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

It Still Can't Do My Job: Four Years of Moving Goalposts (2022–2026)

publicznyprofil.github.io

Post reply on HN
Print view View original

111–120 of 143 posts

Re: It Still Can't Do My Job: Four Years of Moving Goalposts (2022–2026)

#111

Post by blooalien »

tavavex wrote: post #105
captainbland wrote: post #38

Earlier quoted context omitted.

Yeah this idea that the unique value proposition of humans is now our motor capabilities rather than our cognition is unnerving as someone with dyspraxia. Like, oh good, they've figured out how to convert it into a much more limiting disability by commoditising apparently most knowledge work. Great.

If all the AI booster dreams come true with nothing slowing them down, the value proposition for an average human will change to, in this order: 1. At least humans can still do highly specialized physical work 2. At least humans are cheap to run on low-value hard labor tasks 3. At least humans are desperate and highly expendable, in some types of work it's cheaper to 'use one up' and replace them with another than to…

> "4. Humans are useless for all work and therefore have no value"

The part that worries me is when some clown with way too much money and power and time on their hands accidentally really does tip a machine over into AGI and the machines figure this little detail out for themselves... You want "Terminator IRL"? Because that's how you get "Terminator / SkyNet IRL"...

Maybe we can live-stream it all on Pay-Per-View once it starts, to make the ultra-rich idiots who ended us all just a little bit richer before the end, as I'm pretty sure some folks out there actually would pay to watch live-stream feeds of the ensuing apocalypse right up until the machines come to kick their door in too.

Re: It Still Can't Do My Job: Four Years of Moving Goalposts (2022–2026)

#112

Post by inigyou »

dwroberts wrote: post #5

> Call me when it stops making things up. We haven’t moved past this yet

This! It got a much wider pool of templates to copy from, so now if you ask for a 3D web game it gives you a similarly boring game using three.js instead of failing entirely. It still has no imagination and still makes up nonsense all the time. The fundamental problems haven't improved.

Re: It Still Can't Do My Job: Four Years of Moving Goalposts (2022–2026)

#113

Post by bryanlarsen »

solumos wrote: post #50
an0malous wrote: post #4

You should add a parallel timeline of how many times AI CEOs have claimed their next model is too dangerous to release, AGI is months away, or some white collar job will be obsolete within 6 months. I don’t know anyone in the tech industry who thinks AGI will never happen or that software engineering and white collar jobs can’t be automated. We all read sci fi, you’re not some unique visionary for anticipating AI. Th…

I’ll go on record and say that AGI will never happen (in the next 50 years). I think that’s also the timeline for white-collar job automation that requires critical thinking.

The article is about goal post moving. Please provide a clear and concrete definition of AGI that we can judge your prediction by.

Re: It Still Can't Do My Job: Four Years of Moving Goalposts (2022–2026)

#114

Post by inigyou »

jenniferhooley wrote: post #69

Man the one-shot game that is genuinely good in 2027 is crazy. A good game typically takes around 3-8 years of development by multiple skilled people. Maybe 3 years or so for 1 guy that's super dedicated. Right now single prompt with Fable can get us a small protype of like 1 game mechanic that's not even remotely production ready. So this guy thinks it'll 100,000X in a year. AI boosters are something else.

It's like any time I ever wanted to make a game. "What if there was a game where you could drive around and shoot people?" Oh it turns out driving a square to throw circles at other squares isn't actually very fun. AI does the same.

Re: It Still Can't Do My Job: Four Years of Moving Goalposts (2022–2026)

#115

Post by inigyou »

tavavex wrote: post #102
suddenlybananas wrote: post #33

Earlier quoted context omitted.

What on earth does "IQ: Yes" mean at the end??

It's a common joke to reply to something with 'yes' that doesn't make sense for the question asked. When used in place of a number, it just means extremely high, as high as it will go, that no matter what number you need it's a 'yes', and so on.

I heard it with budgets first. If your budget is "yes", it means your boss says yes whenever you ask for money. By analogy if your AI's IQ is "yes" the AI can do anything you can think of. It's also possible to have a budget of "no" and this is much more common.

Re: It Still Can't Do My Job: Four Years of Moving Goalposts (2022–2026)

#116

Post by AndrewKemendo »

henry2023 wrote: post #67
AndrewKemendo wrote: post #31

Earlier quoted context omitted.

What’s your definition of AGI?

Strong AGI would be a system capable of terraforming a planet without any external intervention. I mean, we did it and there have been roughly 117 billion human beings with GI in all of existence.

[dead]

Re: It Still Can't Do My Job: Four Years of Moving Goalposts (2022–2026)

#118

Post by lnenad »

an0malous wrote: post #30
lnenad wrote: post #10

Earlier quoted context omitted.

Literally every CEO AI or otherwise claims bullshit about their product. It's their job. It's got nothing to do with people actually using something and not coming to terms with it having any value. If you've never seen a screwdriver it's a waste of time compared to nails.

“Everyone else is doing it” is a childish justification and there’s a line where this becomes fraud and a felony offense. Do you think Elizabeth Holmes should have been allowed to make up claims about their blood testing technology? Is it OK to grossly overestimate Claude’s capabilities when the US military starts using it to determine strike targets?

I'm not justifying anything nor am I arguing against it. I'm saying it's not a thing worth mentioning as it's not related to AI in any way. If one would try to find AI related negatives go for datacentres, chip shortage/price hikes etc... So many better arguments to choose than just saying "CEO's are lying!!1!".

Re: It Still Can't Do My Job: Four Years of Moving Goalposts (2022–2026)

#119

Post by lnenad »

shimman wrote: post #32
lnenad wrote: post #10

Earlier quoted context omitted.

Literally every CEO AI or otherwise claims bullshit about their product. It's their job. It's got nothing to do with people actually using something and not coming to terms with it having any value. If you've never seen a screwdriver it's a waste of time compared to nails.

It's their job to lie and defraud the public? Why don't you believe there shouldn't be consequences for lying and defrauding the public? Do you believe tech CEOs are above society and decency? Does their wealth make them immune from accountability?

Did I ever say any of those things?

Re: It Still Can't Do My Job: Four Years of Moving Goalposts (2022–2026)

#120

Post by AndrewKemendo »

Ukv wrote: post #110
ofjcihen wrote: post #106

Earlier quoted context omitted.

Sure, let me be blunt. Speed and cost are nothing without quality and quality is partially a product of accountability (not even considering the technical or logistical issues this is enough on its own. An AI cannot be held accountable. It does not desire to feed its family. Your counter argument was outside the context of this articles claims, specifically that programmers and other knowledge workers can be replaced…

> Speed and cost are nothing without quality Quality was what the hypothetical was assuming had reached parity, no? ("humans are as bad as AI", "If AI really is at human level quality/error rate", etc.) > accountability Why could the company not still take accountability? That's already the case for non-ML automated systems, some with high failure rates. As a customer I rarely if ever care about blame being pinned on…

Correct

For the set of [all tasks humans can do] there is a subset of [things automation can do at equvalent or better error rates].

And that’s the baseline the baseline is not some platonic ideal that is never reached the baseline for a business operator is cost per delivery /error.

The idea that existing human systems are optimized or otherwise “not broken”, is the key fallacy that people keep making.

Post reply on HN