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

Horses: AI progress is steady. Human equivalence is sudden

andyljones.com

Post reply on HN
Print view View original

401–410 of 663 posts

Re: Horses: AI progress is steady. Human equivalence is sudden

#401

Post by sotix »

marcus_holmes wrote: post #150
ible wrote: post #5

People are not simple machines or animals. Unless AI becomes strictly better than humans and humans + AI, from the perspective of other humans, at all activities, there will still be lots of things for humans to do to provide value for each other. The question is how do our individuals, and more importantly our various social and economic systems handle it when exactly what humans can do to provide value for each oth…

I'm optimistic. Banks used to have rooms full of bank clerks who manually did double-entry bookkeeping for all the bank's transactions. For most people, this was a very boring job, and it made bank transactions slow and expensive. In the 50's and 60's we replaced all these people with computers. An entire career of "bank clerk" vanished, and it was a net good for humanity. The cost of bank transactions came down (by…

> Banks used to have rooms full of bank clerks who manually did double-entry bookkeeping for all the bank's transactions. For most people, this was a very boring job, and it made bank transactions slow and expensive. > > And the people who had to do double-entry bookkeeping all day long got to do other, probably more interesting, jobs.

I don't mean to pick on your example too much. However, when I worked in financial audit, reviewing journal entries spit out from SAP was mind numbingly boring. I loved doing double-entry bookkeeping in my college courses. Modern public accounting is much, much more boring and worse work than it was before. Balancing entries is enjoyable to me. Interacting with the terrible software tools is horrific.

I guess people that would have done accounting are doing other, hopefully more interesting jobs in the sense that absolute numbers of US accountants is on a large decline due to the low pay and the highly boring work. I myself am certainly one of them as a software engineer career switcher. But the actual work for a modern accountant has not been improved in terms of interesting tasks to do. It's also become the email + meetings + spreadsheet that you mentioned because there wasn't much else for it to evolve into.

Re: Horses: AI progress is steady. Human equivalence is sudden

#402

Post by jakewins »

empiricus wrote: post #380
jakewins wrote: post #286

Earlier quoted context omitted.

I’m deeply sceptical. Every time a major announcement comes out saying so-and-so model is now a triple Ph.D programming triathlon winner, I try using it. Every time it’s the same - super fast code generation, until suddenly staggering hallucinations. If anything the quality has gotten worse, because the models are now so good at lying when they don’t know it’s really hard to review. Is this a safe way to make that sy…

I agree that the current models are far from perfect. But I am curious how you see the future. Do you really think/feel they will stop here?

I mean, I'm just some guy, but in my mind:

- They are not making progress, currently. The elephant-in-the-room problem of hallucinations is exactly the same or, as I said above, worse as it was 3 years ago

- It's clearly possible to solve this, since we humans exist and our brains don't have this problem

There's then two possible paths: Either the hallucinations are fundamental to the current architecture of LLMs, and there's some other aspect about the human brains configuration that they've yet to replicate. Or the hallucinations will go away with better and more training.

The latter seems to be the bet everyone is making, that's why there's all these data centers being built right? So, either larger training will solve the problem, and there's enough training data, silica molecules and electricity on earth to perform that "scale" of training.

There's 86B neurons in the human brain. Each one is a stand-alone living organism, like a biological microcontroller. It has constantly-mutating state, memory: short term through RNA and protein presence or lack thereof, long term through chromatin formation, enabling and disabling it's own DNA over time, in theory also permanent through DNA rewriting via TEs. Each one has a vast array of input modes - direct electrical stimulation, chemical signalling through a wide array of signaling molecules and electrical field effects from adjacent cells.

Meanwhile, GPT-4 has 1.1T floats. No billions of interacting microcontrollers, just static floating points describing a network topology.

The complexity of the neural networks that run our minds is spectacularly higher than the simulated neural networks we're training on silicon.

That's my personal bet. I think the 88B interconnected stateful microcontrollers is so much more capable than the 1T static floating points, and the 1T static floating points is already nearly impossibly expensive to run. So I'm bearish, but of course, I don't actually know. We will see. For now all I can conclude is the frontier model developers lie incessantly in every press release, just like their LLMs.

Re: Horses: AI progress is steady. Human equivalence is sudden

#403

Post by mrwrong »

marcus_holmes wrote: post #150
ible wrote: post #5

People are not simple machines or animals. Unless AI becomes strictly better than humans and humans + AI, from the perspective of other humans, at all activities, there will still be lots of things for humans to do to provide value for each other. The question is how do our individuals, and more importantly our various social and economic systems handle it when exactly what humans can do to provide value for each oth…

I'm optimistic. Banks used to have rooms full of bank clerks who manually did double-entry bookkeeping for all the bank's transactions. For most people, this was a very boring job, and it made bank transactions slow and expensive. In the 50's and 60's we replaced all these people with computers. An entire career of "bank clerk" vanished, and it was a net good for humanity. The cost of bank transactions came down (by…

> There are a ton of current careers that are just email + meetings + powerpoint + spreadsheet that can go the same way.

it's interesting how it's never your job that will be automated away in this fantasy, it's always someone else's.

Re: Horses: AI progress is steady. Human equivalence is sudden

#404

Post by littlestymaar »

billisonline wrote: post #50

An engine performs a simple mechanical operation. Chess is a closed domain. An AI that could fully automate the job of these new hires, rather than doing RAG over a knowledge base to help onboard them, would have to be far more general than either an engine or a chessbot. This generality used to be foregrounded by the term "AGI." But six months to a year ago when the rate of change in LLMs slowed down, and those exci…

> An engine performs a simple mechanical operation

It only appears “simple” because you're used to see working engines everywhere without never having to maintain them, but neither the previous generations nor the engineers working on modern engines would agree with you on that.

An engine performs “a simple mechanical operation” the same way an LLM performs a “simple computation”.

Re: Horses: AI progress is steady. Human equivalence is sudden

#405

Post by sanex »

jsheard wrote: post #18

Cost per word is a bizarre metric to bring up. Since when is volume of words a measure of value or achievement?

There was a time when these models were novel that if use it to write for me. After a year or so the verboseness and lack of personality got old. Now all I have is a decent proofreader. Maybe they'll take over my job but I'm finding the trend going the other way right now.

Re: Horses: AI progress is steady. Human equivalence is sudden

#406

Post by jjmarr »

danw1979 wrote: post #378
netsharc wrote: post #360

Earlier quoted context omitted.

I think, like the Bill Gates haters who interpret him talking about reducing the rate of birth in Africa as wanting to kill Africans, you're interpreting it wrong. The graph says horse ownership per person. People probably stopped buying horses, they let theirs retire (well, to be honest, probably also sent to the glue factory), and when they stopped buying new horses, horse breeding programs slowed down.

I wish the author had had the courage of their convictions to extend the analogy all the way to the glue factory. It’s what we are all thinking.

I have a modest proposal for dealing with future unemployment.

Re: Horses: AI progress is steady. Human equivalence is sudden

#407

Post by Herring »

maciejzj wrote: post #353

I may have developed some kind of paranoia reading HN recently, but the AI atmosphere is absolutely nuts to me. Have you ever thought that you would see a chart showing how population of horses was decimated by the mass introduction of efficient engines accompanied by an implication that there is a parallel to human population? And the article is not written in any kind of cautionary humanitarian approach, but rather…

American culture actively punishes compassion, then gaslights you about it.

https://www.census.gov/library/visualizations/interactive/te... Look at all the professions on the bottom right: Teachers, therapists, clergy, social workers, etc. It’s not a coincidence that cruel people take top positions.

Re: Horses: AI progress is steady. Human equivalence is sudden

#408

Post by myth_drannon »

maciejzj wrote: post #353

I may have developed some kind of paranoia reading HN recently, but the AI atmosphere is absolutely nuts to me. Have you ever thought that you would see a chart showing how population of horses was decimated by the mass introduction of efficient engines accompanied by an implication that there is a parallel to human population? And the article is not written in any kind of cautionary humanitarian approach, but rather…

It's been a decade or so but I'm mostly called "resource" at work, as in Human Resource. Barely collegue, comrade, co-worker... just a resource, a plug in the machine that needs to be replaced by an external resource to improve profit margins.

Re: Horses: AI progress is steady. Human equivalence is sudden

#409

Post by erichocean »

pbw wrote: post #56

This is food for thought, but horses were a commodity; people are very much not interchangeable with each other. The BLS tracks ~1,000 different occupations. Each will fall to AI at a slightly different rate, and within each, there will be variations as well. But this doesn't mean it won't still subjectively happen "fast".

That looks to me like there are ~1000 interchangeable economic human roles for AI to replace.

So I guess we should check to see if computers are good at scaling or doing things concurrently. If not, no worries!

Re: Horses: AI progress is steady. Human equivalence is sudden

#410

Post by 9rx »

ben_w wrote: post #340
9rx wrote: post #256

Earlier quoted context omitted.

Things like trains, boats, and cars exist. Human chess grandmasters can show up to elite tournaments, and perform while there, without airplanes. Computer chess systems, on the other hand, cannot do anything without human oversight.

> Things like trains, boats, and cars exist. Human chess grandmasters can show up to elite tournaments, and perform while there, without airplanes. Those modes of transport are all equivalent to planes for the point being made. I (not that I'm even as good as "mediocre" at chess) cannot legally get from my current location to the USA without some other human being involved. This is because I'm not an American and wou…

> Those modes of transport are all equivalent to planes for the point being made.

The point is that even those things require oversight from humans. Everything humans do requires oversight from humans. How you missed it, nobody knows.

Maybe someday we'll have a robot uprising where humans can be exterminated from life and computers can continue to play chess, but that day is not today. Remove the human oversight and those computers will soon turn into lumps of scrap unable to do anything.

Sad state of affairs when not even the HN crowd understands such basic concepts about computing anymore. I guess that's what happens when one comes to tech by way of "Learn to code" movements promising a good job instead of by way of having an interest in technology.

Post reply on HN