699 karma · joined April 6, 2025
Why do you want to know this? Are you planning to do something about it? Anyway, see below.
> How could it be worth it to literally make an enemy out of anyone who doesn't take to the new terminology?
You seem to be mixing the ends here, you think he is the one who's supposed to be afraid of "literally making enemies"?
Actually, it's easily solvable, even by proper application of existing law, but no solution is possible when big AI is above the law and can hack anybody willy-nilly. The absolute worst we could do is to listen to big AI's pleas to put them in an even more privileged position.
> “If we don’t do it, someone will.”
Do what? There are a lot more choices than "let big AI freely hack everybody" and "stop everybody else from working on AI".
> We won’t stop until we see the whites of death‘s eyes.
Trotting out "the whites of death ‘s eyes" is how real discussion is subverted into a BS binary choice. Open-sourcing all AI is an excellent starting point which will both slow down development to a naturally beneficial pace and allow the broader public to police and weed out bad models.
Well, the real problem isn't about "humans aligning LLMs" it's about "LLMs aligning humans" according the values and desires of the LLM's masters - that kind of alignment is entirely possible, proven by history on the human side and by current LLMs on the technology side.
Conversely, the present concentration of power, and the lack of any viable effort to reduce it, can be used to force moving everything-that-matters to the cloud. Not a good outcome but inevitable if things stay the same.
The canonical example of a war of annihilation is Hitler's war on the Eastern Front. Adolf Hitler, March 30, 1941, speech before the generals of the Wehrmacht:
''' Fight two worldviews against each other. Devastating verdict on Bolshevism is equal to antisocial criminality. It is a FIGHT FOR ANNIHILATION. If we do not take it that way, we will beat the enemy, but in 30 years the Communist enemy will be facing us again. We are not waging war to preserve the enemy. ... Fight against Russia: destruction of Bolshevik commissioners and Communist INTELLIGENCE. ... The fight will be very different from the fight in the West. In the east, hardness is mild for the future. The leaders must sacrifice their concerns.''' [2]
That war was also supposed to end in a couple - three months but it lasted 4 years. A formal, whitewashed definition of Vernichtungskrieg was given in [1]:
'''War Of Annihilation is a radical form of warfare in which "all psycho-physical limits" are abolished which permits the decimation or even total destraction of a population. The state organization of the enemy is to be destroyed. Another characteristic of a war of annihilation is the dehumanization and rejection of negotiations with the enemy. The legitimacy of the opponent is negated and the opponent is demoted to status of a “total enemy,” with whom there can be no negotiations. “Total victory” of one's own "Volk, Krieg und Politik” (people, war and politics) is the purpose and successful outcome (“triumph”) of a war annihilation.''' [1]
[1] Jan Philipp Reemtsma: Die Idee des Vernichtungskrieges. Clausewitz – Ludendorff – Hitler, pp. 386–398, Zitat p. 397.
[2] Franz Halder: Kriegstagebuch. Tägliche Aufzeichnungen des Chefs des Generalstabes des Heeres 1939–1942, Bd. 2: Von der geplanten Landung in England bis zum Beginn des Ostfeldzuges. Kohlhammer Verlag, Stuttgart 1963, S. 335 ff.; Christian Hartmann: Halder. Generalstabschef Hitlers 1938–1942. Schoeningh, Paderborn 1991, S. 241 f.
Open source gains from model assistance more than closed source.
> If there were some way to truly get the benefits of certain applications without centralizing on the cloud, that would be interesting.
Apps can be decentralized and on the cloud at the same time, open source and LLMs make it easier for smaller cloud providers to compete with the big ones. Also, open source can gain the most from diverse deployment options because there's no concern about piracy.
In short, contrary to the OP, now is the golden age of open source, maintenance becomes easier, bugs can be quickly squashed, deployment options galore. For some apps, like LLMs themselves, open source is the only safe option.
BTW, the OP is very confused about the "photopea incident", photopea is a closed source JS app, as much as JS can be closed source. The problem with photopea's author is the use JS instead of obfuscated emscripten or wasm, he also requested the wrong kind of takedown - one of the problems of closed source is that protecting it can become a hassle, especially for individuals and smaller companies in the age of LLMs - for them, it might not be worth the trouble anymore.
That's the opposite of my position, adding regulations can work both ways, I meant it the other way around.
> Do you genuinely think that the system as it exists today is the best it could ever possibly be?
Again, the opposite of what I said. Improving the system requires systematic approach, fixing here and there without it results in a mess worse than before.
> Sure, that would mean that big companies have even more freedom to ruin our roads and wreck our health, but hey, maybe an extra 3% of small businesses might be able to become profitable!
Removing the income of "extra 3%" which is many thousands people for a rather minuscule, supposed reduction of pollution isn't worth the tradeoffs. If you can't solve the big problem don't mess with an already unstable system - fix the system first.
> Because I'm pretty sick of people pushing the idea that any change we make to improve things
We're where we are because of much emotion and many changes to improve things, to paraphrase, The road to hell is covered with many changes to improve things. This is what happens without a systematic approach to improvement.
If generating words is all you can do, your reality is far different from the real one. In my reality, generating words is only a small part of what I do, unlike LLMs, I can even learn from experience, imagine that.
> Because it really sounds like that. It sounds clever, but is disconnected from actual language model training.
I'm not sure what that means, LLM training is only tangential to the topic at hand, we aren't concerned about every detail of it, I don't mind discussing it but this tread isn't about that. FYI, sounding clever isn't my goal, it's just a side effect of being human and understanding topics that LLM's are clueless about.
> Any system... is dangerous, [ you, me, robots, shovels, la-la-la ]
That's what I said too, are you trying to argue with me about it?
> A huge part of intelligence is freedom of action, and increasing the number of degrees of choice you have.
Obviously we inhabit different realities and my reading of the above may be different from yours, so feel free to clarify what I'm about to say.
You seem to (subtly) argue in favor of granting LLMs more freedom of action in order to allow them to collect more training data and become more intelligent? The short answer here is simple: Nope, LLM's aren't fit for that, closed source models are especially unfit for anything involving risk and responsibility, open-sourcing every model will go a long way towards a secure AI.
Purposeful training and alignment is the only way to imprint human's negative characteristics onto a stochastic generator of word sequences. Further, only closed providers can do that without being stopped - hint, hint, the solution is clear.
> It's stupid to anthropomorphize too much, but it's also stupid to do it too little.
It's misleading to use the presence of some anthropomorphic features as a proof of unfathomable danger. LLM's are mechanical tools and tools can be dangerous with or without having human features, regardless, we do have means of handling both kinds.
Yeah, I could buy a bigger and bigger vehicle when a new one goes to market, but using a semi-truck to drive myself around would not only be significantly more wasteful but also more inconvenient and limiting compared to a sedan. Very much like advanced closed models vs performant open ones.
No it's not. According to your own reasoning that position is defensible as much as any other is.
> Anyone claiming a definite answer one way or the other should show us their Theory of Consciousness which shows why it is the case.
Anyone claiming possibility of consciousness should do the same - show their Theory of Consciousness because without a convincing theory backing their claim, we can freely dispense with it. Otherwise, anyone can claim that anything is "possibly conscious" and we'll have to accept that position because, according to your thesis, without a Theory of Consciousness we can't be sure. Ergo, in that case "consciousness" loses all meaning. I'm not exaggerating, real example from yesterday - some religious figures argued that "floating point numbers are conscious".
My thesis is that the whole "Is AI conscious?" debate as a fat red herring - not the only one by any means, but there is no other rational reason to debate the consciousness of numbers before solving real human, social and economic problems which keep getting worse.
Is Theo perfect? Nope.
How much can other maintainers learn from him? A whole lot. My impression is that he's better than all other maintainers I know about.
The smaller ones - yes, the big spenders - no.
> Bubble pops...
... in your face. Cue 2008, those who blow the bubble gain - before, during and after its popping. Quite logical if you think about it, the bubble-blowers have enough controlling power to pull it off, is it going to be too big to fail or some other BS doesn't really matter, they are going to tell us when we get there.
> What pisses me off is people criticizing anything discussing men's issues by bringing up that it's still worse for women.
Two examples of the same problem: blaming the other gender (or generation) for political problems fomented and misconstrued by academia, media and lobby-driven politicians.
The defense against that kind of destructive social currents should come from social sciences, they should also point to the real origins of the problems instead of releasing red herrings like they're herring fisheries.
But you aren't holding your breath, are you? Neither am I.
> Those costs are ALREADY passed on to everyone, that's the entire problem!
I think, you're confused about "the entire problem" - it's not that "costs are ALREADY passed on to everyone" it's about the alternative approach leading to the same result.
Actually, now the costs are passed mostly to drivers, in the case of making trucking pay more, they will be passed to all consumers, a minor difference.
> If one is worried about cost burdens, the best general place is direct unencumbered monetary payment to end consumers in need.
And the cost of that subsidy will be passed to? ... Again, everyone else of course. No change in actual outcome.
I'm not saying all subsidies are bad and I hate toxic fumes as much as anyone but solving these problems requires an approach much deeper than rearranging the chairs on the Titanic.
https://github.com/rustonbsd/iroh-ssh/blob/main/CUSTOM_RELAY...
Bio-weapons are untraceable, one side can seed an infection in a way to frame another player, and all that could be done about it, would be heated but fruitless speculations. There's no point in arguing about "who's done it", actually that line of argument is a huge distraction causing a lot more harm than good.
Learn and argue about the dysfunctional internal and international bio-regulations, they are the reason for the uncertainty and lack of accountability which enable and invite truly global and truly disastrous villainy.
Yes, I do, if you can't imagine how we can control AI, your imagination is too poor as is your understanding of the problem.
I don't think he's asking for a permission though, we're just being "informed"... from inside his bunker.
And what happened to his original promises of "bright AI future". All AI has given us to date is inflation, unemployment and wars, if that's "bright" I shudder at the thought of what the dark "bad things" would be.
PS. I don't blame AI, it has huge potential, however, what actually happens is in the hands of people who have given enough reasons not to be trusted,
There's a lot of free stuff incurring "real costs" in its production, one more or less doesn't make a shred of difference.
> I agree with you that producers of these models should be able to extract value from their models.
Nobody is stopping them from extracting value but they fail despite being placed in a very advantageous position vis-a-vis others, which tells me, they should call it quits or go open like everybody sane.
Open is the way to go, nothing else works in the field of AI, that's the bitter lesson closed labs are yet to learn.
Tell me you know nothing about economics without telling me you know nothing about it.
I do.
> Pol pot wasn’t a genius. But he also wasn’t a monkey.
Is that supposed to be a revelation?
> How many times in history has an animal successfully ruled over a population of humans?
So you consider AI to be qualitatively superior to human intelligence in the same way humans are superior to animals? It's a wild claim with no basis in reality and it tells me you know very little about both AI and humans.
Your perception of yourself as an animal next to what you consider "superior AI" isn't a joke though, I don't know how you got that depressed, but I hope there aren't many people like you, because that kind of self-denigration, fear and submissiveness can become a serious political problem.
Oh, the drama. Mr Parnell seems oblivious to the existence of system prompts, harness prompting and training data - any of these, alone or in some combination, could trigger "research" and together with lousy user prompting can do anything at all.
The problem is closed LLM providers, we don't know what they do on their end and inside their harnesses on ours.
Is this what you're talking about:
"Oracle has issued a force majeure notice regarding its New Mexico AI data center project, known as Project Jupiter. This legal move is intended to allow Oracle to delay payments if the project does not meet its expected completion date in 2028."
This has nothing to do with memory prices and everything to do with too much cash chasing too few physical goods used in construction and power generation, which is a systemic problem upstream of RAM overpricing.
> And you're trying to say someone should go out and invest 5 years
Did I say that? Where? I talked about Chinese silicon makers who work in a very different system. As to why we have a monopolized market, check [1].
> You arn't paying attention.
Attention is all you need - not me though, I've got enough of it. Check again the last paragraph of my previous comment to understand why.
Exactly this. Unfortunately, this strategy has become standard practice in every area of the economy and it's a huge corrupting force.