Well, this is a negative result. Yep, Maths explains turbulence (when things go turbulent, stuff heats up instead of cooperating). If the result went the other way, it would have had much bigger implications, at the very least we would have known we have missed something big.
It is neither a full index of all kinds of turbulence that can occur (assuming such a thing exists), nor is it an explanation of the phenomena we've seen where things refuse to go turbulent (e.g. superconductors, because there small perturbations DO NOT lead to turbulence). Now THAT would have been useful. And given the fact that OpenAI needed $22 million of compute to show this one kind of turbulence, I don't think either of those are forthcoming any time soon.
And, sorry to say, but those prices show that beating mathematicians at Math is a very expensive undertaking indeed at $22 million per problem even with OpenAI's supposedly better-than-Astra internal models. It's another one of those AI demonstrations that make you think if they aren't showing the exact opposite of what OpenAI claims they show (you know, that their AI models are hitting the upper limits of what the algorithm can do with near-infinite compute, rather than showing infinite new possibilities)
What remains is just the fact that this is OpenAI attacking one of their customers, and maybe outright stealing from their chats. Given that the ideas were even discussed in mails with OpenAI employees that admit in those same mails they can't do it, mails which were probably then fed into the model that "discovered" this, followed by Sam Altman threatening the mathematician behind the method with "destroy your career" (he even states that it's because the mathematician works for Anthropic) ...
I'm very much not a mathematician, but I find the meta-discussion about this case fascinating, and I am a messy bitch who loves drama.
I don't know whether I'm just paying more attention this time, but I found the discussion on this be a perpetual game of telephone, where people get small, but important, details just completely wrong.
The person "threatening" the mathematician was _sama_; and the person who the threats were being directed _to_ is not an Anthropic employee!
(And the person who _did_ say these things have come out and explained what they meant; whether you believe them is up to you.)
I don't know if this is worse because everyone is so tired/angry at the big AI Labs; whether something about people's reading comprehension and attention span has gotten markedly worse or if this is just selection bias on my end; but it's _very weird_ to keep seeing this.
Since a week or so everything I ask codex to do, no matter how small, uses at least 1% of my weekly limits and like 5% of my 5h limit. It's getting so bad I'm thinking of just canceling my OpenAI subscription, because this has no use anymore.
I get the same. It sits there and spins for a bit then as soon as it spits out something, my 5h is 5-10% lower, whether it's asking it to do code review over a significant code base or just asking it to change a config value.
I also signed up for a new account and it's right back to working how it used to. They absolutely do not consume tokens equally across accounts. I did TONS of work on the new account and barely made a dent, even on Astra. Old account chews through 20% like it's nothing
That sounds like large amounts of context (maybe from memory or history?) are chewing through your quota, context that hadn't yet had a chance to build up in your newer account. That's just a guess, I'm not an account holder there or anything so I can't try to re-create it on my end. You could try to erase history on your old account too, too see if it makes a difference. 'Course, then you'd lose all that history!
It is distressing that so many groups, usually calling themselves "ideologies" or "religions" demand that the countries where they happen to be cancel their own values. ESPECIALLY when it comes to values like equality.
In this case equality between men and women, but let's not pretend that for most ideologies that is any more than just one of many inequalities they would like to introduce. Sorry, "to have respected", as if that means anything different.
Wasn't making street plans from satellite pictures ruled transformative? So if you get an AI to make street plans from Google maps, satellite layer, you're not violating copyright are you?
Bing Maps allows the OpenStreetMap community to trace from aerial_imagery
https://wiki.openstreetmap.org/wiki/Bing_Maps#Aerial_imagery Google doesn't. How much a tool helps somebody to draw the actual roads varies by editor software used, some have plugins. Using AI in bulk isn't allowed afaik.
If you obtain the satellite images directly from Google Maps, you can still have problems because you might not have permission to use the data. That's not copyright though.
When you've got the lawyer budget of Google/OpenAI/Anthropic, you can twist and bend copyright and licences to take what you want with little consequence.
OSM doesn't have the resources to fight Google or similar if they are threatened, so it's better to just avoid using data you don't have the right license for.
Another reason might be that modifications of the map are entangled and can be hard to revert in case of a dispute. It's much easier for the Internet Archive to just drop snapshots of websites/pages (and they do so, AFAIK).
Violating a ToU is legal, but the corporation might ban you in return because that's legal too. All you give up is your ostensible right not to be banned.
Don't terms of use depend entirely on copyright to work? Without copyright ... why would you need to agree to terms of use to use the software/site/data?
And doesn't that mean that terms of use can never impose restrictions stricter than copyright?
No. ToU is a part of contract law. Basically the website is providing a service for you. In exchange for your use of that service, you agree to be bound by the terms. This implicitly creates a contract between you and the website, with the terms laid out by the ToU. This sort of implicit contract is legally controversial - normally for a contract to be valid there needs to be affirmative signatures from both parties, as well as consideration (money) exchanged. In the U.S. at least, "click-wrap" (where you must affirmatively click "I agree") and "scroll-wrap" (where you must scroll through the whole ToU before the product unlocks) ToUs are generally considered enforceable, but "sign-in wrap" (where creating an account binds you) and "browse-wrap" (where there's a ToU link) are very dubiously enforceable. But we're talking about Google here, and they make you click through the ToU and say you agree whenever you unbox a new Android device or create a Google account, and then have it apply to all usage of Google services.
Enforcement can be quite challenging because typically if you provide a service over the Internet you don't have any idea who is using it beyond their IP address and any registration data they provide, but this also means that when companies do have a known violator and can show significant monetary damages, their legal departments come down like a ton of bricks.
Google Maps has commercial licenses to some of their data. Reusing that data without your own license can cause problems you might not like. This mostly applies in cases where it's provable that the data you used would've required a license that you do not have.
I would worry about trap streets. A collection of facts isn't copyrightable, and arguably the network of existing streets shown in satellite photos falls into that category. But if I were the photo licensor, I'd be tempted to Photoshop in a few fictional features that could be protected as "creative expression," or some such BS.
The ship's computer reacts like an (incredibly good) expert system. It refuses to say anything that isn't verifiably correct, and if you get it into a logical inconsistency it essentially throws and error and doesn't elaborate further.
LLMs don't do have those responses. LLMs are like humans: humans respond, correct or not (with humans hopefully if they only know incorrect stuff the response is to say so, but that is not a guarantee, but they will respond). And if you give a human, or an LLM, a logical inconsistency they will simply proceed, whether they detect the inconsistency or not.
This is how LLMs are designed ... and if you look long enough at the human (or animal) nervous system you will eventually realize that this is also the design of our nervous system: if something goes in, something comes out, guaranteed (in fact that's close to the only guarantee). Very different from expert system's "either something correct (according to the programmed axioms) comes out, or nothing".
Hell, if you then look at insect nervous systems, they are also designed that way. The key is to respond to everything. And while reasonable responses are certainly preferred, an idiotic response is still seen as a lot better than not responding at all by God/Darwin. Exactly like LLMs.
Of course, for humans/insects our bodies are what is called "active stable", like a plane. Meaning our bodies damage themselves, and just outright die without constant neural control. Heartbeat. Breathing. Blood flow regulation. Temperature, probably even the immune system. All need constant neural feedback to stay stable, and if that neural feedback totally disappears, we're dead in 2-10 seconds (heart failure), 2 minutes (breathing), a few hours, maybe a day (temperature regulation). Now we have a distributed nervous system, meaning lots of parts can fail semi-independently, for a short while, but even without your cortex operational you die in a few weeks.
The consequences of this are even explored with "I, Borg" (5x23) and the Datalore episodes after that, where the underlying problem is that the Borg try (and fail) to adapt to and to process a logical inconsistency but Soong's androids have no issue with it (with Data trying to help and his brother Lore trying to control them and Data with it)
> It refuses to say anything that isn't verifiably correct, and if you get it into a logical inconsistency it essentially throws and error and doesn't elaborate further… And if you give a human, or an LLM, a logical inconsistency they will simply proceed, whether they detect the inconsistency or not.
Claude has a `End conversation` tool for that. There was a bit of a fuss over it.
And it'll happily give (apparently) wrong (or bafflingly incomplete/confusing) info.
TNG, S4E5:
Beverly: Computer, what is the nature of the universe?
Computer: The universe is a spheroid region 705 meters in diameter.
1) Because AI models are 1000x better about following a problem to its conclusion than coming up with a genuinely new idea.
2) Because if we accept the facts ChatGPT only came up with its "new" idea after being told exactly what the new idea was by a mathematician (OpenAI doesn't dispute this btw). And OpenAIs story comes down to the usual "We didn't look at it, trust me bro", which is made more hard to believe because OpenAI only started their efforts after receiving news of what the researcher was doing.
Oh and OpenAI emphasizes that part of the researcher's progress was made ... on OpenAI.
3) And, probably, the researchers were likely stopped by token limits, and that's the only reason they were slower than OpenAI themselves, which is very, very unfair.
4) OpenAI's story "smells" (like so many AI stories lately). Supposedly the company's team asked ChatGPT about solving millennium problems, and out of all millennium problems it just happens to pick the one where a solution can be found in its chat logs?
5) Yet again it would be in good taste for these AI companies to just give this to the researchers (no shortage of difficult unsolved math problems, so if AI can solve them all, just find another one). But instead, yet again they're fighting about it.
6) OpenAI admits they only went after this problem, with a team, no less, after finding out which researcher went after what problem, because of how they thought it would affect ChatGPT's PR. They are demonstrating, in other words, their willingness to destroy human researcher's reputation for PR wins.
That's getting close to big tobacco level morals right there.
7) If anyone wants to verify how much OpenAI cares about the truth, just ask on ChatGPT about the copyright lawsuit outcome and how it applies to OpenAI.
You'll get EXACTLY the sort of responses you get from Qwen about Tiananmen ("we didn't do it, you have no data, everyone's lying and if we did do it, it was perfectly reasonable because " style argument. Try it)
I don’t know much about NS problem or Euler problem anyway. Can’t tell how significant it is to go from Tristan approach to OpenAI’s results.
It is 10k agents after all with a model that is 2x-3x more powerful than Astra. So claims some intuition is his alone and the model can’t find it independently is Tristan’s belief not the reality, which I don’t find particularly convincing.
I do think that Levent guy isn’t independent and indeed does have an enormous amount of tokens to spend and also on an internal model from Anthropic.
All things considered, I think OpenAI people are competitive, even aggressive. But I don’t think they steal anything. Ultimately, the model’s capability is the real surprise factor here, the fact they could get a solution after all in 3 days, that is the cause of drama, if it is 3 months, then it is a nothing burger, because by then the original results had already been published.
> Ultimately, the model’s capability is the real surprise factor here, the fact they could get a solution after all in 3 days, that is the cause of drama ...
And there we go, a beside-the-point answer that tries to bring it back to the PR claim. And of course, an unsupported PR claim: if the initial insight really did come from a human, that would indicate the model's capability either isn't there or didn't matter.
And sorry to say, but competitive adversarial rewards for humans + access to the answer always ends in peeking. In kindergarten. In universities. In Engineering firms. In Finance. No difference other than the justification afterwards. Humans peek in competitive settings. They just do. The fix is not "trust me bro", the fix is making sure they have no ability to do it in the first place.
And OpenAI's answer to the "did you peek at the answer?" is trust me, bro ... as per usual.
If OpenAI indeed played fair, then this is a serious "dick move", and not at all how academia is supposed to work. AND if they played fair OpenAI effectively attacked one of their customers here.
If OpenAI didn't play fair then this is stealing from their customers. Which at a different level is their whole business model (get training data from customers, who even pay for that, train better model with that data, charge more, repeat).
And despite how I feel about the US and Israel's government I do have my head straight: regime change in Iran would be a 1000x better outcome than any of your suggestions. And for everyone, by the way. For Trump. For Netanyahu. For Europe. For gulf countries. For Americans. For Europeans. For Bahraini, Emirati, Qatari, Saudi AND for Iranians.
The problem is that empirically regime changes very rarely lead to stable liberal democracies. Yeah, a magical societal transformation according to exactly our standards sounds great to us. But that is not how this is going to end.
But who is there to say that liberal democracies are the desired end of goal for all states? That is the problem with liberal democracies it has very little tolerance for almost any other kind of system.
How many wars in recent history waged in the name of liberal democracies and freedom?
> But who is there to say that liberal democracies are the desired end of goal for all states ...
Everyone. That's the essence of democracy. It quite literally is what everyone wants.
> How many wars in recent history waged in the name of liberal democracies and freedom?
Zero? I guess you mean things like the Iran war, if we disregard entirely that Iran has been attacking almost literally everyone, including the US, but also including, uh, Iraq, Israel, Jordan, Turkey, Afghanistan, Syria, Yemen, ... they attacked fucking Argentina (look at the map, then tell me how ridiculous that is, then check wikipedia to see they did indeed attack Argentina)
Sorry I mean "exporting the revolution", with little or not explanation as to why this particular export required killing 600000 Syrian civilians, and certainly 2 million people total.
What I just can't understand is why tankies support Iran. I mean, how can you get less communist than Iran? I mean is it the words they use and the propaganda? Idiots.
Given a choice I would rather live in democratic systems - but to say that everyone wants democratic systems tell you are not as exposed as you think you are. There are many failed democracies which resulted in highly bloated and corrupt governments, and these tend to make the situation worse. And there are many states running alternative governance system with success.
When you are saying everyone, you are basically projecting your liberal values on others and you are convinced they are superior and that is why it is dangerous.
Regarding wars - Iran, Iraq, Libya, and many other examples (coups) were initiated in the name of liberty and freedom. And notice that all those countries are rich in oil.
Regarding Iran specifically, you won't understand until you study the Shai's faith and history which I'm sure you don't have any knowledge or understanding of.
> Given a choice I would rather live in democratic systems - but to say that everyone wants democratic systems tell you are not as exposed as you think you are ...
Nope, just old. Like 40s old. I'm afraid I'm someone who thinks that either people want democracy, or just haven't been anywhere near an actual undemocratic system, haven't seen what happens.
In reality, either people are (a minimum level of) liberal (voluntarily or "because otherwise I'd get arrested"), or they live under constant violence. Those are the only choices, and the way that works for people like tankies or some religious lunatics is that they've never seen up close what violence does on even a medium scale. In other words, the way it works for them, maybe for you, is that it's a fantasy. Not fundamentally different from what I'd loved to have done with Sarah Michelle Gellar, when even back then she was probably way too old for me.
And yeah your conspiracy theories about oil. THE war about liberty and freedom was Afghanistan. You might notice: no oil.
So yeah I'm one of those really irritating old guys that tells you even you want freedom and democracy, even if you think differently. It's not the only thing you want, and of course you now think you want your little fantasy more, but that will disappear forever the first time a police officer shows you what happens when they decide to use violence against someone you know.
Liberal ideology will crush anyone and anything that stands on it's way until nothing left but total inner destructions, the west is imploding from within with their OnlyFans subscribers and decaying family and cultural values, yet they condemn anything and anyone who show different stance.
Of course there are people who are unhappy with the government, and guess what, there are people unhappy in the US too, but Canada and Mexico doesn't go and arm those people, and if they did, I would be more than certain that the US police and military will slaughter them all.
All the countries you mentioned are nothing but proxy to the US empire and their military complex, and anyone with little knowledge of geopolitics understands this.
The Iranian people will never come out to protest again in masses after how the US betrayed them.
The president said "help is on its way", live on TV, they hijacked radio and tv signals in Iran to broadcast it to as many people as possible.
The people of Iran went out and got slaughtered like cattle.
You think the mothers who lost their sons and daughters, be it us soldiers or Iranians, will just forget that their children died because of an illegal, unnecessary war that was planned by political idiots in a signal group?
Thousands of innocent souls died in this war because the US President is an idiot and a liar. The US citizens are held hostage just as much as the Iranian citizens. It’s not the people’s fault. It’s horrible leadership.
No help was sent, instead, the inability of the US army to end this war put the US into a position where the regime in Teheran will come out infinitely stronger than before the war. They indeed have "all the cards".
Khamenei is not threatening to end the Iranian civilisation and bombing drinking water facilities. It’s the aggressor, the United States of America. They are just missing the R before the US.
The best outcome for the rest of the world is not regime change in Iran. It’s regime change in Washington. Preferably non-violent.
> You think the mothers who lost their sons and daughters, be it us soldiers or Iranians, will just forget that their children died because of an illegal, unnecessary war that was planned by political idiots in a signal group?
I think mothers who saw their children get shot will blame the people shooting. Ie. their own country, the IRGC.
No, because they want to know why this happened, and if they had a chance to prevent it, and getting shot by the IRCG is just the consequence of doing what someone else expected of them.
Go out and protest and I will bring you freedom and peace and prosperity. That was the promise these kids counted on. They paid for those lies in blood.
That blood is on the president’s hands just as much as it is on the hands of the IRCG commanders giving the orders and their henchmen pulling the trigger and you can bet every family who lost someone in this war knows that.
Every Iranian I talk to says everyone knew the risks. Everyone who went protesting even pointed out that they were still killing protestors one-by-one from the "woman, life, freedom" protests and Masha Amini from before, in a much worse way. You know, beating them up daily for a few years, raping the women they're executing first, sometimes the men too. You know, what allah commands according to the ayatollahs.
Iran has been killing protestors for 50 years. Executing their communist allies from their side was almost the first thing khomeini did, and they have never stopped since. The two girls from the famous revolution photo, one was executed by the regime she was celebrating.
But everyone knew EXACTLY what the government would do about the protests.
One Iranian I know said they kill a large number of protestors "every 2 years". So even the basic premise of this thread is totally wrong. The IRGC has mass-killed protestors about 25 times before this protest. That didn't stop the protestors, just like 24 times didn't stop them.
Now they have killed protestors 26 times. That currently stops people, but it won't stop them for long either.
Was regime change in Iraq and Afghanistan worth it despite how bad their governments were? Even the party of Bush & Cheney seems to think that was a bad idea and Trump ran on no more wars in the ME.
The problem with that is that certainly for Afghanistan, you know it's easy to talk to Afghans in Europe these days.
They are angry that Afghanistan was left behind and are absolutely overwhelmingly of the opinion that regime change was worth it.
Now if you ask people in the US if it was worth the cost? That's a different story. The world order effectively depends on free access to the sea and so Afghanistan ... it was never, and will never, work economically. So was it "worth it" in dollars? No. And it won't ever be.
If you ask Europeans who insist they "contributed" to the Afghanistan invasion ... well let's just say I've seen several discussions between Afghans and Europeans about whether it was worth it. Those discussions don't work out very well. Afghans are angry not more was done, Europeans retort that a lot was done and it cost too much (the Europeans discussing never actually contributed anything to the effort, of course. Their argument boils down to that they "gave up" government spending (unemployment contributions, education, ...) for Afghanistan and got nothing back)
... and of course tankies believe islamist propaganda. On marxist.com there was literally published that "If Iran wins this war they will complete the revolution and finally we'll get true communism". I mean, read the site, it regularly has, shall we say, questionable logic. But that one, my jaw hit the floor.
Unfortunately Tesla's great accomplishments do not lie in the car space, but in the getting investments space. In finance. Which means they will have the profitability right, exactly, guaranteed.
Take the Tesla metal casting advance (single-press casting, or as Tesla calls is "Gigacasting"): it is mostly the result of investment, more specifically all other carmakers deciding they weren't willing to risk it, no matter how many papers said the simulations were correct. Tesla risked it, it worked exactly like the (mostly German made) theory and simulations said it was going to work, and now everyone and their mother (ie. even non-car companies) are using it. Result: Tesla has a patent on it, that nobody has licensed since they were working on it decades before the first Tesla engineer thought of it. The idea was literally available in CAD software before Tesla started to work on it (not that I want to claim there wasn't a LOT of design + validation + testing + integrating + ... work left to be done)
This principle is generally true for all Elon Musk companies, and most of their accomplishments/vision. SpaceX is not even close to the first to land a rocket, for example (that was McDonnel-Douglas). Nor were they the second, or the fifth. Oh and only one of the companies that attempted it before SpaceX needed 2 tries, all others succeeded immediately, or the company died, some succeeded, and the company still died. SpaceX, by contrast, needed 7 crashes before succeeding for the first time (one of their rockets never tried to land on account of blowing up during ascent, I'm counting that one), and 17 tries before they first tried to move it to production.
I'm not to say integrating everything, massively expanding the scope/scale of these applications, getting the investments, and getting it commercially operating isn't a gigantic accomplishment, but it is not inventing it at all.
Elon Musk's constant talk about inventing and "work from first principles" ... is therefore kind of total bullshit. He has never done that, and neither have his businesses.
It is neither a full index of all kinds of turbulence that can occur (assuming such a thing exists), nor is it an explanation of the phenomena we've seen where things refuse to go turbulent (e.g. superconductors, because there small perturbations DO NOT lead to turbulence). Now THAT would have been useful. And given the fact that OpenAI needed $22 million of compute to show this one kind of turbulence, I don't think either of those are forthcoming any time soon.
And, sorry to say, but those prices show that beating mathematicians at Math is a very expensive undertaking indeed at $22 million per problem even with OpenAI's supposedly better-than-Astra internal models. It's another one of those AI demonstrations that make you think if they aren't showing the exact opposite of what OpenAI claims they show (you know, that their AI models are hitting the upper limits of what the algorithm can do with near-infinite compute, rather than showing infinite new possibilities)
What remains is just the fact that this is OpenAI attacking one of their customers, and maybe outright stealing from their chats. Given that the ideas were even discussed in mails with OpenAI employees that admit in those same mails they can't do it, mails which were probably then fed into the model that "discovered" this, followed by Sam Altman threatening the mathematician behind the method with "destroy your career" (he even states that it's because the mathematician works for Anthropic) ...
reply