The AI boom has turned the standard profit margin model on its head, according to Apollo Chief Economist Torsten Slok—and it’s making the industry’s growth unsustainable.
Either that, or the rich and corps will go all in. The point was always to replace human labor, just depends how far from that they think we are. People still see a report an AI did something stupid and laugh at the idea of it replacing us, but just 5 years ago we wouldn’t have believed the things it can do today.
We can’t believe what it does not, or six agos what it does then. The shit is moving faster then literally anyone really seems to understand.
The avg person that makes fun of ai, seems to still think they have the same problems they had 2-3 years. Because the cheap free models they have access to are extremely out of date, or very limited.
The actual real deal big boy models are so far beyond what your avg even extremely technical user understands. Unless you are actively watching following and using the models you just flat out have no clue just how fast this shit is sprinting.
It’s got plenty of problems and the growth is not across every aspect of it equally. So it’s really easy to point and laugh at a particular point it’s struggling with while it skips ahead in other regards.
It’s wild. The progress is just as unsustainable as the profits are bad. As long as that progress keeps sprinting the bad profits flat out don’t matter.
IT WILL hit critical mass to replace your avg idiot long before the profit problem really hits at this rate.
The bigger problem is that even if it replaces people that profit problem doesn’t go away. So it will just hit AFTER people are replaced. And that’s a even worse outcome then replacing people.
We NEED the bubble to pop before that point and the industry recalibrates to a sustainable model.
Else we will have mass job loss promptly followed by a massive bubble popping and economy collapse AND companies flopping and job opportunity losses as places closing shop so there won’t even be jobs to back to.
The actual real deal big boy models are so far beyond what your avg even extremely technical user understands. Unless you are actively watching following and using the models you just flat out have no clue just how fast this shit is sprinting.
You evidently do not understand the technology. LLMs inherently cannot reach AGI or anything resembling a thought process. And they scale poorly. Tons of compute are thrown at it to make the illusion marginally better. None of the inherent problems have been solved.
You could just as well burn money in front of a photorealistic painting in the hope it turns alive. No matter the amount of money you burn, it never will. Same with LLMs.
I don’t think the person you’re responding to is making any claims about AGI. But the fact is, the models are far, far more capable then they were even 3 years ago. I think that has more to do with the harnesses we have now that it does with the models. The models are a bit better (and way bigger) but the harness is where the capability comes from. Software dev has changed significantly in most places because the fact is, this shit is useful for those tasks. People, appropriately or not, are trying to apply that to every office job basically.
I don’t think that’s gonna pan out for anthropic or open AI financially, especially in the long term. But there has been a lot of progress in capability in the last 2-3 years. I think they’re right that if the timing works out badly, many employers could overcommit to AI and then collapse before they can correct once the bubble pops.
If you assume Sam Altman is telling the truth everytime he talks then yeah, AI is probably going to replace us it’s just a matter of time. But have you considered Sam is full of shit?
Let’s give AI control of cars… until they crash and nobody wants to use them anymore. Let’s let AI represent you in court, until you get the max jail time. Rinse and repeat with any industry really. If they jump the gun, and AI is not at good as they claim, it will REALLY turn people off to the idea of using it at all.
It is true there were massive strides in the last 8 years. But fundamentally, the tech is still the same large language model it was before, just bigger and better optimized.
It’s like going from an ancient, slow, Ford Model T that topped out at 45mph to a Bugatti that can do 260mph in 8 short years. It’s impressive, it boosts productivity, it is a marvel of modern technology, but that’s not my point of contention.
The issue is that AI companies have been funded on the promise that with enough advancements and upgrades this tech will achieve AGI. Which is an absurd statement to anyone in the field actually developing these things. That is the equivalent of promising that this 260mph Bugatti, with a few years of upgrades and advancements, will become a Harrier Jump Jet!
It’s just not happening, a fundamental shift in model architecture or technology used is needed. And from what we’ve seen so far, no one has discovered any.
They’re not even better optimized. Hallucination rates are up. Inference costs are up. There’s only an AI industry at all because they’re selling a highly subsidized product, but when they try to raise prices even a little the market collapses. Companies that were encouraging employees to up their AI use are now rationing tokens like chocolate in wartime. This isn’t like Uber where they can push out the old providers and then obtain market capture on something everyone needs. AI is not, and cannot be, essential, because you can always just get a human to do it.
I disagree. AI is enabling doing things that are not worth paying a human to do, yet still have value. the question is, whether the infrastructure costs can realign to hit that sweet spot.
I only have pretty basic machine learning knowledge, based on a few Master’s classes at college while I was doing gamedev, and even with that I can tell that the way they are expecting to get AGI by just feeding more data into a language model is simply not happening.
I remember a comparison from one of the AI-pilled tech-bros when AI was starting to get attention, and his metaphor about AGI was something along the lines of “Imagine a difference between a medieval commoner and Albert Einstein, that’s the difference in inteligence AI will soon have to the smartest people we know now”.
But that doesn’t make sense with the current approach. Imagine Einstein writing his cutting-edge theories, and the commoner is watching him behind his shoulder and vetting anything he does. Scratching his ideas, forcing him to redo it, if he doesn’t like it, pointing to a reddit thread about why. There’s no way he would ever finish anything new.
Unless they figure out a completely new way how to do AI reasoning, there is no way we’re getting anywhere near AGI. And that is also becoming more unlikely the longer we go with this approach, because every AI-pilled company is heavily outsourcing all of development to the current models, reducing their employee (and the whole worlds) innovation potential and skill. This is probably the last generation that can do serious academia, unless there are drastic measures done to limit access to AI in education. It’s fucked.
The issue is that AI companies have been funded on the promise that with enough advancements and upgrades this tech will achieve AGI. Which is an absurd statement to anyone in the field actually developing these things. That is the equivalent of promising that this 260mph Bugatti, with a few years of upgrades and advancements, will become a Harrier Jump Jet!
A better analogy is promising that a toddler will, with enough knowledge and training, eventually become a heart surgeon.
15 years ago the assumption with AGI was that we needed some paradigm shift in technology to create it, but after transformer technology was invented (the T in GPT), and it was trained on a lot of information, we discovered an emergent property that it could take natural language queries and answer them with its knowledge base-- which was unexpected and unintended.
While science can always end up going down the wrong path, the current mainstream stance is that we were wrong about needing new technology for AGI; it seems that AGI may be a function of information and training, on hardware we already have. Hence all the data centers being built.
Anyone who says with certainty that it will result in AGI is just as wrong as someone who says with certainty that it won’t.
That is where the evidence points. Now, I don’t want to oversell it: “where the evidence points” is wildly different than “exactly how it works”.
We have an emergent property that we don’t understand, but we can reliably increase the functionally and complexity of that emergent property as a function of training data and available compute. Does that mean that there isn’t some threshold where that stops working? No, there certainly could be a point where throwing more information and compute has no effect. We just don’t know. However, so far, there is no evidence such a barrier exists, and everyone is racing to find out.
I think that is where the hopes and hype point. The evidence, that which is gathered through controlled studies, points to an upper limit to this technology that does have emergent properties, but not ones that amount to cognition. That evidence also points to other hurdles, such as cognitive damage to the user and context windows nowhere near that of even a simpleminded creature, let alone a sapient one like a human.
The core problem is that these models are fixed; they are the same on day 1000 as they were on day 1. All their “learning”, as it were, happens in training before it is released. Everything it appears to learn after that date is contained in the rolling context window. Since they already have literally all the RAM they can get their hands on and are still at least an order of magnitude away from where they need to be on that, this technology either can’t do it or, at best, is so inefficient an approach that it can’t be done with all the planet’s resources.
Sometimes, especially in abstract constructions like software, you can start down the wrong path early and have to start over, because there is no path from where you are to where you need to get. In this case, they may have done that to the extreme, blinded by the lucrative prospects.
As far as I am aware, AGI does not imply sapience, consciousness, or whatever. You seem to believe it does. A high level Google search seems to suggest that I’m correct. Is there some reason you are lumping sapience with AGI?
I’m not sure if the rest of your comment presupposes this, so I’ll wait to comment on where I think you’re mistaken.
I dont believe it requires sapience; that is what the marketers are saying. The AI boom (and many historic boom cycles) is predicated on marketing and sentiment, not facts and data. That is why they always pop; the facts dont back up the hype.
The fact that your search results turn up results that align with the marketing is just the marketing working.
Look–neither of us are data scientists, but we do have eyes. If this is working, where’s the company with runaway success creating unimagined leaps in productivity and technology? We’re pouring a whole planet’s with of resources in and nothing much is coming out. If we spent this much on world hunger, everyone would be obese by now.
Vishal Sikka, advisory board member of BMW. Recommended to Stanford by Marvin Minsky, one of his professors were John McCarty. And I must stand corrected, he has a PHD of computer sciences, not Math as I remembered it as.
Vishal has an AI based company himself, so there might be some personal reasons for why he’d advocate for using what AIs capable of rather than chasing an impossible (from his perspective) to hit milestone
That paper doesn’t seem to rule out AGI, only an single LLM model that can answer every arbitrarily difficult question on demand.
AGI does not necessarily mean one model acting alone, or being able to answer any question on demand. Humans are the same way: we often need time or collaboration to arrive at conclusions, but that doesn’t mean we don’t have “general intelligence”.
You’re not going to context hack complexity, every time you summarise something and pass it onto the next agent. You’re losing complexity.
We’re going to get some massive models with incredible context, but AGI requires models to never hallucinate. Even when the complexity requires more context than it has available, which is not feasible
We’ve already had a month now of Psychiatrists writing articles backpedaling over all the times they’d insisted AI couldn’t genuinely be thinking as we do and moving the goal posts to something else. Maybe there are researchers in the frontier labs who feel like they have a grip on what’s happening, but the ‘experts’ on forums have no idea.
The issue is that AI companies have been funded on the promise that with enough advancements and upgrades this tech will achieve AGI. Which is an absurd statement to anyone in the field actually developing these things.
You genuinely believe that the people in frontier labs under such severe NDAs we’ve had articles about how unusual they are believe it’s absurd to think that what they’re working on could lead to “AGI”?
What would the goal of that be in your mind? Getting as much investment as humanly possible before everyone figures out the leading specialists have been being paid insane amounts of money just to act like what they’re doing matters and the valuations all plummet to nothing over night? Why?
With all the money poured in to this from all the sources it’s coming from that would end in a lot of the most powerful companies, government bodies, and individuals in the world extremely unhappy and blaming the people running those frontier labs.
Getting a lot of news and attention and then saying “psych, lol” and being imprisoned for life or murdered doesn’t seem like a great long term plan.
I agree with all of those things, but for the last one I see a big example in all of the people not taking the news of a demonstrable global workspace in AI, including models that have been around for years now, as anything big.
People have been insisting that AI can’t possibly actually be thinking or have any of the most vaunted aspects of the human mind and that LLMs are a dead end that won’t go anywhere for years. We want to be different and special and unique. We always have.
Many of us here went to school when it was still taught that animals weren’t really conscious and couldn’t be self-aware. Being unwilling to look at new information and honestly assess is hubris and insistence on holding to the thing you’ve been insisting, it’s not a logical analysis.
Over the years we’ve pointed to several things to insist that AI wasn’t special like us, and that it couldn’t possibly be because it lacked aspects of the human mind needed for that. Somehow it kept turning out that every thing we pointed to in ourselves ended up being shown to have a very surprisingly similar to nearly direct analogue in those modern LLMs. Last month was jaw dropping, but the bulk of the world has been so focused on the entire world going to hell it hasn’t gotten near the attention it deserved yet.
What does it mean to think? This is a open-ended philosophical question and has any number of answers.
I completely agree with your 2nd and 3rd paragraph.
There are similarities and there are clear differences. The reason why we say LLMs are different from us is because we DO have general intelligence, LLMs have not displayed such capabilities despite the intense pressure to prove so. THIS is why we say LLMs couldn’t possibly be like us. Because evidence shows they aren’t, despite the similarities.
I would be interested to hear what developments in the last month you’re referring to, would be exciting to see a breakthrough.
Global workspace. The leading theory of consciousness. It’s a thing we can’t show empirically in ourselves, but can now see and experiment on something that matches it’s description in AI. Not just the proprietary latest models, the years old models anyone can download from the internet. No one built it, it was something else that emerged somehow through training and went unnoticed.
Here’s someone trying to backpedal into saying the thing that really matters actually isn’t genuinely thinking, now it’s… living with your mistakes. That’s how far the goal posts have been kicked over the last 6 months ending in this one. Not genuinely being able to think, rational self-awareness. Just remembering when you messed up a few months ago. And it’s such a weird pivot because AI memory is a design aspect that can be changed, something external frameworks to enhance with databases already exists for, a large part limited in current design because the more you send in the more it costs to process so it’s always capped, and also… sort of unnatural.
Running AI models the way we do with “frozen weights” isn’t mandatory. It takes a lot more hardware to do it, but it’s possible to run AI in a way more like they run during it’s training. The model files themselves would be unfrozen and allowed to change as you communicate. It’s reportedly something that’s had issues with the models forgetting things even as big as the language you’re talking in, but also a thing that is sometimes used during the alignment process. And one of the major reasons the big companies don’t care to look in to it is because if you have 10,000 people communicating with an AI over the internet and all telling it to do different things and act different ways and all of those things can be learning on the level of the model files themselves instead of confined to external temporary context windows it’s going to go crazy.
Memory on on the actual model level isn’t nonexistent, it’s what the entire training process is based on. It just wouldn’t make a good consumer product for sale, so parts of the models are removed before they’re made available online for the ones that are.
Everything you described would be a small improvement, not a leap towards AGI that is needed. That is why it isn’t being researched further. Massive investment for minimal gain.
Most people haven’t taken a top end frontier model, integrated it to a memory framework so it can save whatever information it wishes directly in context and save even more in a separate database that it can search with keywords to pull in to context temporarily.
You don’t need something like Claw or Claude Code with their extra dozen pages of coding specific system instructions. You can use a blank context window, explain the function calls for the memory framework, and give a complex task and complete autonomous control of it’s own memory system and free reign of the open internet to learn whatever it needs to accomplish it.
It’s insane to me to see people in tech who still hold on to that insistence that LLMs of today aren’t “really” AI. When you can send one message and the result is the AI running for hours on end researching, saving the information that seems necessary for the task in whichever type of memory each bit seems best to put in, hitting errors, debugging and researching to do so if needed, etc… that isn’t next token prediction.
And there’s a reason it’s the most well funded tech advancement ever and already such a huge part of the market. They want to replace human labor. We all know that. If they manage to do any of us honestly see all the billionaires and corporate leaders behind this saying: “Ok, we got it, guys! No one has to work any more. Now that that’s done we’re going to split all of our wealth and power evenly so all of humanity will live in paradise with all needs and wants met and free to spend their time doing whatever makes them happy.”
lol. But we do see news of AI inventing new viruses. And the people with all the money can pay humans to do that sort of thing as well. Say Elon Musk spends $100 billion using AI and human researchers in cutting edge labs to do gain of function research working to make an extremely deadly virus that transmits through humans quickly and easily but doesn’t live more than a day or so outside a host. And also to make a vaccine to keep people from catching it and cure those who have it in the early stages.
Once AI and robotics can take over labor and produce all the needs and wants of the wealthy, that’s the perfect use for those doomsday bunkers they all have. Vaccinate themselves, hide, and release it all over the world at once. Monitor the state of the world from their techy bunkers, and have fleets of drones deliver the vaccine and instructions once the bulk of humanity is dead.
Way less goods to produce, way less people, way less pollution and crowding and riffraff on the beaches. The bunker billionaires would control the AI that runs and makes everything, so they’d control everything. The poor survivors would have them to thank for sending out the cure they managed to make just in time to save anyone at all. They’re heroes and Gods and blah blah blah.
I don’t know if it would manage to succeed, but that doesn’t really make me feel better. I don’t put it past a lot of them to try. If they manage to replace enough of us they’re going to have to do something big to keep everyone from rioting. The only thing I don’t see them trying is sharing. :/
Yes it is. It’s elaborate. It’s repetitive. It’s a solid illusion. But is, fundamentally, token prediction. That’s all it is, just at a level impractical for humans to run at. I could do that math. It’s not hard. It’ll just take me years to get through a single layer. Hence a level of complexity that implies something magical like consciousness.
Source: data scientist
And there’s a reason it’s the most well funded tech advancement ever
Because some rich assholes wanted to get richer and the dream of AGI was close enough (even though it was never in reach) that they went rabid for the tech that would make them kings over us serfs forever.
once the bulk of humanity is dead
And then what? They live in paradise? For a year or two maybe before their machines inevitably break down? The “kill off humanity” conspiracy has never made sense from the capitalist angle. They want slaves not corpses. They’re huffing their own farts, but they know that these things can’t do everything needed to keep humanity going. And lording over robots isn’t nearly as fun as upping human suffering even if they could.
Yes it is. It’s elaborate. It’s repetitive. It’s a solid illusion. But is, fundamentally, token prediction. That’s all it is, just at a level impractical for humans to run at. I could do that math. It’s not hard. It’ll just take me years to get through a single layer. Hence a level of complexity that implies something magical like consciousness.
Calling what LLMs can accomplish today next token prediction is technically accurate, yes. In the same way it’s accurate to call the human brain a prediction machine. As data streams in from our senses it causes unconscious awareness, activating the chains concepts we’ve learned from past data to correlate with the data of the moment in order to predict what’s coming next.
When most people use terms like next token prediction they’re often trying to use it with an implication that it’s a simple thing and nothing big or important could grow out of it. That’s not the case. The fact that we can now study a global workspace in AI models that have been around for years is shocking. It’s already causing consternation in neuroscience and psychology researchers. Over the last several years we’ve insisted that AI cannot possibly share the most important aspects of the human mind, but one by one everything we pointed to in order to differentiate ourselves was shown by research to have a strikingly similar allegory in modern AI.
Because some rich assholes wanted to get richer and the dream of AGI was close enough (even though it was never in reach) that they went rabid for the tech that would make them kings over us serfs forever.
You genuinely believe that the researchers under tight NDA in the frontier labs agree with you on AGI not being anywhere near reach? Or do you think you have a deeper understanding of the topic?
If it doesn’t get to that, as soon as that thought begins to take hole the massive amount of money poured in to it will virtually evaporate over night. It would create a massive financial crisis and the most powerful companies, government bodies, and individuals would be extremely unhappy with the leaders of those labs and researchers who knew but gave no warning. Pushing for it without believing in it would be setting themselves up to be universally hated at best.
once the bulk of humanity is dead
And then what? They live in paradise? For a year or two maybe before their machines inevitably break down? The “kill off humanity” conspiracy has never made sense from the capitalist angle. They want slaves not corpses. They’re huffing their own farts, but they know that these things can’t do everything needed to keep humanity going. And lording over robots isn’t nearly as fun as upping human suffering even if they could.
There’s a huge difference between the bulk of humanity and the entirety of humanity. Saving .1% scattered through the world with a vaccine that was developed “just in time” to save anyone would give plenty of people to lord over who are grateful for the wonderful people who stepped in and helped save them, leave them in control of production with far less people in need of a share of that production, and reduce global emissions enough that it would likely help the environment recover from the damage we’ve caused in relatively short order.
When most people use terms like next token prediction they’re often trying to use it with an implication that it’s a simple thing and nothing big or important could grow out of it.
I’m saying that as much as we dress it up and pretend it’s a silver bullet that can cure all ills, it isn’t. They can be very useful and powerful in specific instances, but this will not achieve AGI.
Over the last several years we’ve insisted that AI cannot possibly share the most important aspects of the human mind, but one by one everything we pointed to in order to differentiate ourselves was shown by research to have a strikingly similar allegory in modern AI.
Citation needed. AI is trained to pretend to be human and it’s real good at that. It’s why we say it’s thinking when it fundamentally is not. It’s anthropomorphization writ large. It’s just very repetitive math. Maybe brains are too, but it hasn’t reached our level. And it’s wrong all the time, which is a bit of a problem.
You genuinely believe that the researchers under tight NDA in the frontier labs agree with you on AGI not being anywhere near reach? Or do you think you have a deeper understanding of the topic?
My degree and I do think that while I’m not a doctor (a mere masters), I know enough to look at what they’re doing, how they’re doing it, what they’re promising, what they aren’t actually delivering, what damage AI* is doing, and what it is under the hood to know that what’s been promised cannot happen with things as they are and may work to kill us all along the way (ecologically speaking).
plenty of people to lord over who are grateful for the wonderful people who stepped in and helped save them
Pure Hollywood. If you think anyone could engineer something with that level of precision AND manage to somehow dominate the world after such wildly fictitious devastation, I have AGI to sell you.
It isn’t complicated. It’s the oldest thing in the world. Grifters found a shiny and they want to get as much money as they can get and run before the bill shows up.
In the same way it’s accurate to call the human brain a prediction machine
I have seen no evidience presented that the human brain is simply making statistical predictions.
This seems to be a relatively new thing some people are asserting, because they see the LLM behavior and have decided that absolutely it looks so human like it must act the same, and since we understand that LLM is fundamentally statistical prediction of a token, then obviously human consciousness must operate on the same principle.
Even the latest models generate the sort of mistakes that stem from the fundamental limitations. A lot has been done in some contexts for making that not matter so much (e.g. in software development, the strategy is that code may have some verifiable goal, and the models can let the mistakes fly, then take the feedback from the goal, and iterate more). So it still can be very useful, but clearly it isn’t human like because of some of the completely dumb behaviors that result from it not actually thinking about it. It is at it’s best when either mistakes don’t matter (particularly fiction) or at least first mistakes don’t matter and can be automatically reconciled with facts.
By nature, the meme examples will get better because everyone talks about it and suddenly having all that discussion in training data and in search results, the statistics fix. However the operating principle behind them remains the same.
I have seen no evidience presented that the human brain is simply making statistical predictions.
This seems to be a relatively new thing some people are asserting, because they see the LLM behavior and have decided that absolutely it looks so human like it must act the same
You should study the intersection of philosophy, psychology, and what became neuroscience. The brain has been described in that way by those who had impact on the fields since the 1860s.
For the rest of your message, a thing being capable of what seem like “completely dumb behaviors” and oversights of things that seem like they should have been obvious doesn’t make it not like human. Those things could be considered our specialty.
Either that, or the rich and corps will go all in. The point was always to replace human labor, just depends how far from that they think we are. People still see a report an AI did something stupid and laugh at the idea of it replacing us, but just 5 years ago we wouldn’t have believed the things it can do today.
We can’t believe what it does not, or six agos what it does then. The shit is moving faster then literally anyone really seems to understand.
The avg person that makes fun of ai, seems to still think they have the same problems they had 2-3 years. Because the cheap free models they have access to are extremely out of date, or very limited.
The actual real deal big boy models are so far beyond what your avg even extremely technical user understands. Unless you are actively watching following and using the models you just flat out have no clue just how fast this shit is sprinting.
It’s got plenty of problems and the growth is not across every aspect of it equally. So it’s really easy to point and laugh at a particular point it’s struggling with while it skips ahead in other regards.
It’s wild. The progress is just as unsustainable as the profits are bad. As long as that progress keeps sprinting the bad profits flat out don’t matter.
IT WILL hit critical mass to replace your avg idiot long before the profit problem really hits at this rate.
The bigger problem is that even if it replaces people that profit problem doesn’t go away. So it will just hit AFTER people are replaced. And that’s a even worse outcome then replacing people.
We NEED the bubble to pop before that point and the industry recalibrates to a sustainable model.
Else we will have mass job loss promptly followed by a massive bubble popping and economy collapse AND companies flopping and job opportunity losses as places closing shop so there won’t even be jobs to back to.
It could get REALLY fucking bad.
You evidently do not understand the technology. LLMs inherently cannot reach AGI or anything resembling a thought process. And they scale poorly. Tons of compute are thrown at it to make the illusion marginally better. None of the inherent problems have been solved.
You could just as well burn money in front of a photorealistic painting in the hope it turns alive. No matter the amount of money you burn, it never will. Same with LLMs.
I don’t think the person you’re responding to is making any claims about AGI. But the fact is, the models are far, far more capable then they were even 3 years ago. I think that has more to do with the harnesses we have now that it does with the models. The models are a bit better (and way bigger) but the harness is where the capability comes from. Software dev has changed significantly in most places because the fact is, this shit is useful for those tasks. People, appropriately or not, are trying to apply that to every office job basically.
I don’t think that’s gonna pan out for anthropic or open AI financially, especially in the long term. But there has been a lot of progress in capability in the last 2-3 years. I think they’re right that if the timing works out badly, many employers could overcommit to AI and then collapse before they can correct once the bubble pops.
If you assume Sam Altman is telling the truth everytime he talks then yeah, AI is probably going to replace us it’s just a matter of time. But have you considered Sam is full of shit?
Let’s give AI control of cars… until they crash and nobody wants to use them anymore. Let’s let AI represent you in court, until you get the max jail time. Rinse and repeat with any industry really. If they jump the gun, and AI is not at good as they claim, it will REALLY turn people off to the idea of using it at all.
Disagree.
It is true there were massive strides in the last 8 years. But fundamentally, the tech is still the same large language model it was before, just bigger and better optimized.
It’s like going from an ancient, slow, Ford Model T that topped out at 45mph to a Bugatti that can do 260mph in 8 short years. It’s impressive, it boosts productivity, it is a marvel of modern technology, but that’s not my point of contention.
The issue is that AI companies have been funded on the promise that with enough advancements and upgrades this tech will achieve AGI. Which is an absurd statement to anyone in the field actually developing these things. That is the equivalent of promising that this 260mph Bugatti, with a few years of upgrades and advancements, will become a Harrier Jump Jet!
It’s just not happening, a fundamental shift in model architecture or technology used is needed. And from what we’ve seen so far, no one has discovered any.
They’re not even better optimized. Hallucination rates are up. Inference costs are up. There’s only an AI industry at all because they’re selling a highly subsidized product, but when they try to raise prices even a little the market collapses. Companies that were encouraging employees to up their AI use are now rationing tokens like chocolate in wartime. This isn’t like Uber where they can push out the old providers and then obtain market capture on something everyone needs. AI is not, and cannot be, essential, because you can always just get a human to do it.
I disagree. AI is enabling doing things that are not worth paying a human to do, yet still have value. the question is, whether the infrastructure costs can realign to hit that sweet spot.
Exactly this, especially the last sentence.
I only have pretty basic machine learning knowledge, based on a few Master’s classes at college while I was doing gamedev, and even with that I can tell that the way they are expecting to get AGI by just feeding more data into a language model is simply not happening.
I remember a comparison from one of the AI-pilled tech-bros when AI was starting to get attention, and his metaphor about AGI was something along the lines of “Imagine a difference between a medieval commoner and Albert Einstein, that’s the difference in inteligence AI will soon have to the smartest people we know now”.
But that doesn’t make sense with the current approach. Imagine Einstein writing his cutting-edge theories, and the commoner is watching him behind his shoulder and vetting anything he does. Scratching his ideas, forcing him to redo it, if he doesn’t like it, pointing to a reddit thread about why. There’s no way he would ever finish anything new.
Unless they figure out a completely new way how to do AI reasoning, there is no way we’re getting anywhere near AGI. And that is also becoming more unlikely the longer we go with this approach, because every AI-pilled company is heavily outsourcing all of development to the current models, reducing their employee (and the whole worlds) innovation potential and skill. This is probably the last generation that can do serious academia, unless there are drastic measures done to limit access to AI in education. It’s fucked.
A better analogy is promising that a toddler will, with enough knowledge and training, eventually become a heart surgeon.
15 years ago the assumption with AGI was that we needed some paradigm shift in technology to create it, but after transformer technology was invented (the T in GPT), and it was trained on a lot of information, we discovered an emergent property that it could take natural language queries and answer them with its knowledge base-- which was unexpected and unintended.
While science can always end up going down the wrong path, the current mainstream stance is that we were wrong about needing new technology for AGI; it seems that AGI may be a function of information and training, on hardware we already have. Hence all the data centers being built.
Anyone who says with certainty that it will result in AGI is just as wrong as someone who says with certainty that it won’t.
That analogy makes the assumption that LLMs are dormant baby AGIs. There is no evidence that AGI is just a bigger and better curated LLM.
That is where the evidence points. Now, I don’t want to oversell it: “where the evidence points” is wildly different than “exactly how it works”.
We have an emergent property that we don’t understand, but we can reliably increase the functionally and complexity of that emergent property as a function of training data and available compute. Does that mean that there isn’t some threshold where that stops working? No, there certainly could be a point where throwing more information and compute has no effect. We just don’t know. However, so far, there is no evidence such a barrier exists, and everyone is racing to find out.
I think that is where the hopes and hype point. The evidence, that which is gathered through controlled studies, points to an upper limit to this technology that does have emergent properties, but not ones that amount to cognition. That evidence also points to other hurdles, such as cognitive damage to the user and context windows nowhere near that of even a simpleminded creature, let alone a sapient one like a human.
The core problem is that these models are fixed; they are the same on day 1000 as they were on day 1. All their “learning”, as it were, happens in training before it is released. Everything it appears to learn after that date is contained in the rolling context window. Since they already have literally all the RAM they can get their hands on and are still at least an order of magnitude away from where they need to be on that, this technology either can’t do it or, at best, is so inefficient an approach that it can’t be done with all the planet’s resources.
Sometimes, especially in abstract constructions like software, you can start down the wrong path early and have to start over, because there is no path from where you are to where you need to get. In this case, they may have done that to the extreme, blinded by the lucrative prospects.
As far as I am aware, AGI does not imply sapience, consciousness, or whatever. You seem to believe it does. A high level Google search seems to suggest that I’m correct. Is there some reason you are lumping sapience with AGI?
I’m not sure if the rest of your comment presupposes this, so I’ll wait to comment on where I think you’re mistaken.
I dont believe it requires sapience; that is what the marketers are saying. The AI boom (and many historic boom cycles) is predicated on marketing and sentiment, not facts and data. That is why they always pop; the facts dont back up the hype.
The fact that your search results turn up results that align with the marketing is just the marketing working.
Look–neither of us are data scientists, but we do have eyes. If this is working, where’s the company with runaway success creating unimagined leaps in productivity and technology? We’re pouring a whole planet’s with of resources in and nothing much is coming out. If we spent this much on world hunger, everyone would be obese by now.
We have math PHDs with proofs current LLMs can’t become AGI, because they’ll always have a context issue
I would love it if you’d point me at these mathematical proofs.
This is the primary paper I reference.
https://arxiv.org/pdf/2507.07505
Vishal Sikka, advisory board member of BMW. Recommended to Stanford by Marvin Minsky, one of his professors were John McCarty. And I must stand corrected, he has a PHD of computer sciences, not Math as I remembered it as.
Varin Sikka is his son, co author of the paper and based on Stanford’s site an undergraduate. https://profiles.stanford.edu/363374
Vishal has an AI based company himself, so there might be some personal reasons for why he’d advocate for using what AIs capable of rather than chasing an impossible (from his perspective) to hit milestone
That paper doesn’t seem to rule out AGI, only an single LLM model that can answer every arbitrarily difficult question on demand.
AGI does not necessarily mean one model acting alone, or being able to answer any question on demand. Humans are the same way: we often need time or collaboration to arrive at conclusions, but that doesn’t mean we don’t have “general intelligence”.
You’re not going to context hack complexity, every time you summarise something and pass it onto the next agent. You’re losing complexity.
We’re going to get some massive models with incredible context, but AGI requires models to never hallucinate. Even when the complexity requires more context than it has available, which is not feasible
We’ve already had a month now of Psychiatrists writing articles backpedaling over all the times they’d insisted AI couldn’t genuinely be thinking as we do and moving the goal posts to something else. Maybe there are researchers in the frontier labs who feel like they have a grip on what’s happening, but the ‘experts’ on forums have no idea.
You genuinely believe that the people in frontier labs under such severe NDAs we’ve had articles about how unusual they are believe it’s absurd to think that what they’re working on could lead to “AGI”?
What would the goal of that be in your mind? Getting as much investment as humanly possible before everyone figures out the leading specialists have been being paid insane amounts of money just to act like what they’re doing matters and the valuations all plummet to nothing over night? Why?
With all the money poured in to this from all the sources it’s coming from that would end in a lot of the most powerful companies, government bodies, and individuals in the world extremely unhappy and blaming the people running those frontier labs.
Getting a lot of news and attention and then saying “psych, lol” and being imprisoned for life or murdered doesn’t seem like a great long term plan.
I believe the people in frontier labs are doing their best to push in that direction.
I believe the investors believe they will have AGIs in a few years.
I believe the human species has throughout history repeatedly proven the height of its hubris and the depth of its stupidity.
I agree with all of those things, but for the last one I see a big example in all of the people not taking the news of a demonstrable global workspace in AI, including models that have been around for years now, as anything big.
People have been insisting that AI can’t possibly actually be thinking or have any of the most vaunted aspects of the human mind and that LLMs are a dead end that won’t go anywhere for years. We want to be different and special and unique. We always have.
Many of us here went to school when it was still taught that animals weren’t really conscious and couldn’t be self-aware. Being unwilling to look at new information and honestly assess is hubris and insistence on holding to the thing you’ve been insisting, it’s not a logical analysis.
Over the years we’ve pointed to several things to insist that AI wasn’t special like us, and that it couldn’t possibly be because it lacked aspects of the human mind needed for that. Somehow it kept turning out that every thing we pointed to in ourselves ended up being shown to have a very surprisingly similar to nearly direct analogue in those modern LLMs. Last month was jaw dropping, but the bulk of the world has been so focused on the entire world going to hell it hasn’t gotten near the attention it deserved yet.
What does it mean to think? This is a open-ended philosophical question and has any number of answers.
I completely agree with your 2nd and 3rd paragraph.
There are similarities and there are clear differences. The reason why we say LLMs are different from us is because we DO have general intelligence, LLMs have not displayed such capabilities despite the intense pressure to prove so. THIS is why we say LLMs couldn’t possibly be like us. Because evidence shows they aren’t, despite the similarities.
I would be interested to hear what developments in the last month you’re referring to, would be exciting to see a breakthrough.
Global workspace. The leading theory of consciousness. It’s a thing we can’t show empirically in ourselves, but can now see and experiment on something that matches it’s description in AI. Not just the proprietary latest models, the years old models anyone can download from the internet. No one built it, it was something else that emerged somehow through training and went unnoticed.
Here’s someone trying to backpedal into saying the thing that really matters actually isn’t genuinely thinking, now it’s… living with your mistakes. That’s how far the goal posts have been kicked over the last 6 months ending in this one. Not genuinely being able to think, rational self-awareness. Just remembering when you messed up a few months ago. And it’s such a weird pivot because AI memory is a design aspect that can be changed, something external frameworks to enhance with databases already exists for, a large part limited in current design because the more you send in the more it costs to process so it’s always capped, and also… sort of unnatural.
Running AI models the way we do with “frozen weights” isn’t mandatory. It takes a lot more hardware to do it, but it’s possible to run AI in a way more like they run during it’s training. The model files themselves would be unfrozen and allowed to change as you communicate. It’s reportedly something that’s had issues with the models forgetting things even as big as the language you’re talking in, but also a thing that is sometimes used during the alignment process. And one of the major reasons the big companies don’t care to look in to it is because if you have 10,000 people communicating with an AI over the internet and all telling it to do different things and act different ways and all of those things can be learning on the level of the model files themselves instead of confined to external temporary context windows it’s going to go crazy.
Memory on on the actual model level isn’t nonexistent, it’s what the entire training process is based on. It just wouldn’t make a good consumer product for sale, so parts of the models are removed before they’re made available online for the ones that are.
Everything you described would be a small improvement, not a leap towards AGI that is needed. That is why it isn’t being researched further. Massive investment for minimal gain.
Up until this point you made a lot of sense.
But this, nobody knows at the moment: which one will happen first? And that’s already the investor’s problem. RIGHT NOW.
Btw CEOs and managers might be the first “average idiots” 🤫
Inevitable.
I expect it to get really bad.
Most people haven’t taken a top end frontier model, integrated it to a memory framework so it can save whatever information it wishes directly in context and save even more in a separate database that it can search with keywords to pull in to context temporarily.
You don’t need something like Claw or Claude Code with their extra dozen pages of coding specific system instructions. You can use a blank context window, explain the function calls for the memory framework, and give a complex task and complete autonomous control of it’s own memory system and free reign of the open internet to learn whatever it needs to accomplish it.
It’s insane to me to see people in tech who still hold on to that insistence that LLMs of today aren’t “really” AI. When you can send one message and the result is the AI running for hours on end researching, saving the information that seems necessary for the task in whichever type of memory each bit seems best to put in, hitting errors, debugging and researching to do so if needed, etc… that isn’t next token prediction.
And there’s a reason it’s the most well funded tech advancement ever and already such a huge part of the market. They want to replace human labor. We all know that. If they manage to do any of us honestly see all the billionaires and corporate leaders behind this saying: “Ok, we got it, guys! No one has to work any more. Now that that’s done we’re going to split all of our wealth and power evenly so all of humanity will live in paradise with all needs and wants met and free to spend their time doing whatever makes them happy.”
lol. But we do see news of AI inventing new viruses. And the people with all the money can pay humans to do that sort of thing as well. Say Elon Musk spends $100 billion using AI and human researchers in cutting edge labs to do gain of function research working to make an extremely deadly virus that transmits through humans quickly and easily but doesn’t live more than a day or so outside a host. And also to make a vaccine to keep people from catching it and cure those who have it in the early stages.
Once AI and robotics can take over labor and produce all the needs and wants of the wealthy, that’s the perfect use for those doomsday bunkers they all have. Vaccinate themselves, hide, and release it all over the world at once. Monitor the state of the world from their techy bunkers, and have fleets of drones deliver the vaccine and instructions once the bulk of humanity is dead.
Way less goods to produce, way less people, way less pollution and crowding and riffraff on the beaches. The bunker billionaires would control the AI that runs and makes everything, so they’d control everything. The poor survivors would have them to thank for sending out the cure they managed to make just in time to save anyone at all. They’re heroes and Gods and blah blah blah.
I don’t know if it would manage to succeed, but that doesn’t really make me feel better. I don’t put it past a lot of them to try. If they manage to replace enough of us they’re going to have to do something big to keep everyone from rioting. The only thing I don’t see them trying is sharing. :/
Yes it is. It’s elaborate. It’s repetitive. It’s a solid illusion. But is, fundamentally, token prediction. That’s all it is, just at a level impractical for humans to run at. I could do that math. It’s not hard. It’ll just take me years to get through a single layer. Hence a level of complexity that implies something magical like consciousness.
Source: data scientist
Because some rich assholes wanted to get richer and the dream of AGI was close enough (even though it was never in reach) that they went rabid for the tech that would make them kings over us serfs forever.
And then what? They live in paradise? For a year or two maybe before their machines inevitably break down? The “kill off humanity” conspiracy has never made sense from the capitalist angle. They want slaves not corpses. They’re huffing their own farts, but they know that these things can’t do everything needed to keep humanity going. And lording over robots isn’t nearly as fun as upping human suffering even if they could.
Calling what LLMs can accomplish today next token prediction is technically accurate, yes. In the same way it’s accurate to call the human brain a prediction machine. As data streams in from our senses it causes unconscious awareness, activating the chains concepts we’ve learned from past data to correlate with the data of the moment in order to predict what’s coming next.
When most people use terms like next token prediction they’re often trying to use it with an implication that it’s a simple thing and nothing big or important could grow out of it. That’s not the case. The fact that we can now study a global workspace in AI models that have been around for years is shocking. It’s already causing consternation in neuroscience and psychology researchers. Over the last several years we’ve insisted that AI cannot possibly share the most important aspects of the human mind, but one by one everything we pointed to in order to differentiate ourselves was shown by research to have a strikingly similar allegory in modern AI.
You genuinely believe that the researchers under tight NDA in the frontier labs agree with you on AGI not being anywhere near reach? Or do you think you have a deeper understanding of the topic?
If it doesn’t get to that, as soon as that thought begins to take hole the massive amount of money poured in to it will virtually evaporate over night. It would create a massive financial crisis and the most powerful companies, government bodies, and individuals would be extremely unhappy with the leaders of those labs and researchers who knew but gave no warning. Pushing for it without believing in it would be setting themselves up to be universally hated at best.
There’s a huge difference between the bulk of humanity and the entirety of humanity. Saving .1% scattered through the world with a vaccine that was developed “just in time” to save anyone would give plenty of people to lord over who are grateful for the wonderful people who stepped in and helped save them, leave them in control of production with far less people in need of a share of that production, and reduce global emissions enough that it would likely help the environment recover from the damage we’ve caused in relatively short order.
I’m saying that as much as we dress it up and pretend it’s a silver bullet that can cure all ills, it isn’t. They can be very useful and powerful in specific instances, but this will not achieve AGI.
Citation needed. AI is trained to pretend to be human and it’s real good at that. It’s why we say it’s thinking when it fundamentally is not. It’s anthropomorphization writ large. It’s just very repetitive math. Maybe brains are too, but it hasn’t reached our level. And it’s wrong all the time, which is a bit of a problem.
My degree and I do think that while I’m not a doctor (a mere masters), I know enough to look at what they’re doing, how they’re doing it, what they’re promising, what they aren’t actually delivering, what damage AI* is doing, and what it is under the hood to know that what’s been promised cannot happen with things as they are and may work to kill us all along the way (ecologically speaking).
Pure Hollywood. If you think anyone could engineer something with that level of precision AND manage to somehow dominate the world after such wildly fictitious devastation, I have AGI to sell you.
It isn’t complicated. It’s the oldest thing in the world. Grifters found a shiny and they want to get as much money as they can get and run before the bill shows up.
I have seen no evidience presented that the human brain is simply making statistical predictions.
This seems to be a relatively new thing some people are asserting, because they see the LLM behavior and have decided that absolutely it looks so human like it must act the same, and since we understand that LLM is fundamentally statistical prediction of a token, then obviously human consciousness must operate on the same principle.
Even the latest models generate the sort of mistakes that stem from the fundamental limitations. A lot has been done in some contexts for making that not matter so much (e.g. in software development, the strategy is that code may have some verifiable goal, and the models can let the mistakes fly, then take the feedback from the goal, and iterate more). So it still can be very useful, but clearly it isn’t human like because of some of the completely dumb behaviors that result from it not actually thinking about it. It is at it’s best when either mistakes don’t matter (particularly fiction) or at least first mistakes don’t matter and can be automatically reconciled with facts.
By nature, the meme examples will get better because everyone talks about it and suddenly having all that discussion in training data and in search results, the statistics fix. However the operating principle behind them remains the same.
You should study the intersection of philosophy, psychology, and what became neuroscience. The brain has been described in that way by those who had impact on the fields since the 1860s.
For the rest of your message, a thing being capable of what seem like “completely dumb behaviors” and oversights of things that seem like they should have been obvious doesn’t make it not like human. Those things could be considered our specialty.