customers already have datasets on the order of a petabyte, or 2^50 bytes. Thus the 64-bit capacity limit of 2^64 bytes is only 14 doublings away. Moore's Law for storage predicts that capacity will continue to double every 9-12 months, which means we'll start to hit the 64-bit limit in about a deca…

> In particular, it has been shown that 1 kilogram of matter confined to 1 liter of space can perform at most 10^51 operations per second on at most 10^31 bits of information

I believe the Bekenstein bound for holographic information on a 1 liter sphere, using space at the Planck scale for encoding, instead of matter, is about 6.7×10^67.

I confess I got that number by taking round trips through multiple models to ensure there was a clear consensus, as my form of "homework", as this is not my area of expertise.

As far as figuring out energy or speed limits for operations over post-Einsteinian twisted space, that will require new physics, so I am just going to wait until I have a 1 liter Planckspace Neo and just measure the draw while it counts to a very big number for a second. (Parallel incrementing with aggregation obviously allowed.)

Point being, there is still a lot of room at the bottom.

Interesting thought. Space can expand faster than the speed of light over significant distances, without breaking the speed limit locally.

But what happens if complex living space begins absorbing all the essentially flat local space around it? Is there a speed limit to space absorption? If space itself is shrinking, due to post-Einsteinian structures/packing, then effective speed limits go away. As traversal distances, and perhaps even the meaning of distance, disappear. So, perhaps not. I call this the "AI Crunch" end-of-the-universe scenario.

That is the computer I want. And I believe that sets a new upper bound for AI maximalism.

Re: Will I ever own a zettaflop?

#59

Post by danpalmer »

arthurjj wrote: post #14

I just want to thank the submitter. This is the type of internet that I really miss. A very smart person who's a good writer, proud of their interests and obsessions.

Geohot is famous for not being as smart as he makes out. Famously said he'd go to Twitter when Musk bought it and help Musk fix search, because "how hard can it be". Then left in shame 3 months later having achieved nothing except figuring out that It's A Bit More Complicated Than That(tm).

Comma does some cool stuff, if relatively entry-level, and this post is good napkin-maths and was a fun read, but there is so much more depth and a hundred ways in which this post is wrong or over-simplified to the point of near irrelevance.

Re: Will I ever own a zettaflop?

#60

Post by ramon156 »

kombine wrote: post #45
kubb wrote: post #44

Human appetite for more knows no bounds. Imagine what we’d have to do for everyone to have a zettaflop. We won’t have the resources for it. So guys like this one are in competition with normal people who just need a little bit of compute, so that he can feel powerful with a million Claudes. Sad.

Poor mother Earth, this race is unsustainable. In order to satiate guys like geohot we are pillaging the natural resources, destroying ecosystems, fucking up the climate.

I think i know two or three other people who are much, much more to blame... putting it on one dude is weird.
Post reply on HN