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

2027 memory capacity is reportedly sold out

ign.com

Post reply on HN
Print view View original

471–479 of 479 posts

Re: 2027 memory capacity is reportedly sold out

#471

Post by cortesoft »

EliRivers wrote: post #366
tjwebbnorfolk wrote: post #285

Earlier quoted context omitted.

When demand exceeds supply you have two options: make the people who value those goods the highest compete with each other to pay the natural market price for them, or have a lottery with price ceilings. Price ceilings create inefficiency, and they create black markets. In a lottery, the goods are misallocated to a bunch of people who aren't getting the most value out of the goods, creating economic inefficiency. A b…

This all sounds very hypothetical economic universe. In the real world, rich people do buy expensive things and do nothing with them.

Sure, at the individual level lots of people/companies will make irrational decisions. At the macro level, though, the effect is still there.

Re: 2027 memory capacity is reportedly sold out

#472

Post by globular-toast »

folkrav wrote: post #411
globular-toast wrote: post #401

Earlier quoted context omitted.

They didn't say anything about chronic pain, but with that extra context I do understand now.

Re-reading they indeed didn’t say “chronic pain”, that was me extrapolating. However I think it’s still relatively safe to assume “hand problems” prevalent enough to prevent them from typing code for many years to be both painful and chronic.

You're right. I misread it. I'd hate to lose my ability to type and jump at the chance to be able to use a computer again.

Re: 2027 memory capacity is reportedly sold out

#473

Post by kvemkon »

hombre_fatal wrote: post #283
adzm wrote: post #142

Earlier quoted context omitted.

electron, about:blank, 90mb Just tested

That's better than the GTK4 hello world I tried out last year.

GTK4 is 98 MB RSS on Linux. Or just 24 MB for GTK3.

example-0.c (fits both GTK3 and GTK4) from https://docs.gtk.org/gtk4/getting_started.html

Re: 2027 memory capacity is reportedly sold out

#474

Post by bilegeek »

SecretDreams wrote: post #373
noduerme wrote: post #332

Earlier quoted context omitted.

I mean, it's obviously that they think they're going way beyond LLMs to get to something that is profitable, like some ultimate goldmine of intelligence. But if they miss that target by a hair or two, can we buy some of that RAM? I'd like my next MacBook not to cost $10,000.

I'm going to be very pumped when this comes crashing down. It's going to look a bit like oligarchs subsidizing consumer tech as they try to offload more ram than God could own.

Everybody from OpenAI to Micron to the US Government is actively colluding to keep things going. I would love for you to be right, but I'm cynical from the decades that have been spent insulating these companies from consequences.

Re: 2027 memory capacity is reportedly sold out

#475

Post by rickydroll »

globular-toast wrote: post #316
rickydroll wrote: post #99

Earlier quoted context omitted.

Unfortunately, I can't not use AI because it provides the best speech recognition I've ever used and, going along with hand problems, it now allows me to write code after a three-year absence from that skill. I do miss it. I'd never do it again professionally in the large, but I have missed writing code a lot. My solution is to see what four or five-year-old equipment I can buy that will let me run local LLMs. I may…

You can not use it. You spent your life minus a couple of years not using it. You could totally get back into coding but LLMs have you trapped like a tethered elephant.

Speech recognition got the number of years wrong. It's more like 30 years. I was forced to stop programming in 1994.

And hand problems was an aggregate thing. first chronic pain then fine motor control and tremors. The pain flares up again if I try to use my hands too much.

I don't see the llms have me trapped like a tethered elephant anymore than the other layers of entrapment that we live with. I would argue entrapment started when we learned the secret was banging rocks together.

Or did you mean some other kind of entrapment?

Re: 2027 memory capacity is reportedly sold out

#476

Post by rickydroll »

folkrav wrote: post #411
globular-toast wrote: post #401

Earlier quoted context omitted.

They didn't say anything about chronic pain, but with that extra context I do understand now.

Re-reading they indeed didn’t say “chronic pain”, that was me extrapolating. However I think it’s still relatively safe to assume “hand problems” prevalent enough to prevent them from typing code for many years to be both painful and chronic.

You're right, and problems were chronic and painful and as I said elsewhere, have mutated over the past 30 odd years. Speech recognition error misrecognized 30 years as 3 years.

Re: 2027 memory capacity is reportedly sold out

#477

Post by hombre_fatal »

kvemkon wrote: post #473
hombre_fatal wrote: post #283

Earlier quoted context omitted.

That's better than the GTK4 hello world I tried out last year.

GTK4 is 98 MB RSS on Linux. Or just 24 MB for GTK3. example-0.c (fits both GTK3 and GTK4) from https://docs.gtk.org/gtk4/getting_started.html

Exactly. Though I was seeing 130 MB PSS when I did the same experiment last year.

Not sure if it had something to do with my env (NixOS), but either way it was a turn off since I see 20-30 MB with GTK3 which is about the size of my full-featured macOS AppKit apps.

Re: 2027 memory capacity is reportedly sold out

#478

Post by NekkoDroid »

esskay wrote: post #333
tiotempestade wrote: post #329

Hoping all these DCs and AI whatnots turn out to be a flop, so that we’ll get plenty ddr5 at sale price :D

IIRC this wont be possible as a lot of them aren't being packaged on traditional DDR memory boards. They're being put on things like the MI350X. I'm sure there will be a ton of standard DDR5 sticks, but not so much from AI-only datacenters.

> They're being put on things like the MI350X.

Maybe we can get a few MI350P's out of that then :) Surely it will be easy to do :clueless:

On a more serious note, I do wonder if it would theoretically be possible to put HBM on something like CAMM instead of regular DDR? I have not done any research on this.

Re: 2027 memory capacity is reportedly sold out

#479

Post by rickydroll »

tjwebbnorfolk wrote: post #199
theLiminator wrote: post #194

Earlier quoted context omitted.

Yeah, local LLMs are probably an order of magnitude less efficient at a fixed level of "intelligence" if not more.

Inefficient per watt, yes, but local inference capacity is greatly underutilized in aggregate. If a model can run on a machine that already exists, that's a bunch of additional chips that don't need to be built.

You picked up on what I didn't write existing machines don't have to be built, you just use them.

It occurred to me much later that local machines could become a shared resource of a small co-op. Or it only runs off solar or energy harvested at times of excess generation (too much daytime solar, lots of wind, or overnight when demand naturally drops).

Exploring how to partition inference across many machines and shed load for idle sections of hardware would be a problem that could be solved by a community of smart people like us here

Post reply on HN