EPTArtificial Intelligence Thread - Page 5
| Forum Index > General Forum |
|
GreenHorizons
United States24354 Posts
| ||
|
Manit0u
Poland17843 Posts
On July 31 2026 21:59 Jankisa wrote: Also, I'd push back on the "AI can't innovate", I think that Stockfish and Alphago and Alphastar have demonstrated plenty of innovative and novel tactics a loooong time ago, and again, if you read the summary of how GPT attacked Huggingface it literally found 2 new zero day exploits, in a lot of cases innovation is just trying a bunch of things until one of them works, so AI throwing massive amounts of compute at a problem is basically the same thing as innovation. The bugs it found are just old bugs but in different software. So it did what it's best at: pattern matching. Checked old bugs and tried to find software that might be vulnerable to those as well. People are using AI to find plenty of such things nowadays because finally they can leverage AI to do something that would be too time-consuming previously. | ||
|
Jankisa
Croatia1723 Posts
To gain access, the models identified and exploited a zero-day vulnerability (which we’ve now responsibly disclosed to the vendor) in the package registry cache proxy. So maybe what you said is true and you work in OpenAI and have some sort of inside knowledge, but I honestly doubt that. Also you haven't really addressed the fact that AI's have been coming up with never before seen tactics, openings and strategies for games almost 10 years ago. The list of things AI has done is pretty long at this point, so I'll just go with one I find most impressive because I'm terrible at math, it solving a 80-year-old math problem: “No previous AI-generated proof has come close” to meeting those high standards, wrote Timothy Gowers, a mathematician at the University of Cambridge, in commentary solicited by OpenAI. “This is the unique interesting result produced autonomously by AI so far,” says Daniel Litt, a mathematician at the University of Toronto, who was consulted by OpenAI to verify the proof but is not involved with the company. It's not just about things being time consuming, it can genuinely find solutions to things that humans couldn't for a looong time. | ||
|
Uldridge
Belgium5218 Posts
It will also ultimately use the given ruleset in a way no human would ever think about. So in this sense it's innovating, but it's still within the ruleset we created. I'll be truly impressed when we start seeing advanced maths or whatever that are just completely new (just like how imaginary numbers were created), or if it comes up with a grand unifying theory using completely novel physics to explain it all. | ||
|
Manit0u
Poland17843 Posts
On July 31 2026 23:09 Jankisa wrote: You really don't know that, so you are kind of innovating a narrative in order to reinforce your preconceived notions. So maybe what you said is true and you work in OpenAI and have some sort of inside knowledge, but I honestly doubt that. Also you haven't really addressed the fact that AI's have been coming up with never before seen tactics, openings and strategies for games almost 10 years ago. The list of things AI has done is pretty long at this point, so I'll just go with one I find most impressive because I'm terrible at math, it solving a 80-year-old math problem: It's not just about things being time consuming, it can genuinely find solutions to things that humans couldn't for a looong time. Have you checked those stories? For 6 Erdos problems "solved" by AI it turned out 5 were already solved in mathematical literature previously. Just not associated with Erdos but as other things so people just missed them. Most of those problems are also deemed not actually worth solving by humans because the required time investment is not proportional to the complexity of the problem (need to spend a lot of time doing tedious things to prove/disprove something). You don't have to take my word for it, here's a CS professor explaining it: It's like I said previously. AI is great at scouring large amounts of data and connecting the dots there or doing simulations of stuff that would take humans too long to be worthwhile. That's how it finds the bugs in the code, that's how it solves those "unsolved" mathematical problems. Humans simply don't have the capacity to access and keep track of so much data at once so a lot of existing mathematical problems have already been solved by people who simply didn't even know they were a problem while doing something else. And then it gets lost because people trying to tackle the problem don't necessarily follow seemingly unrelated works so the problem stays open. In this way AI is a great tool for assisting with such things but it's not really solving anything by itself. Even for the one problem that wasn't "solved" yet it didn't disprove the original thesis but instead proposed an alternative. And regarding the HuggingFace hack: First, circumventing internet restrictions and hacking into servers are exactly the kinds of things these ExploitGym systems are designed to do. There was no “rogue” agent or revelation of some surprising, devious new capability. Second, the real issue here was OpenAI’s sloppiness. What makes ExploitGym a hard benchmark is that there aren’t supposed to be humans in the loop–you have to let your harness and LLM act entirely on their own, coming up with long-time-horizon plans and executing them autonomously. (When professional programmers use coding harnesses, by contrast, there’s plenty of human oversight, as LLM-based plans are often misaligned with our intentions, or just plain weird, and need correcting.) AI companies know LLM-based plans are pretty unpredictable, so they’re usually pretty careful about how they set up and restrict the harnesses and LLMs used in ExploitGym-style tests. According to reporting from the Financial Times, however, OpenAI recently started playing fast and loose with these safety principles in a bid to catch up to Anthropic in this area – Anthropic having received a lot of cybersecurity street cred from the buzz surrounding its Mythos release. The Financial Times noted that OpenAI had used “increasingly aggressive training methods in its race against Anthropic,” and that it had been warned that their approach could lead to a “breakaway hacking incident” after early tests showed it lacked the right safeguards to block plans that involved bypassing constraints in the test environment. Given these concerns, OpenAI staff were reportedly “unsurprised” by the Hugging Face incident. In other words, the AI companies running these challenges already knew that, without care, these unpredictable autonomous hacking systems might bypass constraints and attack systems you didn’t intend to target. Blindly implementing an LLM-generated plan with a powerful harness is dicey – not because the LLM might develop malicious intent (this is a nonsensical notion given their static architecture), but because as any ChatBot user knows, LLMs are unpredictable. OpenAI wasn’t sufficiently careful, and got burned. | ||
|
BradTheBaneling
44 Posts
On July 31 2026 23:17 Uldridge wrote: It has a lot of time to bruteforce on thing that already exist. It will also ultimately use the given ruleset in a way no human would ever think about. So in this sense it's innovating, but it's still within the ruleset we created. I'll be truly impressed when we start seeing advanced maths or whatever that are just completely new (just like how imaginary numbers were created), or if it comes up with a grand unifying theory using completely novel physics to explain it all. I don't understand how "innovating within a ruleset" is different from complex numbers. Complex numbers are a natural result of the reals and algebra. I also think going from complex numbers to a GUT is like going from fire to the Apollo 11 missions. I don't know if your modern example is a very fair bar to set. Arguably all math after we accept some set of axioms is just 'innovating within a ruleset'. If you wanted to downplay it, I think the better argument to use is that LLMs appear to be far better at doing things like finding counterexamples vs. finding proofs. | ||
|
Uldridge
Belgium5218 Posts
On July 31 2026 23:56 BradTheBaneling wrote: I don't understand how "innovating within a ruleset" is different from complex numbers. Complex numbers are a natural result of the reals and algebra. I also think going from complex numbers to a GUT is like going from fire to the Apollo 11 missions. I don't know if your modern example is a very fair bar to set. Arguably all math after we accept some set of axioms is just 'innovating within a ruleset'. If you wanted to downplay it, I think the better argument to use is that LLMs appear to be far better at doing things like finding counterexamples vs. finding proofs. You say that like it's obvious to to just expand how things work with the "same rules" when complex numbers allow you to do things that weren't possible before they were introduced. The beauty is it having all make sense. If you introduce something new and it breaks something else and then you need to patch that... you get modern game balance teams. But in all seriousness, I think one needs to be quite smart and creative to be able to innovate instead of derive, which, to g8ve credit, can sometimes be quite innovative in its own right lol | ||
|
Jankisa
Croatia1723 Posts
On July 31 2026 23:40 Manit0u wrote: I guess we have a very different interpretation of what is impressive and how innovation works. Even then, you consistently skip over the examples I provided with AI systems coming up with novel ways of playing the games of Chess, Go and Starcraft 2, to varying success. Also, your quote about the HF incident didn't really say anything except what I noted when I first posted about it, that OpenAI and other labs are very lax with their controls. The AI (and Mythos and other systems) are by far, from all human activities biggest experts in coding, and when tasked with breaking systems they do find novel exploits, you asserted that they used known ones without any proof or source, when I asked for it you provided a quote that does not support what you said. Asking for some "breakthroughs that no one saw coming" in the age where these are incredibly rare and have been mostly incremental and "group projects" for us humans is in my opinion, at this point especially kind of silly. LLMs went from having trouble solving extremely easy math problems 2 years ago to solving things that 80 years of mathematicians trying couldn't, it's weird to dismiss them. | ||
|
BradTheBaneling
44 Posts
On August 01 2026 00:29 Uldridge wrote: You say that like it's obvious to to just expand how things work with the "same rules" when complex numbers allow you to do things that weren't possible before they were introduced. The beauty is it having all make sense. If you introduce something new and it breaks something else and then you need to patch that... you get modern game balance teams. But in all seriousness, I think one needs to be quite smart and creative to be able to innovate instead of derive, which, to g8ve credit, can sometimes be quite innovative in its own right lol I mean the idea of complex numbers is just an algebraic closure of the real numbers. Complex numbers are older than the fundamental theorem of algebra. I don't really know what "when complex numbers allow you to do things that weren't possible before they were introduced" means. Complex numbers were discovered because we were trying to solve cubic equations. The beauty is it having all make sense. If you introduce something new and it breaks something else and then you need to patch that Again I just don't understand what this means. Are you suggesting that the mathematical results derived from LLMs so far are being falsely verified by mathematicians? What part of mathematics are you suggesting that is has literally a single iota of relevance to? But in all seriousness, I think one needs to be quite smart and creative to be able to innovate instead of derive, which, to g8ve credit, can sometimes be quite innovative in its own right lol This just feels like a goofy sort of philosophical argument. How do you define innovate and how do you define derive? It feels like you're saying complex numbers were 'innovated' when you could easily argue (and I'm being particularly non-rigorous here) that they were derived from having proven solutions for specific cubic equations and then being able to demonstrate that those equations were also equal to simpler equations of real numbers and negative square root numbers (i.e. complex numbers). | ||
|
Manit0u
Poland17843 Posts
On August 01 2026 01:39 Jankisa wrote: I guess we have a very different interpretation of what is impressive and how innovation works. Even then, you consistently skip over the examples I provided with AI systems coming up with novel ways of playing the games of Chess, Go and Starcraft 2, to varying success. Also, your quote about the HF incident didn't really say anything except what I noted when I first posted about it, that OpenAI and other labs are very lax with their controls. The AI (and Mythos and other systems) are by far, from all human activities biggest experts in coding, and when tasked with breaking systems they do find novel exploits, you asserted that they used known ones without any proof or source, when I asked for it you provided a quote that does not support what you said. Asking for some "breakthroughs that no one saw coming" in the age where these are incredibly rare and have been mostly incremental and "group projects" for us humans is in my opinion, at this point especially kind of silly. LLMs went from having trouble solving extremely easy math problems 2 years ago to solving things that 80 years of mathematicians trying couldn't, it's weird to dismiss them. I'm sorry but finding new way to play chess isn't really impressive in my eyes. It's not the most complex game out there and like it was mentioned it can be brute-forced since there are no random factors involved. Computers have been beating people at chess way before AI. As to not providing the source for the bugs I don't really have time to go through the network security articles and videos I've been through lately to find where it was mentioned because it's also mostly inconsequential. And the claim that AI system are "by far biggest experts in coding" is actually laughable. I work as a senior software engineer and I see what AI can and can't do on a daily basis. If you're writing a blog post or a to-do app sure. But if you need anything sufficiently complex and something that you want to be able to collaborate on with others and maintain in the future AI falls woefully short of expectations. It's ok as a tool to help you analyze a big codebase you're not 100% familiar with yourself and give you some hints about how things work but it is itself unable to produce code that would be up to the quality and standards required for bigger projects. | ||
|
Jankisa
Croatia1723 Posts
On August 01 2026 02:50 Manit0u wrote: As to not providing the source for the bugs I don't really have time to go through the network security articles and videos I've been through lately to find where it was mentioned because it's also mostly inconsequential. And the claim that AI system are "by far biggest experts in coding" is actually laughable. I work as a senior software engineer and I see what AI can and can't do on a daily basis. If you're writing a blog post or a to-do app sure. But if you need anything sufficiently complex and something that you want to be able to collaborate on with others and maintain in the future AI falls woefully short of expectations. It's ok as a tool to help you analyze a big codebase you're not 100% familiar with yourself and give you some hints about how things work but it is itself unable to produce code that would be up to the quality and standards required for bigger projects. Let's do it step by step. OK, you aren't impressed by Chess, how about Starcraft? Is that also a very easy game? Since you are a programmer, you should also know the difference between the Deep Blue way of bruteforcing chess (which is ironically how you seem to understand all AI it seems) and what Stockfish is doing. On the HF incident, come on man, that is such a cope out, it's fine to admit that you made shit up. There is no source for what you claimed (specifically that AI used 2 known exploits for a different system) because you made it up, if you didn't you could just put that sentence in AI and find the source in 30 seconds, I mean you could also find it in your browser history, if it was real. As a senior Sysadmin I can tell you that AI finding 2 zero day exploits just to break containment over a weekend is not inconsequential, it's impressive as fuck. I also did not say that AI's are "by far biggest experts in coding", I said that from all the things that they have been trained to do, they are excelling or the biggest experts in specifically coding, not math, not physics, coding. I've heard this excuse you are using on how AI is not that good for programming from senior developers that are falling way, way behind guys who are actually embracing AI in my company, the difference between them and their peer who does use it is staggering, and we work on a fairly mature product with a huge code base. As long as you know how to approach it, segment it and focus the AI to do exact things, it's amazing. The most senior guy, the boss of both of the 2 guys I referenced above said AI writes about 80 % of his code, and I believe him, because he knows how to use it and the results are spectacular and fast. | ||
|
Manit0u
Poland17843 Posts
The hacking incident only shows the unreliability of AI. Since it hacked the proxy in order to get to the internet to look for answers instead of hacking the target system it had access to which was its original task. Doing new strategies in Starcraft is also not at all impressive to me. Of course you'd come up with new strategies when APM and being able to see only 1 screen at a time is no longer a limitation. I really don't know why people get so impressed by computers doing computer games better than humans. You no longer need the interface and you get access to all the underlying information that humans can't see. Aimbots are banned in FPS games for a reason. Most people probably don't remember this but Virtua Fighter 4 had a system for training your own AI for different characters and you could pit them against other people's AIs on the ladder, they'd get ranked etc. Also, in the arcade mode final bosses were characters with AIs made by top ranked players so ultra tough challenge. That was back in 2001... Personally I don't believe the hype and think we're building and using the AI wrong but time will tell. | ||
|
Uldridge
Belgium5218 Posts
On August 01 2026 02:35 BradTheBaneling wrote: I mean the idea of complex numbers is just an algebraic closure of the real numbers. Complex numbers are older than the fundamental theorem of algebra. I don't really know what "when complex numbers allow you to do things that weren't possible before they were introduced" means. Complex numbers were discovered because we were trying to solve cubic equations. Again I just don't understand what this means. Are you suggesting that the mathematical results derived from LLMs so far are being falsely verified by mathematicians? What part of mathematics are you suggesting that is has literally a single iota of relevance to? This just feels like a goofy sort of philosophical argument. How do you define innovate and how do you define derive? It feels like you're saying complex numbers were 'innovated' when you could easily argue (and I'm being particularly non-rigorous here) that they were derived from having proven solutions for specific cubic equations and then being able to demonstrate that those equations were also equal to simpler equations of real numbers and negative square root numbers (i.e. complex numbers). Well the heart of the argument is a particular non-rigoruous one I think: what exactly is innovation? Would LLMs be able to "invent" scissors to cut paper when presented with the problem that paper needed to be cut, or would it come up with a generic compounded answer of things already existing to hamfist its answer into a thing that does the job (but not really)? Or would it be able to "innovate" the use of steam as a way to move vast amounts of work and come up with the concept of a steam engine? The only thing being asked here is to extrapolate this to our current technological era and ask if it can come up with concepts no other human has come up with so that a new step in the tech tree is unlocked or contribute to our understanding of the universe. I've never claimed anything about the mathematics of LLMs, I'm sure they're quite intricate and complex systems, which I can't be bothered at the moment to understand the architecture of. Innovation and derivation are difficult to pin down, but I hope you more or less understand what I'm trying to say. | ||
|
Acrofales
Spain18442 Posts
On August 01 2026 17:43 Uldridge wrote: Well the heart of the argument is a particular non-rigoruous one I think: what exactly is innovation? Would LLMs be able to "invent" scissors to cut paper when presented with the problem that paper needed to be cut, or would it come up with a generic compounded answer of things already existing to hamfist its answer into a thing that does the job (but not really)? Or would it be able to "innovate" the use of steam as a way to move vast amounts of work and come up with the concept of a steam engine? The only thing being asked here is to extrapolate this to our current technological era and ask if it can come up with concepts no other human has come up with so that a new step in the tech tree is unlocked or contribute to our understanding of the universe. I've never claimed anything about the mathematics of LLMs, I'm sure they're quite intricate and complex systems, which I can't be bothered at the moment to understand the architecture of. Innovation and derivation are difficult to pin down, but I hope you more or less understand what I'm trying to say. Why do you think scissors of all things are the pinnacle of innovation? It's two knives on a hinge. Once you have knives and hinges, you have scissors. Knives were one of the first tools invented due to great necessity. Would AI be capable of inventing knives, having seen teeth and claws? And would it figure out the material science needed to bang rocks together to sharpen one of them? It seems the latter question is one of embodiment: it requires being able to experiment. That is something that is very much up in the air. However, looping repeatedly to solve a coding problem is not dissimilar, just manipulating language instead of rocks. Hinges have also existed for thousands of years, and are, in their basis, two planks with holes in them and a peg through the middle. But going from "I need two things to turn around each other" to having a material solution might be what you'd call innovation. Of course, I think it could be an interesting experiment, but am unsure how to remove from a training set all mention of hinges or things that look like hinges, and seeing if an AI can come up with hinges from first principles. Not easy to do, though: how on earth do you ensure all prior knowledge of how hinges work is removed? Steam power also isn't exactly complex. It's difficult, but not complex. It is also very much underspecified as an "invention": there were hundreds or even thousands of iterations improving the steam engine after the Aeolipile was first conceived of in the first century AD. Advances in science, engineering and metallurgy led to various improvements before and after Watt "invented it". So at what point was the steam engine invented? What existed before that point? Would you call that innovation? Or derivation? E: reading up on scissors, spring scissors predated pivot scissors by a 1000 years or more. So jamming a hinge in there was more an efficiency upgrade (similar to the iterations on the steam engine I discussed afterwards), than the innovation itself: you have spring scissors and hinges and you invent pivot scissors. Spring scissors are simple once you have basic metallurgy and are yet another problem of: would an AI be able to (1) experiment, and (2) recognize useful progress in experiments? I think the answer is a qualified "yes". I say this based on having experimented with a loop autoresearch architecture powered by Opus 4.6. When I have more time I'll go into the qualifications, and also try to respond to Manitou. | ||
|
Slydie
2007 Posts
This was a task limited to the digital world. Afaik, llms do not understand what it types, so they can not understand anything about the real world either. They can fake pretty well, though, as their data sets are large. The next step I am curious about is what happens when the actual cost of running AI needs to be paid by someone. The circular investments between big-tech companies will dry up eventually. Even if AI does real work, there are real costs too, and using humans might often be cheaper. | ||
|
Uldridge
Belgium5218 Posts
| ||
|
BradTheBaneling
44 Posts
On August 01 2026 17:43 Uldridge wrote: Well the heart of the argument is a particular non-rigoruous one I think: what exactly is innovation? Would LLMs be able to "invent" scissors to cut paper when presented with the problem that paper needed to be cut, or would it come up with a generic compounded answer of things already existing to hamfist its answer into a thing that does the job (but not really)? Or would it be able to "innovate" the use of steam as a way to move vast amounts of work and come up with the concept of a steam engine? The only thing being asked here is to extrapolate this to our current technological era and ask if it can come up with concepts no other human has come up with so that a new step in the tech tree is unlocked or contribute to our understanding of the universe. I've never claimed anything about the mathematics of LLMs, I'm sure they're quite intricate and complex systems, which I can't be bothered at the moment to understand the architecture of. Innovation and derivation are difficult to pin down, but I hope you more or less understand what I'm trying to say. You're hyper focused on physical stimuli for a reason that I can't really understand. Would an AI meerkat spend as much time watching for danger as a non-AI meerkat? I just don't know what it's supposed to ask; maybe an AI would invent scissors, the steam engine and spend an equal amount of time watching for danger. I don't understand how any of these could be delineated as a more or less useful question to ask. I know we're on a forum for video games but I'm sure you're cognizant of the fact that human technological progress is not a "tech-tree". It's important because you'll make a statement like "if it can come up with concepts no other human has come up with" and it seems that your base test is: AI must literally invent new physics that describes the entire universe or go back in time and invent scissors. These things were not just invented from a person locked in a cave who, in a serendipitous moment of brilliance, invented complex numbers or scissors or the steam engine. These things were progressions made on already existing ideas; the whole 'if I have seen farther it is by standing on the shoulders of giants' idea. I'll give you examples from your examples: Complex numbers were technically discovered over a decade before they were formally discovered. There are writings that exist where mathematicians derive results for polynomials that are the real and complex roots and they just discard the complex roots because they don't know what the hell to do with them. Complex numbers also relied on previous mathematical formulas about cubic equations that allowed for complex numbers to be derived, these were worked on by entirely separate individuals. Scissors are hopefully a fairly obvious one in that you don't believe that scissors are predecessors to the knife and the idea of a lever or hinge? The earliest evidence for the innovation of the steam engine is from someone born in a year that has two digits in it, you can look up an aeolipile. Steam engines were over 1600 years old as an innovation before they found a practical use thanks to metallurgy improvements that allowed for you to actually build steam pressure inside a container without it smashing a hole through the side. Technological and academic developments virtually never happen 'all of the sudden'. There's a reason why you can find Newton inventing/discovering calculus independently and not releasing anything and then Leibniz inventing/discovering calculus and publishing the first papers on it some ~15 years later as a simple example. I legitimately think you're way too caught up in setting ridiculous requirements (ala GUT or physical stimuli interfacing) for what is, at the end of the day, a computer program. It's probably the most impressive computer program ever built (virtually all of the frontier models meet this category), but if you expect it to either resolve the entirety of the universe or bake you a cake... I just don't know if that's a reasonable place to set the bar. | ||
|
Cyro
United Kingdom20344 Posts
On August 01 2026 20:41 Slydie wrote: Afaik, AI could only learn to play SC2 after it had been fed 1000s of human replays as a starting point. When left alone at the game, it could barely produce a unit. This was a task limited to the digital world. Afaik, llms do not understand what it types, so they can not understand anything about the real world either. They can fake pretty well, though, as their data sets are large. The next step I am curious about is what happens when the actual cost of running AI needs to be paid by someone. The circular investments between big-tech companies will dry up eventually. Even if AI does real work, there are real costs too, and using humans might often be cheaper. Especially when you factor in having to add taxes onto the companies using AI to pay for the no-longer-employed humans to not starve to death (and thus, not vote in somebody who will make the tech illegal or riot and burn down the datacenters). The needs of the humans in society don't actually change with the existence of new technology to automate their jobs, so they have to get "paid" either way. That new technology has to be so productive that it can pay for itself and those human wages, otherwise it's cheaper for society to just employ the human. That's the minimum bar for viability. Tools have achieved this in history, predominantly things like tractors and trucks which replaced horses and multiplied effective human manpower to an enormous degree, but it's not an easy bar to reach. LLM's generally haven't reached that bar and probably won't for most usages IMO. | ||
|
Uldridge
Belgium5218 Posts
On August 01 2026 22:14 BradTheBaneling wrote:+ Show Spoiler + On August 01 2026 17:43 Uldridge wrote: Well the heart of the argument is a particular non-rigoruous one I think: what exactly is innovation? Would LLMs be able to "invent" scissors to cut paper when presented with the problem that paper needed to be cut, or would it come up with a generic compounded answer of things already existing to hamfist its answer into a thing that does the job (but not really)? Or would it be able to "innovate" the use of steam as a way to move vast amounts of work and come up with the concept of a steam engine? The only thing being asked here is to extrapolate this to our current technological era and ask if it can come up with concepts no other human has come up with so that a new step in the tech tree is unlocked or contribute to our understanding of the universe. I've never claimed anything about the mathematics of LLMs, I'm sure they're quite intricate and complex systems, which I can't be bothered at the moment to understand the architecture of. Innovation and derivation are difficult to pin down, but I hope you more or less understand what I'm trying to say. You're hyper focused on physical stimuli for a reason that I can't really understand. Would an AI meerkat spend as much time watching for danger as a non-AI meerkat? I just don't know what it's supposed to ask; maybe an AI would invent scissors, the steam engine and spend an equal amount of time watching for danger. I don't understand how any of these could be delineated as a more or less useful question to ask. I know we're on a forum for video games but I'm sure you're cognizant of the fact that human technological progress is not a "tech-tree". It's important because you'll make a statement like "if it can come up with concepts no other human has come up with" and it seems that your base test is: AI must literally invent new physics that describes the entire universe or go back in time and invent scissors. These things were not just invented from a person locked in a cave who, in a serendipitous moment of brilliance, invented complex numbers or scissors or the steam engine. These things were progressions made on already existing ideas; the whole 'if I have seen farther it is by standing on the shoulders of giants' idea. I'll give you examples from your examples: Complex numbers were technically discovered over a decade before they were formally discovered. There are writings that exist where mathematicians derive results for polynomials that are the real and complex roots and they just discard the complex roots because they don't know what the hell to do with them. Complex numbers also relied on previous mathematical formulas about cubic equations that allowed for complex numbers to be derived, these were worked on by entirely separate individuals. Scissors are hopefully a fairly obvious one in that you don't believe that scissors are predecessors to the knife and the idea of a lever or hinge? The earliest evidence for the innovation of the steam engine is from someone born in a year that has two digits in it, you can look up an aeolipile. Steam engines were over 1600 years old as an innovation before they found a practical use thanks to metallurgy improvements that allowed for you to actually build steam pressure inside a container without it smashing a hole through the side. Technological and academic developments virtually never happen 'all of the sudden'. There's a reason why you can find Newton inventing/discovering calculus independently and not releasing anything and then Leibniz inventing/discovering calculus and publishing the first papers on it some ~15 years later as a simple example. I legitimately think you're way too caught up in setting ridiculous requirements (ala GUT or physical stimuli interfacing) for what is, at the end of the day, a computer program. It's probably the most impressive computer program ever built (virtually all of the frontier models meet this category), but if you expect it to either resolve the entirety of the universe or bake you a cake... I just don't know if that's a reasonable place to set the bar. Likewise I don't understand why you think I'm hyperfixated on physical stimuli? They were just examples.. and since when are complex numbers physical stimuli? In any case, they had to find a solution for being able to solve these equations. Maybe I'm putting too much weight on able to do these kinds of thing. Even if they are collaborative works, or built upon previous works or if they were discovered independently. You seem to think all these things are just self evident. Ultimately, I don't think my point is coming across quite well. The iterative process of tranforming the aeolipile to current day turbines is not what I'm talking about, although it's surely part of the process. One still needs to be able to have the idea of wanting to test the idea of putting a metal container around pressurized steam. Whatever. Maybe I'm again putting too much emphasis on the creative 'divine ideation' process of humans and deem them too intranslatable to machines (for now). Apparently I've done a piss poor job of making this argument. I know innovation doesn't need to be something ultra fancy like GUT or whatever. Perhaps I need to think a bit more about what it is I actually want to say about it before vomiting out a word salad. | ||
|
BradTheBaneling
44 Posts
On August 02 2026 00:21 Uldridge wrote: Likewise I don't understand why you think I'm hyperfixated on physical stimuli? They were just examples.. and since when are complex numbers physical stimuli? In any case, they had to find a solution for being able to solve these equations. Maybe I'm putting too much weight on able to do these kinds of thing. Even if they are collaborative works, or built upon previous works or if they were discovered independently. You seem to think all these things are just self evident. Ultimately, I don't think my point is coming across quite well. The iterative process of tranforming the aeolipile to current day turbines is not what I'm talking about, although it's surely part of the process. One still needs to be able to have the idea of wanting to test the idea of putting a metal container around pressurized steam. Whatever. Maybe I'm again putting too much emphasis on the creative 'divine ideation' process of humans and deem them too intranslatable to machines (for now). Apparently I've done a piss poor job of making this argument. I know innovation doesn't need to be something ultra fancy like GUT or whatever. Perhaps I need to think a bit more about what it is I actually want to say about it before vomiting out a word salad. I mean the complex numbers example was weird because you seem to feel that mathematical academic work doesn't count. A guy working at Anthropic used one of their models to disprove the Jacobian conjecture, which isn't complex numbers level of contribution I suppose, although only Euler gets to have his name associated with the Euler identity and the identity itself was always true given the axioms, it just hadn't been explicitly proven. I don't think they are inherently self-evident, although I think that at the point at which they were 'discovered' they were evident to those who looked and if not demonstrated by the person that demonstrated them it would've been demonstrated by someone else. Furthermore I think the simplest argument that demonstrates that I'm probably barking up the right tree here is that anatomically modern humans have existed for at least ~200,000 years, maybe closer to 300,000, and yet up until extremely recently we more or less knew virtually nothing about the objective reality that was occurring around us for the last quarter of a million years. What changed? Well in ~100 CE China invented paper, and then in ~1000 CE China invented movable type. Would you look at what a brief 1000 years of collaboration has managed? More than the last 200,000-300,000 years combined. Human understanding and technological process is a comically slow 3-4 mm/y climb that didn't really begin in earnest until we invented a writing system, it's that slow even for the smartest people on Earth. It's that if you have thousands or hundreds of thousands of people able to work on something like mathematics and applied mathematics (physics/stats/accounting etc...) you can start to move hundreds of metres per year. We aren't swinging through the trees, we're lying belly down on a cliff-side trying (generally) to inch ourselves ever upward. I mean the idea of a metal container and steam probably happened the very first time people cooked anything with water in anything metal. So you know, cooked food sort of gave us that and then someone just made a little improvement on it by containing the steam a bit more and a bit more. I also think the 'divine ideation' is a weird thing because it suggests that you believe that there is a rational and logical reason as to why humanity has specifically found these things and described them the way that we have, but also that there is a rational and logical reason as to why there is nothing in the entire +90 billion light year wide universe that would be able to conceive of a steam engine or complex numbers or scissors. | ||
| ||
