AIs gunning for our precious freelancers

[malloc]

If you focus on specific cognitive skills rather than intelligence in general, then the picture becomes even more bleak. Artificial intelligence has mastered numerous mental skills over the years and continues to learn new ones everyday as researchers invent new models. Meanwhile humans have not acquired any new cognitive abilities in thousands of years because the structure of the brain depends on genetics and we cannot simply give ourselves cognition-enhancing mutations. That means that AI will eventually master every skill available to humans and even develop skills we could never hope to achieve.

[Raphael]

That means that AI will eventually master every skill available to humans and even develop skills we could never hope to achieve.

AI doesn't master anything (at least at the current point in history - 100 or 1000 years in the future might be a different matter). AI designers might "master" various things. You keep talking about the machines doing this and the machines learning that. No, people who built or work with machines do this or learn that.

[malloc]

AI doesn't master anything (at least at the current point in history - 100 or 1000 years in the future might be a different matter). AI designers might "master" various things. You keep talking about the machines doing this and the machines learning that. No, people who built or work with machines do this or learn that.

Except that the machines have abilities their creators clearly lack. The designers of AlphaFold could not themselves fold proteins nor could the designers of Midjourney draw any subject in any style. They programmed the computer to figure out how protein folding or visual art worked on its own (hence machine learning) and once fully trained the computer can fold proteins or draw without any help from humans.

[keenir]

AI doesn't master anything (at least at the current point in history - 100 or 1000 years in the future might be a different matter). AI designers might "master" various things. You keep talking about the machines doing this and the machines learning that. No, people who built or work with machines do this or learn that.

Except that the machines have abilities their creators clearly lack. The designers of AlphaFold could not themselves fold proteins nor could the designers of Midjourney draw any subject in any style. They programmed the computer to figure out how protein folding

you just said the designers could not fold proteins.

well, neither can a string of binary code.

[Starbeam]

... but i fail to see why we need to drain so much water when nuclear energy could suffice.

Do you think the water is to power the data centers? It's to cool them. And it's not like it's gone forever, it goes back to the water cycle. The main issue is that more pure water (low in salts and other minerals) is also the type of water that life prefers to drink, while less pure water is bad for the pipes and can cause them to clog. So when we use water that is preferable for cooling data centers, we aren't using it for drinking or watering crops or whatever. Still an issue, but not the same you seem to think there is from your phrasing. If you meant something else I'd love to hear. I'm not strongly pro or against in this case, I'd just like for people to be pro or against for actual reasons and not imagined bs. But that was all I could image from you comparing AI water usage to nuclear power.

I honestly didn't realize that, i appreciate the information. I still am firmly against AI on other bases, but it at least seems not as apocalyptically bad as I thought.

[alice]

once fully trained the computer can fold proteins or draw without any help from humans.

Except that the computer does not "know" that it's doing either of these things, any more than my SCA program "knows" it's trying to replicate in seconds something which takes human societies decades to do.

[malloc]

Except that the computer does not "know" that it's doing either of these things, any more than my SCA program "knows" it's trying to replicate in seconds something which takes human societies decades to do.

So what. Airplanes have no conception of flight but they fly just the same. Whether computers have any subjective awareness has no bearing on their ability to replicate human activities like drawing or playing chess. Importantly the impact on the job market and academia and so forth remains the same: humans lose positions to computers that can perform their jobs except faster and cheaper.

[zompist]

Did we really need an Eddy but for the pro-AI side?

2. Has anyone compared the environmental degradation caused by running an LLM for a few minutes vs an author obsessing over a piece of writing for months while consuming resources on a first world scale?

Your contempt for authors is getting out of hand. The median income for full-time writers, from book sales alone, is $12,000 a year. You want to shit on poor people so AI techbros can get richer.

I'll ask you the same question I asked malloc: can you name five novels by AIs you found as moving or thrilling as human-written ones?

We're frying our own ecosphere and if we don't stop, billions of humans are at risk. Some of us leftist bros think this is a bad thing, and that we should literally not add fuel to the fire. According to the MIT Technology Review, the power consumption of data centers will triple, largely due to AI, to 12% of total US power usage. A lot of that will be fossil-fuel-intensive.

3. Blanket opposition to science doesn't come from a principled stance. These are people who are outraged that their livelihood is imperiled. Instead of being luddites, they should oppose the root cause of the problem, capitalism. As long as capitalism is in place, there is nothing we can do to prevent reservoirs from being depleted.

First, opposing the capitalist misuse of AI rather than AI itself is what pretty much everyone besides malloc has been saying. You deciding to cheerlead the AI tycoons and throw authors under the bus is not "opposing the root cause".

Second, moving toward dystopia is not inevitable. All wars since WWI haven't been based on poison gas. All wars since WWII haven't been nuclear. We are no longer destroying the ozone layer with fluorocarbons. We are no longer indiscriminately spraying DDT on everything. We are no longer prescribing thalidomide.

And finally, there are sound capitalist reasons why AI may be a boom that's liable to bust. The techbro dream is that big business gets reliant on AIs and pays them and only them a shitload of money. Last year OpenAI spent $9 billion and lost $5 billion. Now, we live in the topsy-turvy world of Late Capitalism where investors expect a company to lose a fortune for years, but the prospects are not great. It's not clear that revenues can increase enough; AI is still on a loss-leader basis so companies don't know its real price; and the monopoly moat business plan was also torpedoed by DeepSeek. OpenAI may turn out to be as much of a boondoggle as Meta's VR.

Some busts are pure folly, but some leave useful infrastructure behind. If the big techbro AI projects all failed, that would leave the academic and small-scale AI projects in place, as well as cheaper alternatives like DeepSeek. Which would be a vast improvement.

[keenir]

Except that the computer does not "know" that it's doing either of these things, any more than my SCA program "knows" it's trying to replicate in seconds something which takes human societies decades to do.

So what. Airplanes have no conception of flight but they fly just the same.

no, airplanes have no conception of flight, and they are flown. by humans. not by airplanes.

Whether computers have any subjective awareness has no bearing on their ability to replicate human activities like drawing or playing chess. Importantly the impact on the job market and academia and so forth remains the same: humans lose positions to computers that can perform their jobs except faster and cheaper.

Okay, this is something that can be looked into & checked...what humans have lost their professorships to AIs or other computers?

[malloc]

no, airplanes have no conception of flight, and they are flown. by humans. not by airplanes.

Sure but airplanes currently lack the intelligence to fly themselves. Computers by contrast are rapidly developing the intelligence to conduct numerous tasks without human direction. My point of course was that computers can achieve the practical functionality of intelligence even without conscious awareness of their activities. Someone here once started a whole thread asking whether consciousness is even functional as I recall. For the most part, I have avoided the question of AI developing consciousness in this thread because it delves into complicated and unresolved philosophical issues while having little bearing on the practical impact of AI on the economy or institutions.

Okay, this is something that can be looked into & checked...what humans have lost their professorships to AIs or other computers?

None yet to my knowledge, although numerous writers and artists have lost their jobs to generative AI with musicians and actors also increasingly on the chopping block. Nonetheless we hardly need contemporary examples of disaster to sound the alarm about future catastrophes in the making. Consider the threat of global warming, which threatens to wipe out billions of people if left unchecked. Should we wait until people are dying en masse and humans are teetering on the brink of extinction before rethinking our fossil fuel policies? Obviously not and neither should we wait for the worst case scenario with AI to take precautions against it.

[keenir]

no, airplanes have no conception of flight, and they are flown. by humans. not by airplanes.

Sure but airplanes currently lack the intelligence to fly themselves. Computers by contrast are rapidly developing the intelligence to conduct numerous tasks without human direction.

we hoi polloi call operating "without human direction" by the name of autopilot.

Okay, this is something that can be looked into & checked...what humans have lost their professorships to AIs or other computers?

None yet to my knowledge, although numerous writers and artists have lost their jobs to generative AI

lost their jobs?

with musicians and actors also increasingly on the chopping block. Nonetheless we hardly need contemporary examples of disaster to sound the alarm about future catastrophes in the making.

we've got more reason to be worried about a second war against emus, than any battles with AI.

Consider the threat of global warming, which threatens to wipe out billions of people if left unchecked. Should we wait until people are dying en masse

Malloc, even I, with my crappy memory recall, can remember at least a hundred people dying in central France (i think it was) because the temperature was so hot it knocked out air conditioning units in mid-summer.

and that was during a Europe-wide heatwave.

and humans are teetering on the brink of extinction before rethinking our fossil fuel policies?

we've been rethinkiing and debating our fossil fuel policies for decades already.

Obviously not and neither should we wait for the worst case scenario with AI to take precautions against it.

if you don't know what the real dangers are, you can't take effective precautions. you don't handle oil fires the same as regular fires.

[malloc]

if you don't know what the real dangers are, you can't take effective precautions. you don't handle oil fires the same as regular fires.

Nobody has explained why the dangers I am describing are not real, though. Machines capable of producing passable imitations of art and literature already exist. Machines capable of beating any human at chess or folding proteins already exist. Computer hardware and AI models are getting more powerful everyday. Unless the current onslaught of digital innovation suddenly hits a wall, computers will overtake humans and possibly very soon.

[Ares Land]

I work in IT; supposedly computer programmers are a prime target for replacement.
There are absolutely no signs of that at all. The job market is, well, about the same as it ever was.

Looking at job posting online, there are very few position that involve AI at all.
I've worked on 20+ projects in the past ten years. Two of them involved AI -- but neither generative AI nor LLMs.
One product owner I work with insists on using a LLM. We mostly humor him but no one really takes the idea seriously. I'm genuinely curious but I don't expect much will come out of the idea.
Programmers and sysadmins swear by ChatGPT these day. Their use of it is functionally equivalent to using Stack Overflow / Server Fault. I can't say it makes people more productive.

I think artists and graphic designers suffer more. Clients prefer something free -- even though it's invariably ugly and cheap looking -- over paying people.

[Raphael]

Machines capable of beating any human at chess

...have existed for a quarter of a century. For most of that quarter of a century, AI had no impact.

[Ketsuban]

[Machines capable of beating any human at chess] have existed for a quarter of a century.

I feel like this is tricky to pin down. There were chess engines prior to Deep Blue which had beaten humans, and Deep Blue was notable because it played under normal tournament conditions, but the hardware it needed was substantial and highly specialised. I'd put the point where chess engines as abstract software clearly surpassed humans somewhere between 2006 and 2009.

[Raphael]

[Machines capable of beating any human at chess] have existed for a quarter of a century.

I feel like this is tricky to pin down. There were chess engines prior to Deep Blue which had beaten humans, and Deep Blue was notable because it played under normal tournament conditions, but the hardware it needed was substantial and highly specialised. I'd put the point where chess engines as abstract software clearly surpassed humans somewhere between 2006 and 2009.

Ah, thank you for the heads-up.

[malloc]

I think artists and graphic designers suffer more. Clients prefer something free -- even though it's invariably ugly and cheap looking -- over paying people.

So you concede that AI is taking at least some jobs from humans. Finally some progress in this thread.

[Raphael]

I think artists and graphic designers suffer more. Clients prefer something free -- even though it's invariably ugly and cheap looking -- over paying people.

So you concede that AI is taking at least some jobs from humans. Finally some progress in this thread.

zompist made that point two years ago:

https://zompist.wordpress.com/2023/07/2 ... bout-llms/

Which you would have noticed if you would understand the position of those who disagree with you here.

(Part of the problem might be your psychological inability to acknowledge any intermediate steps between "X is nothing to worry about" and "X is the end of the world", so that, when people deny the latter, you falsely assume that they must mean the former.)

[malloc]

zompist made that point two years ago:

https://zompist.wordpress.com/2023/07/2 ... bout-llms/

Which you would have noticed if you would understand the position of those who disagree with you here.

(Part of the problem might be your psychological inability to acknowledge any intermediate steps between "X is nothing to worry about" and "X is the end of the world", so that, when people deny the latter, you falsely assume that they must mean the former.)

Sure but he also says that AI has suddenly gone from one percent of human intelligence to half. That number will only increase as research into AI continues. In the race for intellectual dominance, humans are standing still (Homo sapiens has not gotten any brainier over the millennia we've existed) while AI is accelerating at full gallop. For that matter, Zompist also says that "maybe the human brain isn’t as amazing as we think it is".

[keenir]

zompist made that point two years ago:

https://zompist.wordpress.com/2023/07/2 ... bout-llms/

Which you would have noticed if you would understand the position of those who disagree with you here.

(Part of the problem might be your psychological inability to acknowledge any intermediate steps between "X is nothing to worry about" and "X is the end of the world", so that, when people deny the latter, you falsely assume that they must mean the former.)

Sure

is that a 'sure' in the sense of 'you're right about all of that' or a 'sure' in the sense of 'i don't know what you said & i don't care what you said, i just want to keep panicking' sense?

but he also says that AI has suddenly gone from one percent of human intelligence to half.

the link is right there; but you've made us all instinctively want to know what your link is.

That number will only increase as research into AI continues. In the race for intellectual dominance,

...which only you seem to think is actually being run.

[humans are standing still (Homo sapiens has not gotten any brainier over the millennia we've existed)

theres no need to do so.

while AI is accelerating at full gallop.

I'm an equestrian. what exactly do you think a gallop is?

For that matter, Zompist also says that "maybe the human brain isn’t as amazing as we think it is".

given that we used to think the human brain was one short step away from the brain of G-D, thats quite the development.