upvote
> People make this argument all the time without being able to give specific examples of what jobs will remain. I'm not talking about what jobs will remain in 3-5 years when we are all still interfacing with AIs in similar ways we are today, I'm talking about for our kids in 20-30 years when AI can run billion dollar businesses totally autonomously with robotics.

Businesses must be competitive. If they all use the same AI model, then they’ve lost all competitive advantage by using it. The only advantage left is the inputs that go into the AI model. Who supplies those?

Since you’re asking for examples, let’s talk about marketing. You cannot fully automate this. It’s a zero sum game. It’s like if all business owners decided to pick a really good marketer named Steve. Steve could probably come up with some bangers for a few clients, but once every business is asking Steve for marketing, it all looks the same.

High end fashion brands like Gucci, Prada, Louis Vuitton, are perfect examples of this. Why would I buy such an expensive bag from them? AI is never going to be the reason because fashion is not rational, even if it’s the one constructing the bag

reply
> If they all use the same AI model, then they’ve lost all competitive advantage by using it.

In the next 3-5 years, sure, but that's not the long term vision I see. I think it's more likely we'll see the Anthropics/OAI/Google/Meta/XAIs of the world stop selling tokens. Instead they are the ones offering wider ranges of SaaS products, starting quant trading operations, taking over media (even more than today), developing new pharmaceuticals, etc. GPT 12 or Gemini Francium can and will tighten the RL loop by improving robotics, and have billions of dollars to offer armies of humans to fill the gaps where needed, until it no longer needs humans.

> let’s talk about marketing. You cannot fully automate this

Marketing is one of the most easily automatable jobs there is. AI is already capable of designing commercials, making cold calls, buying ads. Sure, they suck right now and are weird, but in 20-30 years when all of society has been so permeated with AI, AI will be more capable and also people will be way more accepting of AI.

reply
Fully agree: it's not a law of the universe that automated jobs are guaranteed to have a replacement.

I even find contradicting claims within software companies: that agents will boost productivity 10x, so they don't need to hire so many people. But if it made people 10x as productive, the maths say they could hire 10x the people to become 100x as productive! Shouldn't they do this and do circles around their competition?

(a counter argument is the productivity claims mean it's multiplied in general but not exactly per head... But I don't see many nuanced arguments, just "accept or leave").

reply
> I'm talking about for our kids in 20-30 years when AI can run billion dollar businesses totally autonomously with robotics.

This isn't going to happen. You're in the middle of a hype cycle.

reply
Then why are young graduates not getting hired anymore?
reply
It's tough to respond to an unsourced claim with anything meaningful. Here are some possible answers to your question that don't jump all the way to the conclusion that AI will end the concept of employment:

1) Your claim is false. At least at my company, we are still doing the "intern to full-time hire after graduation if we like you" pipeline that we've always been doing. Does my one anecdote defeat your zero anecdotes? I don't know.

2) The companies are caught up in the hype. They're trying to find out what it's good at and what it's not. As reality comes back online and prices go up, they'll find reasons to hire humans again.

3) Money is no longer free (ZIRP is over) so there is less experimenting and less hiring and less mobility between employers, meaning fewer openings completely independently of the AI stuff.

There's many explanations that don't require falling for the AI hype cycle.

reply
The Riemann Hypothesis was just proven yesterday by AI and people still say this is a “hype cycle”
reply
Yeah. We're still figuring out what it's good at and what it's not good at. It seems to be decent at things with concrete correctness claims, like math & basic programming, but it still needs lots of assistance even in those fields, and it's actually quite bad at maintaining large projects. It is absolutely miserable at fields with less concrete correctness, like writing, or creating attractive art & media. It's also very expensive. Once they stop giving it away for free and start actually charging money for it, I suspect reality will settle in and we'll see it used where it's actually appropriate, and abandoned where it's actually not very good and/or too expensive. It's a cool tech, we can do things with it that we couldn't do before, but it's not magic and it's not going to end the economy.
reply
In 20-30 years those NFTs will finally be worth millions too and we'll all be prosperous and trading monkey jpegs, right?

Predictions about the medium term future on HN are ridiculous. 95% of them are mindless drek, 4% are weak claims, and 1% have anything close to reality baked in and might be reasonable.

reply
It was obvious to everyone during the NFT craze that they were worth zero; this is much closer to the creation of the Internet or railroads. I hope I’m wrong about the long term trajectory of AI, but so far, there hasn’t been any sign of it slowing down or concrete areas where we are certain it will never out perform humans at a single task.
reply