zug_zug

Yeah given that I don't see any statistics in this article, I'm not convinced of a systemic problem.

So what?

Let's take aside the charged component of this and just suppose hypothetically that 100% of the assaulting students belonged to some subgroup. So what?

They may be 100% black, or 100% male, or something, but it's not like you can reasonably just say "So no males at our school." Regardless of whether you can find a subgroup who this belongs to with a very low specificity, you still need policies that tackle the behavior itself rather than trying to find some really weak predictor.

"Labour shortage" is just the business's perspective on "plenty of jobs!"

If there truly is a shortage of outstanding jobs, then they can compete and raise salaries and the least significant roles ultimately won't get filled, which is optimal.

That's their problem to solve however they want, if they want. The population shrinking by some large percentage doesn't seem to be inherently bad to me, and probably is good.

It's so laughable to me when people say "BUT THE SHAREHOLDER VALUE!" when we talk about things like the important life trends of millions of people.

I agree it's worth looking into, potentially very emphatically. But that means doing some hard science.

That means don't just look at test scores by county, it means looking at DELTAs in test scores across time by numerous dimensions (income, poverty, did the kids get breakfast, school, etc) and an open mind to every possibility. Perhaps comparing against the parents' test scores might help elucidate.

Given the rise of so many unexplained mental health phenomena, I personally suspect there are environmental factors, but again, we need hard science and a lot of it.

The challenge here is that you're assuming performance by race should be constant over time. Maybe that's true, maybe it's not, I think a richer analysis with wealth, parental education, english as a first language cofounders might be helpful.

Also if you want to entertain the theory that it is racial/genetic, it'd be important to do that analysis on a continuum since a lot of people who are "minority X" may be more than 50% white or there may be trends in that over time.

> If you haven't used it, how do you know if it's winning?

I'm judging on benchmarks, and whether anybody or any company I know has ever suggested using it (not yet).

Well I "tried it out" I asked it one question, and it gave no answer and said "Sign up to use more!" I don't think I'll be doing that, no.

I can't think of a single dimension grok is winning on (capability, cost, voice), but want to stay open-minded -- anybody want to vouch for its capabilities in any domain?

I think this is a bit of a simplistic mental approach. I've certainly seen a lot of "The engineer owns the outcome, AI is just a tool, don't release anything you don't vouch for."

However, I just don't think that's realistic. It's asking an author to suddenly become an editor. It's asking somebody who writes code to now read and debug others code.

It can actually be harder to find the the bug in a tricky piece of code than it can be to write your own correct code from scratch. I see AI introduce all sorts of bugs all the time in my personal projects that I would never introduce, and would never think to test for, especially around anything graphical.

I feel like this is a willfully confusing article. So autism may involve disrupted genetics, but you may not have inherited those genes from your parents, they may be what are called "de novo mutations" which are mutations to the sperm or egg cells in the parent. This is evidenced by autism being more prevalent when the parents are older and those mutations are more common.

Also it clearly must be to some degree mutations if you believe the rates are going up, which even the government is no longer disputing (though that may not be conclusive proof, to me it seems evidence leans to more likely than not currently).

De novo mutations can also be caused by environmental factors, this is very well established and called "genotoxicity"[1] and certainly some people have been exposed to known genotoxic chemicals to varying degrees over time (formaldehyde) and who knows what undiscovered ones there are, newer standards of pesticide testing are starting to look for reproductive health effects but many chemicals that have been in use for a long time haven't been held to the highest standard.

Another reason you might think it would be "de novo" mutations is that this could be any sorts of random errors in the genome. If autism is an umbrella label for various genome errors (which it very well may be, since it correlates with other conditions very often), that may explain why it's so hard to pin down a few genes and why symptom expression varies so drastically.

1. https://en.wikipedia.org/wiki/Genotoxicity

Did you read your link? It says that 33% of people adopt AI, not that AI makes all workers 33% more effective, that would be absolutely insane. For that to be true we'd expect to see either companies who use it have revenues all suddenly jumping 33% (which we have not seen) or laying off 33% of their staff (which we have not seen and would also be catastrophic in the short term at least).

I'm not really going to belabor this point more, because it really sounds to me like you haven't even done the most cursory exploration into this and the people who have give an estimate of 7-10%[1] (and only a small fraction of that value would be captured by the AI provider). All the best.

[1] https://www.goldmansachs.com/insights/articles/generative-ai...

Okay but like... so what? Like does that even really pass the common-sense test that that is enough?

Like my bank would give me a phone call before I wired $20k oversees. Google should by default send you a sms confirmation on your first $10k+ purchase. This is the type of common-sense stuff that isn't an issue in-person, but somehow online we get the most barebones protections for consumers because "technically the EULA said so and so."

Basically if we make the corporations liable for it, we give them an incentive to prevent fraud, identity mistakes, etc. Whereas if they are not, we give them every incentive to avoid unintended purchases.

You said 10s of trillions of dollars of revenue, which when the sum of all salaries is 65T, is minimum 15.3% of all salaries best case. It really feels like you're just pulling numbers out of nowhere here.

That's what the 30x instead of 20x multiple is, for a "growth" tech company.

There's no strong evidence that openAI or anthropic will have non-linear revenue growth, so I'm not sure what point you're trying to make. Unless you think they'll fire all their engineers and replace them with agents or something crazy?

Well if we make AI that learns at the rates humans do, it'll fundamentally undermine and destroy the relevance of all existing AI. It sounds to me like you're saying "we basically are there it's just a matter of degree" and I'm saying "No it's orders of magnitude off and probably won't be using LLMs at all and maybe a very fundamentally different type of neural net technology that hasn't been invented yet."

I blame youtube. I feel like any online service provider has a duty to verify your identity on transactions over $10k USD by default and that should be enshrined in law. Honestly maybe lower.

Well except AlphaZero played 44 million chess games in that time (and actually played with a 44 core computer). So I'd like to point out that the human is still just a few orders of magnitude more efficient.

I looked at the math and I think it's true. Remember revenue is just sales, not profit. These labs are shooting for > $1T valuations, which traditionally means your PROFIT is at least 1/20th or 1/30th of that (so let's say minimum 30B$/year PROFIT).

These companies however are LOSING money (anthropic tries to make it sound like it's profit by deviating from accepted accounting principles) and subsidizing these models. When accounting for all the engineering salaries, training, GPUs, etc, what's their best-case realistic margin three years out, 10%?

So to we'd need a scenario where companies are spending a collective 300B annually on AI (believable) but ALSO that these companies jack up their margins WITHOUT companies switching to the cheaper open-source models (even when there's a $300B incentive to do so).

> The "value" of most knowledge workers -- based on what enterprises currently pay for them -- is $50 - 70 trillion annually.

What do you mean? The sum of ALL US salaries is $13.4 Trillion per year. According to google $65T is the sum of ALL salaries Globally (not just knowledge workers). It's not reasonable to assume AI is a drop-in-replacement for any job yet (perhaps bottom tier customer support from oversees?).

> So if their hypothetical revenues are double-digit trillions and valuations are some multiple of that

So you're sort of premising here than more than 16% or 1/6 of all the world's jobs get replaced by AI. Hopefully you can understand that's both not the current AI capability and also would be a terrible (unprecedented?) economic shock.

So you can see an actual game on that website, and the play seems pretty decent to me for a while (~1700 lichess = 1300 elo) until move 28 when black throws away their queen for absolutely no reason in an incomprehensible blunder.

In some ways this is reflective of the AI experience at large, sometimes shockingly competent but then also sometimes ludicrously incompetent.

I put some thought on this in this comment https://news.ycombinator.com/item?id=49469729

100%. There's an implicit threat as a journalist -- I'll write this story up if I don't get my way.

Correct me if I'm wrong but isn't there an inherent vicious cycle in that if the price gets too high our ability to pay it off becomes uncertain, at which point the premium for the risk would shoot up dramatically (and so on)?

Hopefully it’s apparent to you now at least, that providing names and political affiliations to the current administration is dangerous.

I spent a few hours trying to run all the numbers myself from first principles (e.g. sum of all American salaries is 11.7T, plus international), to see what I could come up with.

What I came up with was Nvidia is maybe slightly overpriced currently, but a bad stock to buy or hold onto because the risk of it shrinking is significantly higher than the risk of it ever doubling again (and it trades at a tech/growth multiple). The case is far worse for Anthropic/OpenAI who, at current pricing, capture such a tiny slice of the pie it's hard to see how they will stay in business long term.

In an alternate universe where all research is completely auditable and we all have infinite time, yes that's a valid approach. And you're welcome to spend your life going that route, but the rest of us are gonna trust our common sense on this.

We need to start using AI to hack our own devices and make our own TV os, sort of like was done for routers. Probably relatively doable these days.

+1 .

I think it takes a certain type of mental or emotional deficit to forfeit all of your free time and the majority stake in your company for funding for a tiny shot at hitting a grand-lam.

"Not everything needs to be a platform". I remember a 10 person company that did translation dispatch for medical needs in California. And the CEO was insisting we needed to someday "become a platform."

I've looked up animal farm in more detail, and it seems like you're fundamentally undereducated on the topic.

Animal Farm entered Anglo-American classrooms in the late 1950s, and its adoption was heavily influenced by the Cold War political climate . The book was seen as "useful" for teaching, but its political utility was part of the appeal—it provided "an easy use... as an anti-Communist and anti-revolutionary lesson" . This wasn't just a grassroots phenomenon; it was actively encouraged by intelligence agencies.

Both British and American intelligence services promoted the book as an anti-socialist message and supported its distribution . The CIA even arranged for a 1954 animated film adaptation that altered Orwell's story—emphasizing that animals on neighboring farms were content and changing the ending to have the animals overthrow the pigs . This film was widely shown in US and UK classrooms, embedding a distorted version of the story in the educational system for years

I agree that a lot of what I didn't like about Animal Farm could have been fixed in a different context (deep dive into the theory of socialism as a classless society, overloading of the word communism, actual history of Russia and how it didn't achieve a classless society, history of Orwell as a person and his own political beliefs, exploration of Animal Farm as polemic tool and analysis of its effectiveness as a tool, analysis of how it's misunderstood today as a criticism of the idea of communism, McCarthyism and how Animal Farm was introduced to classroom during the coldwar)

But here's my main thesis -- we have a situation where not-reading books is incentivized:

In an information-rich environment, I just learned more about animal farm in 10 minutes by asking AI all of the above questions than I did in 10 hours of reading without proper context back in highschool.

If reading itself (and not shaping the values of youth) is really just the goal, then we could solve this by assigning materials to kids that are likely to match their interests (I don't know slenderman, twilight, harry potter, ender's game, whatever).