@RandomRanger's banner p

RandomRanger

Just build nuclear plants!

3 followers   follows 1 user  
joined 2022 September 05 00:46:54 UTC

				

User ID: 317

RandomRanger

Just build nuclear plants!

3 followers   follows 1 user   joined 2022 September 05 00:46:54 UTC

					

No bio...


					

User ID: 317

GDP is a number that correlates pretty directly with the ability of the state to purchase goods and services, such as military equipment

Someone should remind the North Koreans their 'GDP' is small, so they can't provide more shells to Russia than Europe (huge GDP!)

In any case, I'm not sure this is any more evidence of Israel controlling the US government than Ukraine does

Israel gets the most advanced US weapons to fight a few Arabs, while Ukraine gets second-rate equipment, F-16s rather than F-35s, in a war with Russia.

it’s no different to what the US would do if missiles were fired towards Saudi Arabia, the UAE, or any of their other middle eastern allies

The distinction is that all other US allies bring something to the table. Saudi Arabia brings oil and money. Turkey controls a key strait and sends troops to help the US occasionally, though they're not a great ally. Britain, Australia, Canada will send troops to help America too.

Israel only takes. They create enemies for America, they harm collaboration with the Islamic world, they sell military technology to China and frustrate US diplomatic efforts to pull out of the Middle East and focus on Asia. They never send troops to help America, they send shoddy intelligence and suck up aid like a leech. They even got the US to pay off their neighbours too, Egypt and to a lesser extent Jordan get billions in aid for being nice to Israel, the aid started as soon as they signed a peace treaty with Israel.

given how unlikely the claim that the US went into Iraq primarily for Israel’s sake, just a bailey and a slightly less rickety bailey

Go tell that to the neocons, generals, and officials who were there when decisions were made and describe their reasoning perfectly clearly. Did the US go into Iraq to seize the oil, which ended up mostly in the hands of Chinese companies? Or did they go in to fight the Islamist terrorists, mostly of Saudi background and who Saddam was suppressing anyway? Clearly not, instead there's silly discourse about aluminium tubes and other shoddy intelligence, much of which came from Mossad or was used to justify a pre-determined decision. If it weren't for Israeli influence, the war wouldn't have happened.

hasn’t deployed ground troops to take out any the modern threats facing Israel in Yemen, Lebanon, Gaza or Iran

The US has bombed Yemen and Iran, given Israel munitions to bomb Gaza and Lebanon. US troops were infamously on the ground in Lebanon before getting blown up and departing. Just because the Israel lobby doesn't get everything they want all of the time, it doesn't mean their influence isn't excessive.

I think they could've made a better Snow White film than the original, it's just that they didn't want to. They wanted to make a bad film and did so.

Mastery isn't the problem, it's bad people using great resources to achieve bad goals. Now I see it, there's a pleasing symmetry in our tags "Just build nuclear plants" and "nuclear levels of sour" and what we're saying.

However I do agree that there are serious risks with progress and power concentration, it will probably end in tears for the vast majority for us for the same fundamental reason, bad people wanting bad things.

I don't see a collapse pathway though, only greater acceleration. Technology forms society. Writing and agriculture enabled settled states, steam engines enabled modern society. Powerful AI will enable transhuman or posthuman society. Maybe that does look more like an oligarchy where a few enjoy limitless technological power and can suppress everyone else. It may well be bad for those who aren't a chosen few or a singular one. Nevertheless I expect that it'd be much more highly developed than modern civilization in technological sophistication and scale.

Even if there's a full nuclear exchange induced by destabilizing technology, would the survivors really give up on securing more wealth, more power, more security through technological superiority? I believe they'd think 'damn, we should've struck first' or 'this time let's hide our schemes more effectively' or 'at least we've got the most remaining resources, we can try again'. They'd still know all the things we'd know, they'd be back at it again sooner or later, probably sooner and with a more ferocious sense of determination. A full nuclear exchange isn't certain either, it's hard to foresee what happens. I agree that there will be ever-greater instability and disruptions but that's just part of the transition from one kind of society to the next. The general trend is that even occasional setbacks (using rooted in social decline) are overcome - the Bronze Age Collapse, the fall of Rome and the Black Death only temporarily inhibited a larger trend of acceleration. Ideally acceleration should be channelled in a more pro-social way than it is but it seems an irresistible trend. Only if this time is different should we expect it to fail.

It can be true that Microsoft is replacing workers with AI, slimming down a bloated management apparatus and that they're also hiring more cheap H1Bs to watch over the AIs or for internal factional reasons.

I may be in the minority here but I still believe that AI is a big, big thing and advancing fast. No plateau, just steady growth and new avenues to advance down. Two recent papers.

https://arxiv.org/abs/2506.24119

Using SPIRAL, self-play on zero-sum games produces reasoning capabilities that transfer broadly. Training Qwen3-4B-Base on Kuhn Poker alone achieves 8.6% improvement on math and 8.4% on general reasoning, outperforming SFT on 25,000 expert game trajectories. Analysis reveals that this transfer occurs through three cognitive patterns: systematic decomposition, expected value calculation, and case-by-case analysis. Multi-game training (TicTacToe, Kuhn Poker, Simple Negotiation) further enhances performance as each game develops distinct reasoning strengths. Applying SPIRAL to a strong reasoning model (DeepSeek-R1-Distill-Qwen-7B) can still lead to 2.0% average improvement. These results demonstrate that zero-sum games naturally develop transferable reasoning capabilities, highlighting a promising direction for autonomous reasoning development.

https://arxiv.org/abs/2505.03335

we propose a new RLVR paradigm called Absolute Zero, in which a single model learns to propose tasks that maximize its own learning progress and improves reasoning by solving them, without relying on any external data. Under this paradigm, we introduce the Absolute Zero Reasoner (AZR), a system that self-evolves its training curriculum and reasoning ability by using a code executor to both validate proposed code reasoning tasks and verify answers, serving as an unified source of verifiable reward to guide open-ended yet grounded learning. Despite being trained entirely without external data, AZR achieves overall SOTA performance on coding and mathematical reasoning tasks, outperforming existing zero-setting models that rely on tens of thousands of in-domain human-curated examples. Furthermore, we demonstrate that AZR can be effectively applied across different model scales and is compatible with various model classes.

Two major papers on fully autonomous learning, in coding and reasoning via playing simple games respectively. Admittedly both are on a small scale, using small models and little compute. The former used maybe $1000 worth of compute when the big boys have billions to spend. However, recursive AI self-improvement isn't science fiction, it's eminently possible, presumably it's being done right now on a large scale.

Are people really going to be better at programming than a machine that has thousands, millions of years of experience? Surely not, just like they're not going to be better at chess or go or starcraft.

I directed Sonnet to make some software. It's doable. It works. Gets revenue. Not sustainable as yet though.

I think it's a decent employee and very cost-efficient, albeit not perfect.

Anyone else reading that excerpt and thinking 'Based'? Wouldn't it be excellent to carve out a new artificial world, make better animals and plants according to one's wishes? Live as long as one likes without regard for age?

Not the specifics of perfectly cleaning the world, that could take many angles. One might make a jungle of talking animals, or an endless lived-in leafy suburbia or a Willy Wonka wonderland or all of those things simulated within a ball of computronium. But isn't that the logical endpoint of ever increasing mastery and control of the world? What's the alternative, stasis?

I can sense that many people don't like this vision but isn't this what we're doing, irregardless of objections? Unless you think 'no people mustn't live forever' or 'we mustn't have children' or 'technological advancement must stop' then you endorse indefinite growth in numbers and in power of worldshaping and knowledge ("All stable processes we shall predict. All unstable processes we shall control"), so eventually something like this will happen.

less than 1% of Israel's GDP

Forget GDP, GDP is just a number. You can't just go out and buy large numbers of artillery shells, JDAMs, advanced missile defence, spare parts. It's not a liquid market, buying more can just make the price go up. Few countries make these things. Israel can't produce munitions at scale because they're a small country, they don't even have a domestic steel industry. They rely on the US for this because America is actually large and has huge stockpiles that are reliably used to replenish the Israeli arsenal. Otherwise they'd just run out of munitions or Israel would have to wait ages to restock, inhibiting their military capacity.

Furthermore, military aid always roars up whenever Israel actually needs it, it went up to about $22 billion in the year after October 7th. See here: https://watson.brown.edu/costsofwar/papers/2024/USspendingIsrael

Technically speaking, the there is a law against America giving any military aid to nuclear powers who haven't signed the non-proliferation treaty like Israel but they just pretend it's fine.

Was this last bit about the USAF some sort of typo?

No, the USAF and RAF literally, directly, provide air defence for Israel directly. US F-16s shot down Iranian missiles attacking Israel. Plus US warships nearby fire their expensive ABMs to defend Israel.

the US has never invested ground forces in taking out any military group directly opposing Israel

Saddam's Iraq was an anti-Israel force. Israel bombed their nuclear reactor in the 80s. In the Gulf War Iraq Scudded Israel. In 2003, the US invaded Iraq, in large part due to false intelligence about WMDs which the Israelis contributed. Plus a bunch of US policymakers talked about how the real reasoning was that Iraq was a threat to Israel. See my comment here: https://www.themotte.org/post/765/culture-war-roundup-for-the-week/162796?context=8#context

Well I have a vague theory that China will demolish the US military in Asia and create the actual conditions for real political change in the US and elsewhere (military defeat + huge economic crisis are a tried and tested combo), whereupon previously unthinkable options become possible.

But the problem with basing a theory on a hypothetical is that it feels like wishing, the infamous 'my ideology will be the one to arise from the ashes'. Trying to predict the world after an epoch-changing event is like trying to look inside or beyond a singularity. Maybe Trump gets the blame for fooling around and the old regime capitalizes it. Maybe the military gets blamed for losing and the US doubles down on democratic-socialist isolationism. Maybe there's a nuclear exchange. Maybe there's an AI singularity. Nothing is inevitable, even assuming a contested hypothetical.

Of course it'd be good to have more accurate, adaptive ideas flowing more widely. The US does not, in my opinion, need more Haitians, quite the opposite. The US shouldn't be spreading multicultural propaganda around the world, that's not a recipe for good outcomes. America isn't screwed, it's powerful and innovative in many areas. But it's running well below peak performance, there are fractures and internal weaknesses based on unsound ideas of human equality.

As for personal advice, well I've read Nightmare Vision's Rosedale thread https://x.com/GodCloseMyEyes/status/1414619671056297984 and 'Don't make the Black kids angry', it seems pretty clear that black parts of the US, London and elsewhere are dangerous and one shouldn't go there or live there. The author of the latter has seemingly been driven into this state of insanity where he just goes on and on, listing all these grievous attacks and perverse instances where white racism gets blamed for black misbehaviour, one after another after another.

How do you change this state of mind, where people speak in code to realtors because they're not allowed to ask about crime, because it's too racist and discriminatory? Who knows, it's bizarre and weird.

I'm not even American and so my theories about US politics are really limited in skin-in-the-game beyond having a lot of money tied up in US shares. Lots of cool stuff is happening in America, it's a country of contradictions.

if the position is that all US interests are subordinate to Israeli interests

Israel doesn't own the US government but they have enormous influence especially in foreign policy and anything pertaining to Israel. Occasionally the US tries to do something that actually prioritizes American interests over Israel's, the Israel lobby usually nixes this in the end: the Iran deal for instance. Now the US does have huge leverage over Israel in terms of capabilities. Merely shutting off aid would be catastrophic for their military, who relies on US provided weapons, satellites, communications and USAF for air defence. I've said before that the US could annihilate Israel at will with sanctions alone, the state would quickly disintegrate.

But in terms of mental, political, lobbying power, Israel enjoys a huge advantage. The warrior can easily demolish a succubus in battle but it's irrelevant if she has her charm spell running.

I'm not dogmatic on whether it's arm-twisting or owning hearts and minds, there's a mix of both going on. Nor can I give you a date where it suddenly happened, it's not a switch that was flipped on but a gradual process peaking around the 2000s.

But something, surely, has to be off when you've got big figures like Ted Cruz going on interviews about how the Bible says God will bless those who bless Israel, curse those who curse Israel, how he got into politics aiming to be the biggest defender of Israel. Some of this is Adelson money and other Israel lobby cash/threats, some of it is weird Christianity, nevertheless it's unusual and indicates powerful influence.

Low murder rate, relatively rich. HDI is 'very high' what are you complaining about?

I admit that I didn't define the difference between low and high performance but I do strongly think there's a difference between more or less rich, developed countries and places (like South Africa) where the health minister might declare that HIV vaccines are some kind of imperialist plot, or where raping virgins to cure aids is widespread. You can have bad economic policies but still be high performance, all that means is that your abilities are hampered like taking an exam in a loud room. And accordingly Argentina is still decent and safe, they score OK on the test, could be better. The retarded students though, it doesn't matter if the room is loud or quiet, the results aren't going to be good.

Is the US really losing much by banning them from office? All that would happen is some rioting, which can be quickly and easily put down with a little effort. West Africans are notoriously bad at fighting, disorganized and inaccurate marksmen. Of course it's a totally moot point since as bad as West Africans are at fighting, US whites are even less willing to force the issue.

I already answered this. There's no practical program because you'd need a game-changing event for this to be possible. We may as well theorize about the balance of power between Earth and Mars or how to restore the Bourbon Dynasty to the throne of France. Maybe I think the Bourbons would be amazing for France. But I obviously have no practical idea to make this happen because it's impractical and would require an incredible turn of fortune to be even conceivable.

Really don't understand the point of trying to get these 'damning' confessions of wrongthink out of me.

Russia is a high performer, not the best but still clearly in the top category. The US was relying on their spacecraft for the ISS at one point (which Russia helped to make) plus they produce a wide range of advanced technological products - drones, jets, tanks, warships, nuclear reactors. There are little robots transporting food and parcels on the streets of Moscow. Ukraine is similarly a high performer, also possessing advanced industry, they exported an aircraft carrier to China back in the day.

The whole 'Nigeria with snow' argument is profoundly silly. How hard would it be for the US or any major power to wreck Nigeria? Is anyone really worried about Nigeria? How do Nigerian industries affect the world, what ramifications do decisions in Lagos have on anything? Now, how about Russia?

Colombia is not white, it's 50% mestizo, 26% white, the rest being black or indigenous according to estimates.

entirety of the Balkans

Not amazing but still pretty rich and capable all things considered. Serbia is fine, they manufacture cars and pharmaceuticals. The whole 'former Ottoman Empire' part of Europe is less developed and orderly than one might expect from Europeans but it's not a barren gulf of civilization. That's what happens if you have non-European input into a country, you get less European output.

Sure but they're the best of the lot (still no STEM Nobels though). It's not that easy to get to the US from West Africa.

I don't deny that whites who voted for this guy are fools but there is at least potential for good things amongst a broad, non-cherrypicked white population. Build up a power base of elite West Africans at your peril, see what happens if they get you to open the floodgates.

The US can clearly do fine with a modest number of West Africans dragging it down. But if you want first-world performance... If you want safe, efficient, orderly public transport... If you want a lower burden of progressive taxation and affirmative action... If you want crime at civilized, first world levels...

Then you need to address the problem at the root cause. If you let them have political power they'll cause all kinds of problems, they'll West Africanize the country to a lesser or greater extent based on their number, admixture and so on. Bloated and corrupt government is just one and not even the worst problem necessarily.

Consider a thought experiment - what if all the politicians and powerful officials in America had to be black? Give it 20 years for the effects to settle. What do you expect the outcome would be in terms of performance? Would it look more like a high performance country (Japan, Switzerland) or a low performance country like South Africa? Naturally the US has plenty of capable demographics to squander so the decline wouldn't be as severe as South Africa, whose murder rate is actually comparable to the death toll in the Russia-Ukraine war. Nevertheless, there are no white poor performance countries and no black high performance countries. Even on a city level one can observe that having politics dominated by blacks is not a recipe for good outcomes: Detroit.

Now consider the reverse. All the politicians and powerful officials in America have to be non-black. Give it 20 years. Would the outcome be better than the alternate? Is the US really losing much by banning them from office? All that would happen is some rioting, which can be quickly and easily put down with a little effort. West Africans are notoriously bad at fighting, disorganized and inaccurate marksmen. Of course it's a totally moot point since as bad as West Africans are at fighting, US whites are even less willing to force the issue.

I think political solutions can encourage personal/political virtue. Imagine a really intense anti-corruption campaign, where high ranking people were actually given long prison sentences or executed for corruption? Wouldn't that work on the simple, clear level of 'cant commit crime if dead'? China has become less corrupt since the mid 2000s after pursuing this approach.

How does a culture become virtuous in the first place if not severe punishment crushing the bad elements? If the bottom-up anti-corruption from virtue angle isn't working, then one may as well try top-down. In the US this kind of approach is complicated because there are certain groups that are innately clannish and corrupt or so inclined in that direction that it's nigh-impossible to correct. I don't know why anyone expects West Africans to perform well in anything. You can look at West Africans in West Africa and uniformly it's a mess, regardless of history or laws (Liberia stands out here). You can look at West Africans in Haiti - standard West African demographics and outcomes but in the Western Hemisphere instead. And you can look at West Africans with a non-trivial amount of white admixture in the US, plus a constant inflow of white money - much less of a mess but still a mess. Certain parts of Chicago, Detroit, Philadelphia, even areas of Washington DC... these are not places one wants to be!

If you don't want a bloated, grossly inefficient, corrupt government, don't let them have any political power.

Having big muscles does change people's perceptions of you. I doubt he would've gotten this banger song made about him if he looked like a nerd: https://youtube.com/watch?v=vTyeZjo7n_M

He has a space program, I don't believe it's beyond his power to get someone to manage dinner parties easily, or at least more easily than a massive 80 million dollar wedding. If anyone is elite human capital, it's bezos. He can learn!

Nobody goes that far as a matter of convenience. Bezos is not marrying out of convenience, there must be some deeper reason.

First, will it change the outcome? So you ban the big free-to-view US sites. Does this mean that teens will go back to jerking off to pictures of women in swimsuits, as god intended?

They'll hopefully move onto the boorus or hentai or whatever, where at least no real people suffer. It's not exactly educational content but it doesn't weirdly push incest and it's harder to confuse with reality. As far as I'm concerned, killing pornhub would be an unalloyed good. It's not like there's any prosocial value like Fedex or whoever else.

Furthermore, I've read a fair few stories about their business practices that really do resemble the mustache-twirling villain. The verification system they have from producers doesn't seem very effective. From the NYT:

Indeed, one private memo acknowledged that videos with apparent child sexual abuse had been viewed 684 million times before being removed.

Even though it's not possible to crack down on internet pornography, it is at least possible to wipe out the biggest and most obnoxious offenders and rake in a bit of cash too. You're confusing my 'loot and burn' gunboat action with nation-building.

A median porn video teaches a teenage male that of course a woman will be enflamed with desire as soon as you touch her, enthusiastically give you oral sex for a while, then be ready to get fucked however hard you want to fuck her, then happily switch to anal and finally let you cum on her face. Communication about consent, boundaries, or birth control?

Surely people are aware that there's a difference between reality and fantasy? Movies teach me that with the power of friendship and snarky quips I can overthrow giant conspiracies and evil empires. But I don't try that IRL because the evil empire is actually very strong.

I guess there are a bunch of retarded people like the guys who openly masturbate to lingerie adverts on the street or try to rape girls in the park who don't appreciate this fine point. But they're not going to get the message 'don't be creepy and rapey' since they're already locked in on the creepy, rapey lifestyle. 'Why would I not rape when I can, how am I supposed to fuck' they might say and be right, since surely nobody consents to sex with ugly, poor, retarded men. The answer here is using force, not education.

Also, anything teachers or the state try to do will be extremely uncool and cringe. It'll be just like the 'informed consent, no means no' training that nearly every institution has but worse. Can you even imagine how groan-inducingly awful official state-sponsored pornography will be? How woke and diverse and uncool and stilted the dialogue is?

All this would do is teach the more autistic men insanely uncharismatic bedroom talk and make them more nervous. Anyone who wants to actually get off will go to real porn, or hentai, or impossible AI girls. I think it would actually make girls more pressured too, they don't want to be an uncool, frigid, lame bitch like that girl in Safe and Respectful Physical Connection Part 4, do they?

A better solution would be punitively obliterating Pornhub and co with massive fines and lawsuits so they stop profiting off people trafficking and child rape.

Exactly. Korean tests are, by their very nature, actually not good at assessing intelligence. It's all multiple choice. You can't do writing as multiple choice, you need an essay response to assess real comprehension. But if there's an essay then it becomes subjective and open to endless appeals and lawyering, whereas multiple choice is consistent, uniform and totally fair. Just get the right answer loser!

I remember having to study the 'theme' of 'belonging' in some awful book English teachers probably think is profound. That did not help me understand or wield English any better. The only way to get better is to read and write widely, not obsessively study a single text at 10x the intensity of what I was doing just because it's on the test.

Dunno if it's fun but phailyoor will probably blow his gasket at the PM of Albania using ChatGPT so obviously to write out his argument: https://x.com/ediramaal/status/1938739319168024656

Albania's an irrelevant country and English is probably not his native language but really... See this is why I think it's so naive to go 'just tear out the plug'. If ASI or AGI or powerful AI is hostile, it'd be like ripping out your own spinal cord. It'll be AI in your comms, AI needed to even execute the order, AI needed to verify the security of your transmissions and your identity as a legit commander.

I agree, one can also see elements of the pre-WW1 crisis slide ('22 war in Ukraine, '23 Israel war, '25 Iran-Israel conflict), a gradually heightening sense of hysteria about foreign threats and this looming drama of '27 being the year when it all kicks off: AI and China-Taiwan.

I've seen Musk and others doing I/O. I/O is not read/write. The difference is one involves with a widget and your mind that you could otherwise do with your hand and your mind, and the other involves directly reading or changing your mind. When Musk has a working, rigorously accurate lie detector, let me know.

That's like saying you need to have admin access to truly read/write. Just because we can't inspect every part of the memory of a computer, doesn't mean we can't read. You can't go from 'show me even weakly possible' to 'show me a rigorously accurate lie detector'. We can't make a rigorously accurate malware detector for a computer even with admin access!

But in fact, we do not know how to make significant positive changes to the human brain, and we have no idea if significant positive changes to the human mind are possible even in principle.

It's obvious that significant positive changes to the human mind are possible, you can prevent down's syndrome for instance. Or you can find genes that induce aggression and remove them. The Mao-a warrior gene for instance could be altered. That's not a silver bullet but it is something. Genes do things!

The simplest way to improve minds is not to be incestuous, that's a good start. If your benchmark for superior minds is 'everyone being supremely good people out of some morality fable' then sure we don't know how to do that, it would require very sophisticated understanding and practice of genetic alteration. But there's no qualitative difference between simple changes like 'make people less retarded' and 'human perfection', only quantitative differences in understanding and sophistication of approach.

Maybe those things will exist in the future, and alternatively, maybe Jesus Christ will appear in the sky tomorrow to judge the quick and the dead.

You shouldn't compare technologies grounded in progressive realized development to 2000-year old Jewish schizobabble. Wait a few centuries and if there's been no progress whatsoever on the interpretability of the brain (as with the return of Christ), then it might be time to reconsider how realistic these predictions are.

Shame sure can enforce anti-social norms but the problem then is the norms, not the shaming. Every society does shaming in one way or another.

For instance, consider the complex built up in Britain that it was racist to look too closely into Pakistani grooming gangs or consider what exactly was going on with these naked, drunk 12-13 year old girl 'prostitutes' spending all this time with much older men, despite an otherwise powerful feminist apparatus. Shame works both ways. It can support cover-ups and abuse just as it can produce clean, cities full of orderly and considerate people.

Easy paths are all well and good but sometimes one has to do things that are hard, that's where shame comes in. 'Hard' can be doing very good or very bad deeds.

Well we can be pretty sure genetics is the substance behind heredity. I see no reason to give up on mechanistic models when good progress has been made. It's just difficult. Certainly not helped by the amount of fear and politics involved. Gene-editing is functionally illegal: https://en.wikipedia.org/wiki/He_Jiankui

Reading the above wikipedia article is soul-destroying given the context of what we now know about other genetic bioresearch in China in the late 2010s.

To me, this assertion is evidence of the glaring blindspot which materialist rationalists such as Alexander have - they assume that materialism / genetic determinism is right, and then reason backward in order to make their fundamental assumptions fit the data. While the genetic framework is clearly helpful and has had some limited success in new medical breakthroughs, it's beyond obvious to anyone with an ounce of common sense that compared to the hype in the early 2000s, the new branches of genetic science have been a massive let down.

Just because the train is late, it doesn't follow that it will never come or that trains don't work at all. There is a lot of trackwork and bad weather plus the conductor has been derailing it!

Genetics is just really complicated. It is not at all simple like the 'blue eye gene recessive, brown eye gene dominant' charts you might have studied in school. It is not designed to be comprehensible, it's a giant mess that somehow works most of the time.

Why don't everyone's kids look like models? Why are some people born retarded goblin-creatures with gruesome, deformed faces? Why are people dying of old age? Because we don't have a good understanding of genetics, because it's just very difficult. Nobody even knows what 98% of the genome does, it was thought to be 'junk'. We know about as much about genetics as we know about the high-level structure of the universe, nothing of any significance. There too, 95% is 'dark'.

Not to mention that measuring intelligence is complicated, whether it's people or AI. Intelligence is a vibe, a fuzzy, qualitative thing. You can tell the difference between smart and dumb, that is immediately obvious. But quantifying it is very hard.

It is completely understandable for the genetic basis of intelligence to be very murky and unclear. Meanwhile, heritability is possibly the oldest branch of biology. Animals were being bred millennia ago, we know it works, few things have a stronger basis in fact.