> If AI obviates the need for human labor, then obviously those who control AIs will become the elite while the rest are left to rot.
Absent political intervention, I think we agree here.
> Therefore, if we ensure everyone controls AIs, the power differences will not become so staggering as to be irreversible.
This part isn't clear to me though, but I'm open to being convinced (and frankly, would like to be convinced?). Right now most people (indirectly, via money) trade their labor for access to essentials like food/housing. If we can't do that, and everyone has access to roughly equivalent AI capabilities, how do I monetize my own access to SOTA AI? It only seems possible if you already have a lot of physical capital that the AI can manage as a business.
I guess if the endgame is instead very good non-AGI AI that doesn't entirely obviate human labor, your scenario makes a lot more sense to me. But not in the case of total replacement. In that scenario it seems like ownership over physical capital (land, data centers, energy, factories, robots, etc.) would become the only remaining source of power.
On a side note, somewhat optimistically, I think "absent political intervention" is carrying a lot of weight. Unemployment during the Great Depression peaked at <25% (iirc) and incited a lot of political change that advantaged much of the working class. AGI would be capable of inducing much higher unemployment and it would start (is starting?) with the relatively more political powerful white-collar segment of the working class.
> for ushering in the technofeudalism that will put us all in the permanent underclass.
Why is unlimited access to SOTA AI less likely to put us here? If AI obviates the need for human labor, how does having GPT-5 Sol help me get food or shelter any more than GPT-3.5 would?
> Arxiv is full of pre-prints that anyone can upload.
You now (at least for some categories) have to receive endorsement from someone who has multiple recent papers on arxiv in the same (or adjacent) category.
Agreed. I know nothing about nuclear physics. I still doubt you could pick a random person off the street and have them convincingly pose as a nuclear physicist to explain a "nuclear physics" concept to me. I doubt you could do it with a random PhD from a non-physics field. An LLM could probably convince me even if 90% of the content of the explanation is subtly or blatantly incorrect.
I get 0% (accurately) on my latest paper. Not super surprised, as I intentionally avoid some LLM-isms that I used to use because I don't want reviewers to have even the slightest indication that text is LLM-generated (even if in principle I'm not opposed to polishing or even wholesale generating academic text if it can convey the original research well, especially for non-native speakers).
I don't think the problem is as bad as a naive reading of this article suggests. I'm highly skeptical that anywhere near 65% of recent CS papers that I've read (mostly systems papers) are substantially AI-written. I threw some recent papers I've read into the system and they come back as 0-7%.
> As a software developer, I am used to finding and fixing the underlying problem instead of relying on the quick fixes these doctors were offering me.
I'm skeptical that avoidance of "relying on the quick fixes" generalizes to software developers as a whole :)
> I mean, you can read them even without the colors
I'm not colorblind and I was depending on the textual context implying Sol was better than Terra. I had to zoom in quite far to actually differentiate between the colors.
If they insist on terrible colors would it be so hard to differentiate by marker shape or line dashing too?
I was wondering the same thing. From textual context it is clear enough that Sol should be above Terra, but I had to zoom in really far to actually differentiate between the colors and I'm not colorblind. I saw a light mode version of the plot on twitter that was better but still not great.
OpenAI's plot design has been consistently awful and inaccessible, it seems like they're optimizing for something other than readability because I find it hard to believe they aren't putting in any effort for such major announcements. If the colors have to be awful they should at least differentiate with marker shapes or line dashes.
At least it isn't as bad as the stacked bar chart where the 50-something bar was higher than the 60-something bar.
I agree with the general sentiment of this comment, but national labs do hire foreigners/non-citizens, albeit possibly not from all countries with eligibility for all roles.
The funniest one I've noticed lately is a bunch of Capital One ads saying "We built a multi-agentic system for finding a car to buy!"
I'm not saying I 100% wouldn't use AI to help me in product searches, but isn't one of the main selling points of AI that it is general-purpose? Why can't I just boot up ChatGPT and ask it what cars have XYZ things I need? Certainly being informed that Capital One's system is "multi-agentic" doesn't tell me much about what is being offered.
> There are some people that believe that writing is an act of creative expression.
I think "some people" might be underselling it, as evidenced by the borderline innumerable fiction books in existence?
> and as such, it's a quite selfish activity
"quite" seems a bit harsh, surely "writing because you enjoy it" is pretty far down the list of all "selfish" activities? I'd imagine many authors also write because they think others will enjoy their works.
> I have yet to see a "error" that modern frontier models make that I could not imagine a human making
I mostly agree if "a human" is just any person we pluck of the street. What I still see with some regularity is the models (right now, primarily Opus 4.6 through Claude Code) making mistakes that humans:
- working in the same field/area as me (nothing particularly exotic, subfield of CS, not theory)
- with even a fraction of the declarative knowledge about the field as the LLM
- with even a fraction of frontier LLM abilities suggested by their perf in mathematical/informatics Olympiads
would never make. Basically, errors I'd never expect to see from a human coworker (or myself). I don't yet consider myself an expert in my subfield, and I'll almost certainly never be a top expert in it. Often the errors seem to present to me as just "really atrocious intuition." If the LLM ran with some of them they would cause huge problems.
In many regards the models are clearly superhuman already.
> because other individuals, organizations and nation states are not going to stop, and not going to leverage their AI if they get ahead of us.
I don't think that it is likely AT ALL, but it is probably only necessary for China and the US to agree to stop, not all organizations and nation states. It is at least possible given leadership in both countries that see AI as an existential threat.
The hardware needed to run and train SOTA AI can only be made by a very small handful of companies in a small handful of countries that either the US or China have significant influence over. Making AI R&D illegal would stop 99% of it overnight, most of the researchers are in it for money rather than some ideological commitment and there are plenty of other well-paid jobs they could take. Doing local inference in secret with existing models and GPUs would be possible, but training new SOTA models probably wouldn't be.
> Job loss is likely to have statistics more comparable to the Black Plague.
Maybe this is overly optimistic, but if AI starts to have negative impacts on average people comparable to the plague, it seems like there's a lot more that people can do. In medieval Europe, nobody knew what was causing the plague and nobody knew how to stop it.
On the other hand, if AI quickly replaces half of all jobs, it will be very obvious what and who caused the job loss and associated decrease in living standards. Everybody will have someone they care about affected. AI job loss would quickly eclipse all other political concerns. And at the end of the day, AI can be unplugged (barring robot armies or Elon's space-based data centers I suppose).
> LLM's are better at keeping consistency at details (but not at big picture stuff, interestingly.)
I think it makes sense? Unlike small details which are certain to be explicitly part of the training data, "big picture stuff" feels like it would mostly be captured only indirectly.
> As it turns out Nvidia's H100, a card that costs over $30,000 performs worse than integrated GPUs in such benchmarks as 3DMark and Red Dead Redemption 2
> you'll also end up with scores of people who "correctly" followed the signals right up until the signals went away.
I think this is where we're headed, very quickly, and I'm worried about it from a social stability perspective (as well as personal financial security of course). There's probably not a single white-collar job that I'd feel comfortable spending 4+ years training for right now (even assuming I don't have to pay or take out debt for the training). Many people are having skills they spent years building made worthless overnight, without an obvious or realistic pivot available.
Lots and lots of people who did or will do "all the right things," with no benefit earned from it. Even if hypothetically there is something new you can reskill into every five years, how is that sustainable? If you're young and without children, maybe it is possible. Certainly doesn't sound fun, and I say this as someone who joined tech in part because of how fast-paced it was.
LLMs and AI more broadly certainly seem to have upended (or have the potential to upend) a lot of white-collar work outside of technology and art. Translators are one obvious example. Lawyers might be on the chopping block if they don't ban the use of AI for practicing law. Both seem about as far as you can get from "careers in technology," and in fact writing has pretty much always been framed as being on the opposite end of the spectrum from tech jobs, but is clearly vulnerable to technological progress.
Right now I can think of very few white-collar jobs that I would feel comfortable training 4+ years for (let alone spending money or taking on debt to do so). It is far from a guarantee that almost any 4-year degree you enroll in today will have any value in four years. That has basically never before been true, even in tech. Blue collar jobs are clearly safer, but I wouldn't say safe. Robotics is moving fast too.
I really can't imagine the social effects of this reality being positive, absent massive and unprecedented redistribution of the wealth that the productivity of AI enables.
Absent political intervention, I think we agree here.
> Therefore, if we ensure everyone controls AIs, the power differences will not become so staggering as to be irreversible.
This part isn't clear to me though, but I'm open to being convinced (and frankly, would like to be convinced?). Right now most people (indirectly, via money) trade their labor for access to essentials like food/housing. If we can't do that, and everyone has access to roughly equivalent AI capabilities, how do I monetize my own access to SOTA AI? It only seems possible if you already have a lot of physical capital that the AI can manage as a business.
I guess if the endgame is instead very good non-AGI AI that doesn't entirely obviate human labor, your scenario makes a lot more sense to me. But not in the case of total replacement. In that scenario it seems like ownership over physical capital (land, data centers, energy, factories, robots, etc.) would become the only remaining source of power.
On a side note, somewhat optimistically, I think "absent political intervention" is carrying a lot of weight. Unemployment during the Great Depression peaked at <25% (iirc) and incited a lot of political change that advantaged much of the working class. AGI would be capable of inducing much higher unemployment and it would start (is starting?) with the relatively more political powerful white-collar segment of the working class.