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

Everyone feared AI taking over; the real danger is AI serving just the few

news.ycombinator.com

Post reply on HN
Print view View original

81–90 of 97 posts

Re: Everyone feared AI taking over; the real danger is AI serving just the few

#81

Post by zcw100 »

appplication wrote: post #71
zcw100 wrote: post #58

I'm done. The cognitive dissonance over AI has jumped the shark. It's not fair that it's everywhere. It's not fair that only a few have it. It's not fair that it uses so much electricity. It's not fair that it's killing my environment. It's not fair there's a datacenter in my back yard. It's not fair there isn't a datacenter in my back yard. It's not fair that it threatens my job. It's not fair that my employer doesn…

I don’t think it’s as cognitively dissonant as you make it out to be, the dominating sentiment outside the corp world is quite negative. Yeah there are a few positive perspectives to be had, and not to mention complexity of perspective does not imply cognitive dissonance. I think you may be building up a straw man.

A large number of logically irreconcilable perspectives (in one's own mind) is the precise definition of cognitive dissonance. You also say that I may be building a straw man without even trying to point out where you think I've done that. You've already conceded my argument that there are many competing perspectives and I make no attempt to poke any holes in that so there is no straw man.

Re: Everyone feared AI taking over; the real danger is AI serving just the few

#82

Post by TurdF3rguson »

dragontamer wrote: post #30
TurdF3rguson wrote: post #27

Earlier quoted context omitted.

Nuclear weapons development hasn't stopped outside of US. Maybe you mean it hasn't spread to any new countries lately... which is true but I wouldn't count on that lasting.

Both USA and Russia, have decided that ~Megaton Hydrogen Bombs are the biggest we're going to get and we don't plan to build anything bigger than that ever again. The USSR wanted to make sure they were the ones who built the largest bomb of all time (the Tsar Bomba at 50 MTon). And after that, development on more ferocious weapons has stopped.

Right so now China, India, Pakistan all need to catch up. And France for some reason.

Re: Everyone feared AI taking over; the real danger is AI serving just the few

#83

Post by hodgehog11 »

KellyCriterion wrote: post #78
hodgehog11 wrote: post #12

Earlier quoted context omitted.

Because if spread equally and used responsibly, it raises quality of life for everyone. Generally speaking, the vast majority of people alive today have far easier lives than hunter-gatherers. There is less famine and starvation. There is less fatal disease, etc.

So how does quality of life increase overall for everyone, if Nvidia builds now HA200 instead of HA100? (or whatever they are called)

GPU innovation in the past led to the development of new artforms and allowed researchers to solve tougher computational tasks. Today, not so much. You're missing the "spread equally and used responsibly" part here. Nvidia's customers now are the data centers, which have debatable impact on quality of life.

Re: Everyone feared AI taking over; the real danger is AI serving just the few

#84

Post by pgryko »

jwally wrote: post #54

Maybe I'm not paying close enough attention (probably) - but I feel that China is providing a glimmer of hope that this (only the uber elite can get good AI assistance) isn't true, and we're maybe at a weird inflection point of sorts wrt the leverage that tech can now provide. China gets sanctions and stale chips - fine, they just DIY the algorithms through CPU instead of GPU and open source it. American warships hav…

[dead]

Re: Everyone feared AI taking over; the real danger is AI serving just the few

#85

Post by robocat »

jwally wrote: post #54

Maybe I'm not paying close enough attention (probably) - but I feel that China is providing a glimmer of hope that this (only the uber elite can get good AI assistance) isn't true, and we're maybe at a weird inflection point of sorts wrt the leverage that tech can now provide. China gets sanctions and stale chips - fine, they just DIY the algorithms through CPU instead of GPU and open source it. American warships hav…

  NASA wanted to avoid pencils because the lead could easily break off and float away, creating a hazard to astronauts and sensitive electronics on the spacecraft In fact, a pencil is such an impractical alternative in space that cosmonauts also have been using Space Pens since 1969
https://spinoff.nasa.gov/space-pens

You don't want conductive graphite particles shorting out electronics nor do you want flammable dust or wood flakes.

Good engineers learn to smell out simple but wrong answers.

Re: Everyone feared AI taking over; the real danger is AI serving just the few

#87

Post by lr1970 »

robocat wrote: post #85
jwally wrote: post #54

Maybe I'm not paying close enough attention (probably) - but I feel that China is providing a glimmer of hope that this (only the uber elite can get good AI assistance) isn't true, and we're maybe at a weird inflection point of sorts wrt the leverage that tech can now provide. China gets sanctions and stale chips - fine, they just DIY the algorithms through CPU instead of GPU and open source it. American warships hav…

NASA wanted to avoid pencils because the lead could easily break off and float away, creating a hazard to astronauts and sensitive electronics on the spacecraft In fact, a pencil is such an impractical alternative in space that cosmonauts also have been using Space Pens since 1969 https://spinoff.nasa.gov/space-pens You don't want conductive graphite particles shorting out electronics nor do you want flammable dust o…

> You don't want conductive graphite particles shorting out electronics nor do you want flammable dust or wood flakes.

For decades mechanical pencils use non-conducting polymer based refills that eliminate the problem you mention.

Re: Everyone feared AI taking over; the real danger is AI serving just the few

#88

Post by softwaredoug »

I don’t worry about access to amazing AI. GLM 5.2 is pretty amazing. I can easily see it becoming a norm to own beefier personal hardware to run your own AI.

I worry, like other emerging industries, the administration will make hamfisted decisions that let the Chinese take the lead. See also EVs / clean energy and the war in Iran which has only accelerated that transition.

There’s currently cronyism. There’s no laws that give companies certainty about a regulatory framework (it’s all just being made up). Add to that deep suspicion of expertise at the White House. And a huge disconnect from the economic reality.

It’s a great opportunity for the second tier US labs to release open weight almost-as-good models. Meta kicked off this race, maybe they can release something? Microsoft wants to develop more of their models. Maybe Google should go open weight?

Re: Everyone feared AI taking over; the real danger is AI serving just the few

#89

Post by JeremyNT »

jwally wrote: post #54

Maybe I'm not paying close enough attention (probably) - but I feel that China is providing a glimmer of hope that this (only the uber elite can get good AI assistance) isn't true, and we're maybe at a weird inflection point of sorts wrt the leverage that tech can now provide. China gets sanctions and stale chips - fine, they just DIY the algorithms through CPU instead of GPU and open source it. American warships hav…

> Maybe I'm not paying close enough attention (probably) - but I feel that China is providing a glimmer of hope that this (only the uber elite can get good AI assistance) isn't true, and we're maybe at a weird inflection point of sorts wrt the leverage that tech can now provide.

I think it's easy for China to be more open right now (and Western governments to essentially not worry about it yet) because they're playing catchup.

If we reach some endpoint where American models stop improving for the general public and only party apparatchiks and sycophantic oligarchs can access them, the situation is different. Chinese models could exceed the publicly available SOTA, and surely they would close off their own models in the same way.

Re: Everyone feared AI taking over; the real danger is AI serving just the few

#90

Post by robocat »

lr1970 wrote: post #87
robocat wrote: post #85

Earlier quoted context omitted.

NASA wanted to avoid pencils because the lead could easily break off and float away, creating a hazard to astronauts and sensitive electronics on the spacecraft In fact, a pencil is such an impractical alternative in space that cosmonauts also have been using Space Pens since 1969 https://spinoff.nasa.gov/space-pens You don't want conductive graphite particles shorting out electronics nor do you want flammable dust o…

> You don't want conductive graphite particles shorting out electronics nor do you want flammable dust or wood flakes. For decades mechanical pencils use non-conducting polymer based refills that eliminate the problem you mention.

Polymer leads were invented by the Japanese late 50's.

Pentel use manga (漫画 whimsical/informal drawings) to write the history of a burnt rice ball being the inspiration: https://pentel.ca/docs/e_manga.pdf (page 7 with the title "Pentel's Sharp Pencil Refills").

Previously there were clutch pencils using the historical fired clay and graphite https://www.bl.uk/stories/blogs/posts/the-history-of-the-pen...

As usual Germans claim inventing it first https://www.computerhistory.org/siliconengine/the-european-t... ). Why are there no teutonic Space Pencil jokes?

Post reply on HN