Casual Conversation

No because LLMs are actually quite dumb and nowhere near close to AGI yet.

Some jobs will be lost, especially I can see concept artists getting reduced everywhere they are used. But nothing catastrophic yet.

Anyways I’m not afraid, socialism won’t be viable until we have the capacity to automate all labor.

I think the more you sit down and learn about current AI models (LLM models) the more you will realize a lot and I do mean a lot of the promises and pie in the sky rhetoric is just tech bros out to raise capital and take money from dumb rich people.

The more you learn about current AI, the less you will fear it because it is not some magic bullet. Every single AI out there is riddled with issues just bubbling under the surface.

They don't "think" like people do, they're basically very fancy autocorrect - yes that's a gigantic oversimplification but it isn't too far off the truth.

Even white collar work it isn't great at straight replacing people, it's a useful tool but it isn't a person and AI makes a lot of mistakes still.

Anyways I think it's a good tool, I use it, but I can tell you I'm not scared of AI taking everything away all at once. I do think present day automation tech paired with AI will sooner or later eliminate some jobs but I'm hopeful we see a systemic shift to handle that. I am concerned one day we're going to get efficient enough at enough stuff that we'll have like a permanent unemployed underclass that's 20 to 30% of the population but things like that are solvable crises if we choose to solve them. That's where universal basic income, healthcare as a right and easily accessible, upping the min wage and reducing the normal working hours week from 40 to say 30 will help. Or would help if we lived in a functioning country not teetering on the brink of all our facism (USA).

I don’t think we will see AGI until at least a decade later. Right now LLM is no longer making significant progress (diminishing returns on scaling), AI research will need to find new ways to achieve AGI. This will probably require breakthrough in quantum computing and energy infrastructure.

I’m not worried about AGI because I don’t think we’re anywhere near that level of complexity or programming skill yet. If we ever do get it right, there will be numerous false starts before then.

LLMs are stupid and lack basic common sense. They can’t even take fast food orders reliably.

The hype will continue, but people are seeing through it more and more. What’s currently marketed as “AI” can’t live up to its promises.

1. AI is an extremely overloaded term. We use it to mean any kind of machine learning (most of which has been around a while), generative AI, and even things like NPC logic in video games
2. All the fear mongering you've heard about AGI/super intelligence is based on LLMs. LLMs just process language based on statistical probabilities from their training data and aren't anywhere near close to "understanding" in the sense that you might think of it.
3. LLMs don't seem like the path to AGI, so I sincerely doubt we're anywhere close.
4. LLM based tech isn't even that good at replacing workers for the things it kind of sort of works for. Companies keep trying to replace call center staff with LLM based systems and walking it back after it sucks. Fast food restaurants are trying but people hate it for much the same reasons, so we'll see how it goes.

Tldr: AGI being on the horizon and AI destroying the job market is all marketing to make people think AI is more capable than it is. It is good for some things, but not for everything under the sun like they'd have you believe. The narrative that AI is capable enough to replace humans is especially insidious because that is many CEOs fondest wish, so many of them will drink the coolaid and try and fail.

Tldr tldr: don't panic it's not as big a deal as they'd have you believe.

There is absolutely no sign that AGI is anything approaching reality. LLM technology is a dead end for achieving it. 80% of jobs absolutely will not be gone in ten years. I would be shocked if it were as high as 10%. Productivity will increase in certain fields and that will have a trickle down effect where jobs are slightly reduced but never eliminated. I wouldn't expect AI to be any more disruptive than when PCs or Internet became ubiquitous.

I think an actual super intelligence would fix a lot of shit the stupid version of AI we have now has and continues to fuck up. Among other things caused by massive stupidity.

Too bad we are nowhere near such a thing.

I know this sounds crazy but I really think by 2030 we will have 80% jobs replaced by AI and the new jobs that might evolve out of jobs that have vanished will only be doable for AGI/ Super Intelligence.

I don't think we'll lose 80% of jobs by 2030, but 50% of jobs by 2035 is realistic.

I really think I will be fck.d up in 5 years. A job is annoying, but it’s also a humans purpose. A way to express itself and be useful to society. If this is being taken away we will find 95% of humanity in a huge depression and suicide rates are going to rise by 2030.

But you're spot on with this. Humans earn wages because companies need humans to do things. If companies can do without humans, they won't see the need to pay wages. An economic crisis ensures, and also people lose hope and purpose in life, and that creates a large-scale depression that affects all of humanity.

One of the ways out is to give people resources sothat they at least have an income [Universal Basic Income]. It does not solve the "purpose"-crisis though. The best thing would be if people have fewer children, so there's fewer workers and they have an easier time finding jobs.

All it seems to do is create dumb shit.

"AI" are still tools.

The issue is their underlying technology, as of now, is way more fundamentally limited than 'Tech Bro' types will tell you. Don't get me wrong, they're neat tools, but they are fundamentally incapable of taking over intricate decision making processes. They're just a layer of human assistance and automation.

I'm as big of a local LLM enthusiast as you'll find, and I'm telling you: the AGI scaling acolytes are full of shit, and the research community knows it.

Imagine finding out that you won’t be able to pay off your debt cause most fastfood restaurants will use AI/ Bots that can serve, prepare, clean etc. 24/7 while a useless human needs breaks, wants money and needs days off and can only work 8 hour shifts.

This sucks.

...But honestly, in the long run, it's not so bad. Working fast food sucks and it would be great if people could do something else instead.

As a little silver lining, there's a good chance 'AI,' as it is now, is goin to 'race to the bottom,' and a lot heavy lifting will be done on your phone or some computer you own. So you'll have a little assistant to help you with stuff, self hosted, not corporate cloud controlled. Think Lemmy vs Reddit in that regard.

Marvel's X-Men Comic got it exactly right.

(Ignoring, for this conversation, that current generations of AI are rubbish.)

Nimrod, trained on all of the knowledge of humanity, was able to plan millions of ways to defeat any of the X-Men who he calculated had any chance to defeat him.

The key insight is that infinite computer cycles can do lots of impressive things, but will always have critical flaws and weaknesses.

Computers are great at striving for perfect, and shit at achieving it.

Imagine a Boeing going at the speed of flight.

That would terrify... Airbus?