quiet_NaN
5m ago
Agreed. My feeling is that OpenAI is burning through venture capital faster than any company in history. If they are selling inference for more than what it costs them in chip deprecation and electricity, that is only because they have a moat in the form of good models. If they ever decide to stop burning through money to make more powerful models, they will quickly find that without that moat they will only be able to charge the same as any rent-a-chip company.
For the most part, the investors do not care about OpenAI being able to sell anything at a profit in 2025. They are simply purchasing stakes in the ASI race. If OpenAI wins that race and alignment just happens, they will be the nobility under god-emperor Altman. If LLM progress plateaus and the singularity fizzles out, their stock will likely crash like the internet companies in the dot-com bubble.
FirmWeird
Randomly Generated Reddit Username
7m ago
I'm talking about Labour movements and politics (i.e. how the modern day Anglosphere Labour parties got started). Left wing populism gave us the 8 hour workday and 5 day workweek, and I'm personally glad that I don't work the 12-hour shifts and 6-7 day alternating workweek that private industry would prefer. As for state owned businesses I don't think that you can really say they all perform poorly - there are plenty of them that do incredibly well. Singtel has done so well that it has actually bought and acquired a decent portion of the private cell companies in other countries, for instance. And as for Bezos, isn't a large portion of his workforce reliant on welfare to survive anyway? Amazon is the worst of all worlds - the public purse is subsidising all their expenses in exchange for no return at all.
OliveTapenade
10m ago
Well, he said, "The United States should stop with this half ass shit... If the US decides that you are deserving of its wrath there is no resistance, there is capitulation or everyone dies."
I asked a clarificatory question: "Er, are you advocating that the US should only do nothing or destroy its enemies utterly? And if the standard for utter destruction is astronomically high, doesn't that imply that most of the time the US should do nothing?"
His response to this question was: "Errr... um...errr.... ummm....uuuuur... Correct."
I took that to mean that, yes, his position is as I described it - that the US should either do nothing, or completely annihilate its enemies with nothing in between.
I believe that the point in the Starship Troopers passage, and the metaphor of punishing a baby by cutting its head off, is an effective argument against that position. Sometimes a military should enact a level of destruction that stops somewhere short of "everyone dies" (zoink's words) or "utter destruction" (mine), because the policy goals that a nation might wish to achieve with military action might be, well, something other than complete annihilation of its foes.
Now to his credit zoink seems to back off from his statement and say that he was using bombastic rhetoric. I'm not entirely sure what his actual position is - he rejects the child comparison but concedes he was using extreme rhetoric, but does he concede the actual point of controversy, that is, that some mission profiles call for less than maximum force, and that is desirable for the US military (or any military) to be able to exert controlled force for limited effect? But I stand by what I said as being a reasonable interpretation of what he had said at the time.
George_E_Hale
insufferable blowhard
18m ago
I got it, just going for low comedy.
birb_cromble
26m ago
You've reached a level of shitpost where I can't tell if you're ironically saying that WWII was civilized or sarcastically implying that it was not. Would you mind clarifying that a bit so that I can understand you better?
quiet_NaN
29m ago
I think that one aspect is the question which performance you actually require from the model.
A fundamental difference between free / open source software and open weight models is that for software, the bottleneck is mostly developer hours, while for models, it is computing power on highly specialized machines.
For software, there have been large fields of application where the best available options are open source, and that has been the case for decades -- for example, try even finding a browser whose engine is proprietary, these days. (Of course, there are also large fields where the best options are all proprietary, because no company considered it strategically important to have open source software, nor was it a fun project for nerds to play with, e.g. ERP software or video game engines.)
For LLMs, tens of billions of dollars worth of computing power have to be sacrificed to summon more powerful shoggoths forth from the void. For the most part, the business model of the AI companies which produce the most advanced models seems to be to sell access to it. If Llama or DeepSeek had happened to be more advanced than OpenAI's models, their owners would not have published their weights but charged for access. (The one company I can imagine funding large open-weight models would be Nvidia, as part of a commodize your complement strategy. But as long as no AI company manages to dominate the market, it is likely more lucrative to sell hardware to the various competitors than to try to run it yourself in the hope of enticing people to spend more on hardware than on model access instead.)
That being said, for a lot of applications there is little gain from running a cutting edge model. I may be nerdier than most, but even I would not care too much what fraction of IMO problems an AI girlfriend could solve.
birb_cromble
39m ago
And I'd feel bad responding to a long text, just to discuss a minor part of it, as that's just nitpicking or directing the conversation towards what interests me but which may not interest the other person.
Please don't feel bad about that. I frequently find that the most interesting content is 1 - 2 branches off of a top level post.
aqouta
39m ago
Left wing populism is personally advantageous for everyone who does not have so much wealth they never need to work again.
This is straightforwardly not true. State owned businesses perform poorly. Europe which has much more left populist crap is a decaying retirement home. Like most populism leftwing populism is very specifically selected for what scratches the grievance hindbrain of the most people listening to just so stories about how homelessness is really caused by the fact that Bezos has a really big yatch.
TwiceHuman
56m ago
Isn't this limitation a part of the map rather than part of the territory? Language is limited, logic is limited, math is limited, etc, but reality doesn't particularly care about the mental jails which we create. I disagree with your earlier comment that understanding aspects of the world in depth is impossible, but I do believe that knowledge alone is insufficient. A condition you might accept for "understanding aspects of the world" is being able to predict the future, and some great people of the past have made eerily good predictions (I believe Tesla predicted phones and computer monitors, and Nietzsche predicted communism and its death toll. Less impressive works are ones like 1984, but that still requires a good intuition to notice an approaching problem before others). Maybe it seems like a nitpick, but my claim is "0.01% of people have a solid understanding of some aspect of the world", and with how statistics work, the vast majority of people who claim to have these abilities are wrong.
I hope you get to experience something which breaks your models of what's possible. It's a refreshing experience and a great blow to limiting beliefs
MaiqTheTrue
Renrijra Krin
56m ago
I think it depends on the flavor of Protestant. If you’re talking about low church Bible thumping evangelicals, I get it, but I think most high church Protestants respect the councils and the dogmas of the early church. The Anglo Catholic movement actually accepts the dogmas and canons of the first seven councils so they’d be pretty in line with the Roman Church and the various Orthodox Churches. Lutherans still informally accept quite a bit of that dogma through the Augustine Confessions and Book of Concord.
HermeneuticsOfVibes
57m ago
This has been ongoing for far longer than that. Tristan Harris's TED talk outlining how he as a google employee explicitly aimed to manipulate you to maximize your "Time On Site", came out in 2016, and his original internal talk on the subject dates back to 2013.
A few additional data points:
- Twitter started phasing out the chronological timeline in favor of the engagement algorithm in February of 2016.
- Instagram switched from chronological feeds to "best posts first" in the summer of 2015.
- Facebook as of 06/15/15 acknowledged that news feeds action on how long a user reads a feed item, as well as how likely they are to engage with it.
chooky
58m ago
Trying to defund woke by instead funding creationism honestly seems like a pretty likely outcome!
ThomasdelVasto
Κύριε, ποίησόν με ὄργανον τῆς ἀγάπης σου
1hr ago
I figured it was a bit of a joke, heh. It's ok I'm a strange Christian. Even amongst the Orthodox. I've made peace with that fact.
erwgv3g34
1hr ago
You know what would work even better? Destroying academia and burning the university system to the ground.
Declare war on college. Remove all government funding of higher education. Stop issuing student visas. Make hiring on the basis of degrees illegal.
College delenda est. This will not end until the last university building is reduced to ashes, the last college campus has been sowed with salt, and the last tenured professor's head sits on a pike.
Or is it a scam? Are its promised rights lies?
Whether we have a right to life, liberty, the pursuit of happiness, guns, free speech, or whatever else seems pretty clearly to be an axiomatic moral argument, not a factual one. Ought, not Is.
This comment from two years back lays out what I think is a pretty solid argument for the nature of the problem:
Do you believe that it's practical to build and enforce a set of rules that ensure acceptable outcomes so long as they're followed, regardless of the behavior of those operating under the rules? Put another way, do you think loopholes are a generally-manageable problem in rule design?
And the answer elaborated in the rest of the comment comes down to explaining why loopholes are in fact not a generally-manageable problem in rules design. The lie and scam comes from the idea that you can write down a legible definition of rights, and then right down a legible set of rules about how to adjudicate disputes over them, and then by following these rules the rights will be secured, and thus the processes and outputs we observe are simply The Way The Rules Are. Our society is built on the idea that rules work this way, but they really don't.
You can make a set of rules that work when people generally want them to work. Making a set of rules that work when people don't want them to work is probably impossible. Incompatible values results in a lot of people not wanting the rules to work any more, so they don't. There's a term that Moldbug came across awhile back: "manipulation of procedural outcomes". It's one of the most perfect political terms I've ever encountered, and the perfect encapsulation for the nature of the problem. Rules, procedures, exist to secure outcomes, but can be manipulated. Once you grok that, everything else follows with the crushing inevitability of a glacier.
TwiceHuman
1hr ago
I think you're missing the point. If you wanted to talk to your mother, would I be okay with deciding what you were allowed to say? Would Google? The government? As far as I'm concerned, nobody has the right to hinder communication between anyone else. The fact that Google can even read my emails is already a disaster, and I'm quite sure reading your physical mail is highly illegal, and that the reasons behind this decision aren't invalid for digital mails.
The one who listens has as much freedom as the one who speaks
This sounds like the freedom of association? I like that concept. What I dislike is when companies try to decide who I can associate with, as well as who can associate with me.
The internet didn't work like this before the fallacy of association began. The form of the fallacy is "If illegal content ends up on Google, Google is guilty" or "If a person writes a slur in your game chat, your game is guilty", "If you're friends with a sexist, you're likely a sexist yourself", etc. You might have heard other versions of it, like "Roblox is guilty because pedophiles use it" and "Guns should be illegal because criminals use them". The idea is sometimes mocked as "Hitler drank water once, therefore you're a nazi for enjoying water". I believe that a large chunk of all conflict in the world, and the biggest reason that ideological bubbles have become such a problem, is this very fallacy.
OliveTapenade
1hr ago
Likewise letter #29 (1938) where he "should regret giving any colour to the nation that [he] subscribed to the wholly pernicious and unscientific race-doctrine", or letter #100, where he confesses, "I know nothing about British or American imperialism in the Far East that does not fill me with regret and disgust", or the way in letter #71 he speaks of having "a curious sense of reminiscence about any stories of Africa, which always move me deeply" (he was born in South Africa), and so on.
There is a very strong tendency for fans of Tolkien, in the public sphere, to misrepresent what he said and believed, or just ignore it. From more left-wing or progressive sides it tends to be about avoiding the way that he was a devout traditionalist Catholic with everything that implies about things like social policy, sexuality, or gender. From more right-wing sides it tends to be about the way he was equally anti-imperialist and anti-racist.
Perhaps more relevantly to the current discussion, one part of his writing that springs to mind is from The Two Towers, in the confrontation with Saruman. Théoden's response to Saruman's offer of peace runs thus:
...we will have peace, when you and all your works have perished – and the works of your dark master to whom you would deliver us. You are a liar, Saruman, and a corrupter of men’s hearts. You hold out your hand to me, and I perceive only a finger of the claw of Mordor. Cruel and cold! Even if your war on me was just – as it was not, for were you ten times as wise you would have no right to rule me and mine for your own profit as you desired – even so, what will you say of your torches in Westfold and the children that lie dead there? And they hewed Háma’s body before the gates of the Hornburg, after he was dead. When you hang from a gibbet at your window for the sport of your own crows, I will have peace with you and Orthanc. So much for the House of Eorl. A lesser son of great sires am I, but I do not need to lick your fingers. Turn elsewhither. But I fear your voice has lost its charm.
I think the aside is significant. Intelligence or technical wisdom does not confer any right to rule another. Even if the other appears, to outside eyes, to be ruling themselves incompetently or in squalor, superior wisdom does not create a right to dominate. No, not even for the other's own good.
netstack
Texas is freedom land
1hr ago
Reversed stupidity is still not intelligence.
In this case, I’d also expect it to be a flagrant 1A violation in a way which the status quo is not. But I’m not up to date on my “viewpoint neutrality” jurisprudence, and I’d be quite willing to believe that there’s some awful precedent here. It would still be an extraordinarily petty, expensive, short-sighted way to imitate the Cultural Revolution.
TwiceHuman
1hr ago
The shopping mall analogy turned out to be a poor choice on my end, I meant the feeling of the space itself, not any of its functions. A bar, a mall, an airport, a school.. They all feel public, in a way that your bedroom, or a house in the forest 50 miles from any other civilization does not.
Neither problem can be helped, I think. Some people (high IQ non-conformists) are 10 times less common today than they were in the past, so asking for a community with a higher concentration of them than this is a rather unrealistic demand. Also, this website cannot improve much further than this, because the surrounding world wouldn't allow it to be much more based than it is already. As the world gets more connected, the difference between everything decreases, on every scale (cultures, countries, websites, people, ideas, genes, you name it), and then it faces "pressure" from the outside to the extent that it is different. So if your surroundings degrade, you degrade as well. Fighting this is like keeping your house cold in the middle of the summer, or trying to keep a child from learning any swear words.
What I seek may become possible if/when web3 becomes a thing
BurdensomeCount
Unable to escape TheMotte's cycle of Samsara...
1hr ago
I have a Ryzen Max 395+ with 128GB RAM and it runs pretty well; granted I don't use it for LLMs but the humongous amount of RAM is useful more often than one might think.
The_Nybbler
If you win the rat race you're still a rat. But you're also still a winner.
1hr ago
Ken ham
Well, certainly not him, he's Australian. I'm sure we can find some American creationists. Maybe Anthony Watts would like a sinecure in the Climate Studies department.
Almost every claim in this paragraph is incorrect. Every model Openai has trained was profitable, gpt3 and gpt4 both almost certainly made back their training cost. They are pouring an incredible amount more into R&D of course, but that's kind of the point, the market for this stuff is actually red hot. And I genuinely have no clue why you think NVidia only sells to one company, estimates put the amount of chips sold to microsoft(including openai) at
23% of nvidia sales, google at12%, amazon at13%, meta at11%, and Xai at around 10%.More options
Context Copy link