Generative AI is still terrible at producing content people would actually pay to access. It’s hit the commissions market because the commissioner has some idea they want and someone who knows how to do the promos can make a lot of images quickly. Yet, Disney animation isn’t laying off anyone these images don’t suck but they’re also not good enough yet. Give it 40 years and who knows but that’s not tomorrow.
In terms of programming yes in very specific situations these things can improve productivity, but so can new libraries etc. It’s a long way from actually replacing programmers because what they are good at is what there’s examples that are close to it in the training data. Ask for a basic weather simulation using different atmospheric gasses and it’s not going to know what to do.
It is reshaping society in much the same way the industrial revolution did. Luddites resisted change, for precisely the same reason people aren't happy about AI, and it's now a derogatory label.
It's much better to adapt to a changing world but precisely what that entails is still up in the air.
Why does tech believe there should be no constraints whatsoever on what we build? Either our "change the world" pitches were completely inflated--what we do is meaningless dross--and we don't need to be regulated, or our mission statements were bang on and like any world-shaping thing we desperately need need to be regulated. You can't have it both ways.
I'm not against AI regulation on safety grounds but I'm not convinced it's a good idea to try to use regulation to preserve jobs. That feels like it would have much more damaging longer term consequences.
Not at all what I'm saying. What should they have done? Not trained on code they weren't explicitly licensed to do so, and only given access to workers.
The message of companies offering "productivity gains" (or outright replacement: "Hire Devin") to workers is "we're only interested in putting you out of work; how you pay your mortgage and get health care is your concern; call your congressperson". The fact that they're pushing hard to replace workers and flat out stating that their business models are illegal under current copyright law says everything about their offering: hey big business, if we put a carve out in copyright law your SWE headcount goes to zero; call your congressperson.
Hence my original--earnest--question: do you all who are working on AI honestly think the government will fix this?
People will adapt and find new things to do just like they did in the first, second and third industrial revolutions already.
EDIT:
> first, second and third industrial revolutions
These killed and maimed millions of people
Humans still need to do agricultural labour, the only difference is that machines mean less humans need to do that as a percentage of the population.
AI doesn't improve human labour productivity, but makes it obsolete in a similar way to how the combustion engine made the horse obsolete.
What exactly are you going to do of economic value when an AI can do everything the human mind can do?
This right here, this is 90% why.
With 10% of "what this button do?"
Thankfully I believe this is largely just overblown hype. Silicon Valley has gotten way too high on it's own supply of bullshit and the AI craze is that being taken to the next level of insanity. Regardless, the tech industry is amoral at best. We are so obsessed with "can we" (or rather "can we convince a VC that we could") that we never take the time to stop to consider "should we".
Ultimately, watching tech find new and innovative ways to fool and to fleece people has, at times, almost made me feel shame that I've been part of this industry for almost 25 years now.
We are addicted to the idea of disruption and the monetary rewards that come with it. Generally we don't care if we sink the rest of the world in the process either.
Damn shame, that.
Agree that we don’t think enough about it.
In such an economy, it may be that we can become communists! One of Marx's ideas was that technology will liberate man from labor. In actual historical Marxist societies, slave labor and coercion were widespread. But, in the future, if all labor is done by robots, far-left, restributionist policies may be viable because there's no humans being exploited (except, ironically, the humans who invented AI -- the people you're criticizing). In such a communist utopia no one will have to work, people will just have hobbies and hang out.
Are we heading towards this sci-fi scenario? I doubt it. I'm just trying to make two points: 1. when we imagine a future world with AI, there's no reason to imagine that it's a bad one and 2. it could be the case that "from each according to his ability, to each according to his needs" becomes a viable credo in an imaginary post-AI world.
This is a total reshaping of society at worst, and a huge transfer of wealth with almost no steps at best. How does anyone working on this justify causing unemployment on a mass scale while collecting a huge check? Is it just "progress gonna progress?" Do you really believe that justifies all the consequences? Is it, "if not me/us then someone else (worse)"? Then why aren't you giving control of these tools to the workers? Why are you profiting so handsomely?
This is nuts. Normally I don't go full communist rant, but a few hundred people (presumably some of whom read HN) are about to destroy our society without asking a single one of us, and with absolutely no regard for the consequences.