aqouta
Friends:
User ID: 75
Wasn't Mistral 7B developed for $1 million or so, and Mistral Large for less than $100 million?
Mistral is not a frontier lab and no model they've ever trained would trigger any of the proposed regulatory scrutiny.
Didn't DeepSeek or whoever launch a successful model with a low-end stated cost of a few million dollars?
No they didn't. The number you're thinking of was the money they spent getting an already existing model to work with chain of thought.
but if compute costs really are dropping it stands to reason that developing new models will be cheaper in the future, not more expensive...right?
Ok sure, but in the coutnerfactual world where there is no regulation the current frontier labs will have substantially better models.
I realize the obvious counter-response is "well but those aren't cutting edge" but as far as I can tell, most people using Anthropic and ChatGPT aren't using their cutting edge models.
If they're not cutting edge then the regulations do not apply. They're trying to pace the frontier, not dinky merely hundreds of millions of dollars models.
The entirety of your case is assuming a boat load of contradictory incentives. You've not bothered to actually make a case for how this regulatory capture could work or why it would be preferable to just continuing the status quo.
But it clearly cannot work in this case for the reasons I've listed. So the idea that they're going to try to do that instead of continuing to the trajectory that has them rising faster in value than any other companies in history is hard to justify on grounds other than the belief that going forward without regulation is legitimately dangerous. These are at the very least dual use technologies, the idea that we're not going to regulate them is crazy.
So the dastardly plan here is that the frontier labs destroy their own leads on the theory that they'll be able to control the regulatory state so thoroughly as to choke out all other competition and avoid becoming a commodity. Except when the frontier is paused they'll instantly become a commodity with several other players in the same tier already. This just doesn't make any sense. If progress stalls then the margins on current models get competed down to the marginal inference cost with or without other market entrants. Half or more of the nation hates these companies, the idea that a regulatory state is going to be particularly sweet to them is hard to imagine and you're making a general argument against all regulation. I'm sorry but one way or the other we're going to be regulating the production of models with nationstate level hacking ability. The public and state will not stand for any random with a bone to pick having the ability to shut down the power grid.
As was pointed out by @confidentcrescent big companies typically start as small companies.
None of the proposed regulations apply to small companies. They're about frontier training runs which cost many millions to billions of dollars.
For example, suppose a license is required to buy more than a modest amount of computing power.
If people are doing frontier training runs we indeed want to see them regulated, that's the point. Of course we should not do the regulation badly so that legitimate runs take months or years to get approved but you're just making a general argument against all regulation. These are big boy companies spending big boy money, we can let them speak for themselves.
The delay and prospect of being investigated could easily scare off a lot of potential investors and/or founders.
If being investigated scares you off from being involved in building the kind of thing that has nation state hacking capabilities then I would call that mission accomplished.
I know what regulatory capture is. I'm asking how the type of regulatory capture being discussed could possibly apply to a product class that takes billions of dollars to bring to market in any scenario. Stop pattern matching and look at the actual fact pattern it's absurd. If a small business was able to bring a frontier model to market alone then we have a whole lot of other things to worry about.
Because any regulatory regime is likely to impose costs that they can bear but competitors may not be able to.
Like what? Need a remind you that the table stakes for being a frontier lab are billions of dollars worth of compute?
Nobody has a gun to their head telling them to improve as fast as possible.
The game theory here is obvious. Unilateral disarmament is not a serious proposal.
I don't really see how this is signalling giving up their lead. Giving it up to who?
Google? Meta? Chinese labs? All pauses I've seen proposed are against the training of models past the frontier, not preventing trailing labs from catching up to the frontier.
There are already multi-billion dollar hurdles for new players in the industry to put together enough compute to train a frontier model, the idea that having a legal department is going to be the difference is absurd. This isn't a haircutting industry.
But the reality is, the bigger problem is this group of AI people, who advocate for guardrails, have no idea what they are even asking for. I dont think they know what a proper morality for an AI is, because they largely dont have one themselves.
So you're in full agreement with the Yuddites? We do not even know what to align a smarter than human intelligence with yet alone how to do so? We must pause and find a way to do this? Right?
Why would the current winners of the race want this? This would be catastrophic for their companies as others caught up. Ironically its the current losers that would cynically want this as it would allow them to catch up. I'm so tired of these takes that hallucinate some cynical interest that falls apart in two seconds of thought because it's impossible to imagine that the type of people who have been publicly posting about and discussing the dangerous ai since before the industry existed could possibly care about the singular thing they've been warning about for over a decade.
We the current frontier mousetrap makers think it should be ILLEGAL for anyone to build a better mousetrap faster than we can, for the good of humanity.
Who? Who the fuck are these other better mouse trap makers that they're afraid of? it doesn't make any sense. If signaling that they're willing to give up their lead isn't enough to show you that they are serious about the problem then what possible action could change your mind?
I see the additional religion and race elements as intentional motivators to keep up the fight over the territory, since you want your fighters to believe its an existential conflict. I'm not so sure this conflict would persist if Israel were to cede the contested zones entirely.
The contested zones that the hardline Palestinians want them to concede is all of Israel Palestine from the river to the sea. Surely the Jews of Israel earn themselves no favor from the Palestinians with their flagrant abuses in the west bank but approximately no one in the region actually wants a two state solution.
The implication is either that she wouldn't have a problem if espousing hatred was done by non-whites, or she doesn't consider it to be a problem in the first place. (Non-whites espousing hatred and violence is just a hypothetical.
I'm sure if I dig into this person's post history or knew their views I would find them despicable but this simply doesn't follow from the above quote. There is a more charitable, and I honestly think in this case correct, reading that she views it as a hypocrisy because she, incorrectly, believes that America would not expect people "tear [their] clothes and smear ashes on [their] faces in performative mourning for a [black] man that espoused violence"
I think they've not advanced it an have instead advanced idiotic data center bans.
We are only a hundred years out from the state of nature in the west
?????????????
If you think the tech is inherently going to cause existential crisis or dystopias then the correct policy prescription is an international bilateral treaty to halt scaling a la Yud. Opposing data center construction is like trying to protect your sand castle from the coming high tides by digging a foot deep moat. If we're not going to do something that works I'll take Dario over Xi getting to make the pivotal action. And in the mundane worlds I definitely want the US to be ahead on this critical technology.
I'm describing a trajectory. As we've found ourselves more wealthy and able to automate labor we have gone from occasional cold indifference to deaths of those who can not contribute to willing to spend heroic amounts of effort and expense to wrestle 6 more months for grandma against nature. The callous world you scare yourself with is a fiction of your imagination.
The idea that we're sending disabled and elderly people enough money to live on every month merely so that some shadowy elite might be able to use them for violence is a monument to the insanity caused by populism. There is a much simpler explanation for why we don't cull people who are net fiscal drains. Our nation is actually meaningfully a democracy and we do meaningfully, to the degree it's possible, usually get what we vote for. Now of course politicians lie about what's possible or promise multiple mutually exclusive things so we don't always get exactly what we're promised on the campaign trail, but we don't usually get exactly the opposite, at least not durably.
I think the question whoever in charge of AI would ask is if AI can do our work trivially, what's the point of having so many people around at all?
It is already the case that half of all people are not net tax payers. Where is the culling?exist
So I can only reach as far as my immediate circle.
Organizing against local datacenters, or worse American data centers in general, will do less than nothing against xrisk. It's acting frantically impotently to discharge your anxiety. We don't e on a sprectum where relative resistance to American data center buildout moves any levers on chance unaligned ASI is reached.
Somewhere between 40 and 80% of Americans are already not net tax payers. Being redundant and superfluous is a spectrum that all of us are on. There are precious few who if they disappeared today would break the system badly enough to cause a blip. We're a long way from when most people worked on the farm and had a tangible hand in their own survival. As far as most of us are concerned food production is already functionally automated. Most of the basic necessities are already functionally automated. Besides housing, because it's rivalrous, what it took too survive in the 1800s could be provided by the state pretty trivially and with a bit of focus for a couple years with minimal human input. It's just not the case that humans need to be on the beating pulse of survival to find meaning in their lives. We'll be perfectly find with whatever social hierarchy game we entertain ourselves with once our survival is guaranteed.
I think one of the big problems here is, ostensibly they might be, but people saw how these companies will threaten lawfare against small towns
The small town is the legal system. I'm sure you can marshal some individual situations from the past but seriously, you think the kind of society that is willing to vote for data center moratoriums is going to have judges and juries who will just allow corporations to bully small towns? Anything unreasonably would be national news and have many millions of dollars in legal funds even if it was half misinformation. You can't get screwed over by a lego reseller without being able to raise millions in a go fund me.
And if you're in a historically industrial area, there is probably still the memory of things like '08 where the companies just left for good, leaving the community to clean up the mess.
I'm sure I've read the same articles you have. The biggest example of this is where the prospective Datacenter is the only entity since this event willing to fund the clean up after almost twenty years. Is your argument seriously that small towns should just give up economically and die rather than ever taking a gamble at industry again? Seriously, what possible investment opportunity passes this test? How is this view different from degrowth?
It's a very low trust system, and at least in my opinion, that distrust is more than earned.
Distrust can't be cumulative and global. Someone somewhere is always going to be being betrayed. There's eight billion of us. Yes, investments have failed in the past, it used to be that whole mining towns found nothing and died. That rate of failure has been decreasing over time but we'll still have failures. You can't just cancel all progress forever because it's not a sure thing. We'd never do anything. It's just safetyism. The logical end is everyone being locked in a padded cell as until the last of us dies out because any risk might lead to a tragedy.
We evolved from creatures that hunted in bands and were probably murdered by our kin shortly after we could not longer provide for ourselves. I'm confident in a post work society we can conjure up meaningful games to substitute for the drudgery of what is currently required to upkeep society. Tons of people today make their living in some way derivative of professional sports and many people generate a great deal of their identity from sports teams. Same for celebrities or entertainers. A lot of the economy has already been separated from what is required to survive.
I think a lot more people than you think derive meaning and value from their work. Maybe not a majority, but I'd say a large minority.
I think this is a lot of cope. If we didn't have washing machines I think a lot of people might derive a sense of personal definition from being responsible for keeping people's clothing clean. Now that this labor isn't necesary it is a trivial thing. People will find a way to self define and fine meaning without the need to do labor.
I, and a lot of people, don't even believe AI will ever allow the majority of us to live in luxury. Personally, I believe that if AI were as good as it's touted as being, there will be a culling in the best case scenario.
Why? If it can do our work trivially why would there be a culing? We don't cull people currently who can't or won't do value added labor. Many of these people live lifes of relatively luxury to those in the distant past and that's with few voters feeling direct kinship with them. The idea that people who are themselves threatened with unemployment are going to vote to cull unemployed people beggars belief.
Worst case scenario would still be Skynet.
If you're taking this seriously then data center moratoriums are trivially not sufficient and you should absolutely be advocating for bilateral treaties with china to cease frontier development. If you're not advocating for that then I don't want to hear about random data center blocks which will not and cannot prevent this.
But if nothing changes, and things simply continue to be as they are, it would still be a disastrous result as that would mean that hundreds of thousands of people have lost and will lose their livelihoods and futures because we as a society were sold a bill of goods by an entire industry of charlatans.
I'm sorry, are they delivering the goods or not? Because if they are delivering the labor replacement then they are not charlatans. Your lot have this weird way of accusing them of being hype artists while your premises buy entirely into what they're selling. If they can indeed substitute all labor then that is incredible and they are fully vindicated for the rhetoric. If they aren't then we don't have to worry about any of this. It's a very clear sign that what you're actually concerned about is hurting them because you hate them and are not thinking clearly at all.
Further, it's not like the vision of the future that these heads are selling (the best case scenario being that humans are some form of pet or cattle, to be dispensed on the whim of an AI overlord and the worst having already been tastefully adapted by James Cameron) are particularly attractive to these people either.
Not spending a third or more of your life in toil should be attractive to everyone. Cattle, give me a break, the median American is working for some company whose mission they don't care about doing some cog level work. We work to live not live to work. A world where AI does all the meaningful work and lest the people living in luxory is a strict improvement to 99.9% of people and the remaining people that must position themselves as superior to other for self definition can do some neosport for the same sensation of superiority.
This is notably contrary to your initial statement calling them idiots and claimed there were no rational justification.
I called them barely sentient, not non-sentient. There is a vague and difficult to clean up cloud of noise behind the eyes that still marks them as among the living. The noise is basically ludditism in a pure form that they're too embarrassed to or historically illiterate to own.
If you are trying to get someone to do something for you (in this case let you set up datacenters) it is common sense that you should have a clear answer when they ask what is in it for them.
We're not even talking about the perspective own talking to the perspective data center, we're talking about some third party declaring that the town should be disallowed by gunpoint from even negotiating.
- Prev
- Next

They call for "pacing the frontier" and all the previous proposed bills have had frontier scale compute reporting requirements. Essentially if you aren't using frontier levels of compute and don't believe you're pushing the frontier then you're exempted. We can go over any specific calls for state intervention if you want but as far as I know the labs are not proposing any regulatory framework that applies to smaller models.
More options
Context Copy link