81 points•gotmedium•8 days ago•85 comments•

85 comments

1293486818 days ago
The actual report is by Epoch AI, which is financed by totally independent funds like (Eric) Schmidt Sciences.

Professor Tabarrok uncritically praises the report, which might lead to the conclusion that either human intelligence or moral standards are "falling rapidly".

oklahomasports7 days ago
What possible ideological message could that blog post be pushing?
simianwords7 days ago
Its run by a secret cabal that wants to push AI to the masses and cash from IPO before the bubble pops
mchusma7 days ago
What? The report is about ai performance and cost.
caycep7 days ago
I mean, this is more or less true of Marginal Revolution for over a decade now...
silverFork8 days ago
Price of human intelligence is also falling rapidly. I suspect looking pretty is going to go up in value more rapidly.
jordanb8 days ago
In my career I've worked with people with PHDs in Physics. People with advanced degrees in fluid dynamics and turbine engineering, people with chemistry and ChemE degrees, postdocs in Geology.

All were working alongside me, a self-taught programmer making websites.

I've long thought that there's an oversupply of smart people in the world but for the field of software development, which was hiding that fact by soaking all of them up.

If SDE is collapsing as a field, a whole bunch of people who thought their intelligence and credentials guaranteed them a comfortable life are in for a shock.

mrdependable7 days ago
This is really looking at the problem from the wrong angle. It’s not that we don’t need all those degrees. The problem is that tech companies are making so much money that it makes other industries practically irrelevant when it comes to return on investment. It has thrown the economy completely off balance.
margalabargala7 days ago
> If SDE is collapsing as a field, a whole bunch of people who thought their intelligence and credentials guaranteed them a comfortable life are in for a shock.

Even if we accept your precondition as true, that SWEs contain multitudes of smart people who will jump to other industries, I don't think the conclusion holds.

In the US at least, SWEs are among the most entrepreneurial industry members. They'll bring that attitude to whatever their new industry is. I think there will be an explosion of new startups in different areas, and we'll Jevons Paradox our way forwards.

rendang7 days ago
South Korea would be an interesting point of comparison. High IQs, after-school tutoring, 70% of young people get degrees: result is a college wage premium of only 31% vs 54% average in the OECD

https://www.oecd.org/en/publications/education-at-a-glance-2...

naveen997 days ago
I would rather have SpongeBob making my burger patties than squidword. More intelligence is better than less intelligence.
sph8 days ago
Price of human intelligence is going up actually, because there's ever-decreasing amount of it, and we've invented a cheap substitute that makes people even stupider.
bryanrasmussen8 days ago
Can you prove to me that the price of human intelligence is going up? At best it is as it always has been, intelligence in some fields is going up.

It's generally true that the rarer something is the more valuable that is, but only if there are enough potential buyers of the thing that can recognize it. If there is an ever-decreasing amount of intelligence there is probably an ever-increasing amount of stupidity - so will the stupid be smart enough to reward the intelligent with high paid positions?

What I suppose is actually that human intelligence will become more monetizable, by allowing those with it to better take advantage of their unrecognizable ability. But only if the interlocking systems of humanity do not fall apart and destroy the world due to there not being enough intelligent people left to make things work, as per various works of science fiction on the subject.

silverFork7 days ago
In schools there are massive changes. I was told that kids are more relaxed and care less about education and that the value of the smart kids, the know it all kids, is going down as more students can access Ai and they are not as helpless as they were when failing to tune in while in the class. The dynamics are changing and students are less suicidal too apparently because the pressure to deliver scholastically is going down because no one knows what the future holds and what will be valued.
mrshadowgoose7 days ago
I frequently have speculative conversations about what (good-leaning) post-AGI societies might look like.

When asked about how one might attain social mobility in such a society, where the intrinsic value of our intelligence and ability to do labor is gone, I unironically respond with "be hot".

pocksuppet7 days ago
some people think that's how humans got so intelligent. actually if you think about it, it's the only way humans could have got so intelligent. brains were our tail feathers for a while.
onlyrealcuzzo7 days ago
> Price of human intelligence is also falling rapidly.

Where?

What's your definition of "rapidly"?

I could see an argument for price of human intelligence going down - we're essentially in a global stagflation - which feels like a recession (sometimes moreso) even if it isn't technically one...

Rapidly? I think we have different definitions, or you're only looking in a niche.

I'd also argue that global stagflation doesn't yet have anything to do with AI. My crystal ball doesn't work. Things could change quickly. But I don't see any evidence now.

silverFork7 days ago
High priced intelligence is currently the doctors, engineers, professors, scientists, analysts, bankers, technical staff.. etc. Many are already replaceable with AI and many others sooner or later you know. In my opinion that is.
IAmGraydon7 days ago
If human intelligence stops being the trait that propels one up the socioeconomic ladder, it will not be good looks that takes its place. The ability to project physical force will become currency. This is why AI could cause civilized societies to completely unwind.
silverFork7 days ago
We will have the robots that will beat them up; LOL.. My robot will know the Brazilian Jiu-Jitsu and will beat up their robot!
_superposition_7 days ago
I love marginal revolution. But I'm not fond of the unit stated. I dont believe we're buying intelligence. I'm sure marketing will run with it though.
zaphar7 days ago
I have been searching for a good term or name for what we are actually buying. A "probabilistic search for a reasoning chain through a semantic space" is about as close as I've gotten and I have a feeling it's not exactly accurate either.
judahmeek7 days ago
"Plausibility"
cs7027 days ago
Yes, the price of inference is falling rapidly.

What about the cost? Inference looks like a viable business model for those operators that have access to SOTA model and serving infrastructure, but the investment required to have it is enormous, and appears to be never-ending, because if an operator stops investing aggressively, its model and serving infrastructure quickly become non-competitive, and customers will quickly leave for alternatives.

epolanski7 days ago
I went from 408 euros/month of AI spending to 28 in the span of one year.

The 20 euros GPT plan with GPT 6 Luna (and Astra for doing the tough stuff) and OpenCode Go (8 euros) offer me both way more intelligence and tokens than spending 400 euros just at the beginning of the this summer did.

Today I worked two hours with GPT 6 Luna, have been very productive tackling some hard stuff and...my weekly usage went from 57% to 56%?

You really don't need to use the strongest model for everything. And even chinese open models are effectively few months old sota. You can use mimo 2.6 or DS 4.1 flash on 8 $/months OpenCode and barely ever have to worry about consumption.

cs7027 days ago
Thank you. Yes, the price of service is falling rapidly. I agree.

What I don't know, and was asking about, is whether the cost of providing that service is also falling rapidly, including the cost of training.

prpl7 days ago
I’d happily take Opus 4.8 xhigh at a reliable 10k tokens per second over oversubscribed SOTA models.

At some point this will probably happen, but we’re not quite there yet.

hyperionultra8 days ago
Well, since llms need matterial for self-education, someone still had to be inteligent enough. Or llms will freeze their “intelect” at some point.
bonoboTP8 days ago
No, they can just solve tasks verifiably. Reinforcement learning from verifiable rewards. Pure next-token supervision is only in the pre-training phase for modern LLM agents. They can also define new RL environments and pose new challenges to themselves and train themselves to solve them faster. Yes, at some point some human steering and judgment comes in as to what sorts of tasks to train towards. But there is no obvious taper-off at human level.
hiddencost8 days ago
Surprisingly, this isn't necessarily true.

To take an intuition: there are lots of problems that are easier to pose and to check than they are to solve, in mathematics.

esseph8 days ago
> Well, since llms need matterial for self-education, someone still had to be inteligent enough.

Not necessarily. https://arxiv.org/html/2609.26457v1

hlynurd8 days ago
Well, someone. But not necessarily anyone else.
bilbo0s7 days ago
This is the key.

In the age of AI, human super-intelligence, is extremely valuable. But yeah, just being intelligent? Not so much.

Just to illustrate, many have complained that big AI labs "buy" startups only because they want to acqui-hire the Stanford/CalTech/MIT PhDs who started it. Then they pretty much $#!t-can the rest of the people who were, to be fair, mostly just run of the mill UTexas/Umichigan/UofwhateverState TF/PyTorch monkeys. But the point is that in the old days, the second tier guys would have made out like bandits as well.

I think that was, kind of, the start of human intelligence becoming a winner take all game. The AI's only need to learn the latest and greatest from one person before they can leverage that value everywhere. And business is not likely to care too much about who that person is. They will care even less about the people who are not that person.

BenzeneDream8 days ago
Why would it freeze if they can create more training material. Its already been shown synthetic training data is just as good if not better.
timcobb8 days ago
> Its already been shown synthetic training data is just as good if not better.

Source, please! I'm a machine learning layperson but how can this not be overfitting?

Read the full thread on Hacker News →

Related stories