Exactly. He has a dishonest and predictable schtick. We need skeptics for any industry, but he is just grifting in an ecological niche as a knee-jerk reactionary. There is near-zero value to anything he writes or says in that role
Yet nothing he wrote in that article was wrong. Schtick or not, if you read the article, it's all true. OpenAI has no moat, they have no killer app; this was a huge PR blunder for them.
A business that burns money at the rate OpenAI does, without any clear path to profitability, will eventually die.
This is a fundamental misunderstanding of what a moat is in this industry.
If GPT 4.5 had nothing of value to offer other than a flex of opening eyes ability to scale, it’s still a signal that they have the chops to throw the most amount of compute at the upcoming reinforcement learning race.
If you’re actually selling enterprise solutions in this space, you’ll quickly learn that a large number of Enterprises have their own private deployment of open AI on Azure and are pushing their vendors to use that even if it means they get lower quality outputs for certain use cases.
Data and model quality aren’t the only moat-able features.
This is a clear signal that the people remaining to run the show at OpenAI may not actually know what they're doing. The big names building the technology have fled from Altman's power grab. They all have startups of their own.
Meanwhile, the people at Anthropic keep making Claude better and better. No bombast, just tech. And then there's the model not on the list of many of these shoot-outs: Grok 3. That one has scale, and it's actually good.
"We can sorta kinda also make a big AI!" is not the message you want to be sending. It has to clobber the other "main" models out there.
I was going to reply to this snarky comment with a non-snarky reply but I changed my mind, so I will instead laugh riotously at the idea that aviation worked great continuously from its inception to today.
What year do you place “2.5 years after mainstream public buy-in” for aviation? I think you’ll find that in most periods, people with the means were paying a lot for the ability to “get there faster”, despite a bad early safety record. It did work great when it didn’t kill you, and despite that risk it was a hit (for those that could afford it).
I don't know if Kyle is right but "any good after mainstream public buy-in" and "worked great continuously from inception to today" is quite the goalpost move.
It's been less than 6 months since OpenAI last released a SOTA model architecture that promptly caused every other AI lab to scramble to copy them (o1 preview). Maybe OpenAI is running ashore, but it's too soon to tell yet.
I wonder if we reached the limit of what a single-shot large models can do? Or we're just waiting for another breakthrough? If it's the limit, then probably the chain of thought is the answer. I think models are so good now that given the right context, tools and time, they can pretty much do anything.
I would have thought that is obvious. Chain of thought is clearly the right path forwards. People don't stare at a thing and immediately come to a solution. Without chain of thought, you are mostly dealing with a statistical regurgitation machine. Effective for some tasks yes, but clearly not enough to meet the bar of general intelligence.
I think you are wrong that this is the limit for single shot though. We have reached a limit due to data, but I expect what will happen next is that we will transition into a slow growth phase where chain of thought models will be used to effectively create more training data, which will then be used to train a better single shot model, which will then be extended into a better chain of thought model, which will then produce higher quality training data for another single shot. And so on. Kind of like what happens as knowledge passes from teacher to student across successive generations. Effectively a continuing process of compression and growth in intelligence, but progressing rather slowly compared to what we have seen in the last 5 years.
Because AI LLMs are actively ruining the education of children, the actual retaining of information coming from writing your own words and thinking your own thoughts. Because AI is only being used as a cudgel by executives to reduce the workforce of humans with no intention of additional economic justice.
I wouldn’t give up on teachers just yet. The 5-paragraph essay is probably dead as are mathematics homework assignments that are boring variations of the same problem over-and-over. The field is being cleared for new ideas and I bet some of them will prove to be very good.
How in the world did you get from "1:1 student-teacher ratio" to "ruin kids education"? AI is amazing for education. Everyone has their own teacher in their pocket.
The exact same thing is said for every new form of tech. All tech has its good parts and bad parts. It was true of the black and white television, it was true of the Nintendo, it was true of the cell phone and it will be true if this technology as well.
You either accept that change happens and use your life experience to help shape it in a positive direction or, well… I dunno. Become a old curmudgeon and watch the world blow by you.
We don't allow people to use mustard gas willy nilly as their technology of choice. But sure, your "there's good and bad to everything!!! change is inevitable!!" platitudes are useful too.
And your tech examples don't offload actual process of thinking to others the way AI does. The comparisons are surface level and ignore what I actually said.
Except you didn't offer any evidence for your position, provide any references, or provide any sort of logical argument. You offer your own platitude and then condemn other commenters for doing so, hypocrisy.
The current state of LLMs _can_ be helpful for education. Millions of people use them as such and benefit from it.
A far bigger problem with the technology IMO is the generative aspect. We already have a large problem with disinformation and spam on the internet, and generative AI will increase this by many orders of magnitude. Discerning fact from fiction is already difficult today; it will be literally impossible to do in the future, _unless_ we invent more technology to save us from it. This is a problem we haven't even begun to address. The public is collectively blinded by the novelty of the technology, while entrepreneurs are jumping over themselves trying to profit from the latest gold rush. Very few people with the power to change anything are actually thinking about the long-term, or even mid-term, impacts.
I have school aged children and it’s not AI that’s ruining anything
Curriculum isn’t moving fast enough. Just a few years ago every teacher had to adapt for COVID and go 100% online/remote in many areas. Kids are still turning in every assignment online, even in classroom settings in my district.
So, yeah, kids can just paste the question into ChatGPT and copy the answer. Nobody learns anything.
This isn’t AI ruining education, it’s schools being under-resourced and unable to move as quickly as society is changing.
Teachers are still buying their own supplies, how can they adapt their entire curriculum in the course of a couple years to work under this entirely new paradigm of LLMs?
Give it a bit and I am convinced schools will go back to oral reports, handwritten essays and whatever is needed to make sure children are not just pasting garbage back and forth
Honestly, I think this is what we need to kill toxic social media and phone addiction as well. If AI forces us to talk and interact as a community again, it’s a win. Leave the internet to the bots and AI.
Because this needs to be toned down a serious notch. I’ve spent the last year and a half in AI land just to do 95% data work and pretend that it’s now somehow magic AI while OpenAI and the rest are seen as the magic that makes this all work.
Well put. They are useful building blocks but ultimately a lot of the magic is in the data modeling and data engineering and shaping happening behind the scenes. And because it's still slow and costly it's hard to do that well ... and the democratizing frameworks for doing it well haven't been born yet
Bingo - we can focus on the energy on actually modernizing industries and leveraging the (pretty good) AGI we have rather than racing with each other to incinerate more and more money for little to no gains.
I was talking with someone in a non-tech industry and we have such a long way to go for even decades-old information system improvements. They don't even have basic things like an effective system of record or analytics. They have no way to measure the success of any particular initiative. If revenue is down - they don't really know why, they just randomly change stuff until it hopefully goes back up.
Because regardless of what we're talking about, a bubble around a thing is still a lie. The faster one bursts, the faster truth lays bare and one can make an actually informed decision. LLMs are here to stay and have probably already found a growing place in our lives. But much energy is currently spent speculating about their future significance. The bubble is about downplaying those are just speculations, while inflating the perception of their importance in our current or upcoming reality.
Well, for a start, the longer it takes to burst, the worse the burst is going to be. Right now, the fallout if it bursts would be fairly limited. Give it another few years of hype, and it may be structurally dangerous when it does go.
Also, it’s inefficient allocation of capital. Every cent being spent on this is money that could be spent on something useful (of course, absent the AI bubble, not _all_ of it would be, but some of it would be).
The pressure to hit quarterly targets seems to, in this case, caused quite the opposite outcome intended. Gotta love corporate America and the overvalued AI sector. (:
These are powerful building blocks, they're just not the be all end all. The building blocks up stack that use these as part of a broader architecture, that shape these and many other traditional techniques in software... this is where the real value will be unlocked.
This layer itself will inevitably see its cost come down per unit of use on a long road towards commoditization. It will probably get better and better and more sophisticated but again the value will be primarily up stack, not accrued primarily from a company like this. It's not to say they couldn't be a great company... even Google is a great company that has enabled countless other companies to bloom. The myopic way people look to these one size fit all companies is just so disconnected from our economy works.
It took us nearly 70 years to get to this architecture and the processing power to support it.
If we are waiting for a new breakthrough architecture, it could be decades. Our brains send signals with very high concurrency to individual processors (neurons). Each neuron is massively more complex than a single ReLU function and we have billions of them. If we need that kind of parallel processing and data transfer to match human thought and flexibility, it could be another 70 years before we see AGI.
That said, I do think LLMs are one of the biggest AI breakthroughs since the inception of AI. And I am sure that it, or something very similar will be part of an eventual AGI.
> These are powerful building blocks, they're just not the be all end all. The building blocks up stack that use these as part of a broader architecture, that shape these and many other traditional techniques in software... this is where the real value will be unlocked.
I could not agree more. There is much value still to be gained from blockchain technology!
While OpenAI is certainly in trouble (zero Hinton students, partnership with Azure is straining, while Brin is back at Google!), its demise is still uncertain. They are still absolutely amazing lab, pushing both product and research. Execute. And make all these naysayers obsolete ;)
My biggest hope is that home computing catches up to the point where you can buy a sub-thousand-dollar card, slap it into your desktop, and run this stuff locally. Take the power away from massive, well-funded corporations with armies of lawyers and “safety weenies” and put it back in the hands of makers and tinkerers.
Right now, these models are built by the establishment to serve the establishment—and that’s a load of shit that needs to change. The real fun starts the day some random group of “anonymous” on 4chan can train one of these models to generate incredibly convincing deepfakes of world leaders, all using the compute power in their own homes.
Power to the people. Fuck the system, and all that jazz.
Open AI has a brand, it has talent, it has some really solid models, and it has the compute to serve inference at scale. They stand as good a chance of covering their inference costs as anyone else.
The compute for training is beginning to seem a poor investment since it is depreciating fast and isn't producing value in this case. That's a seriously big investment to make if it's not productive but since a lot of it actually belongs to Azure they could cut back here fast if they had to. I hope they won't because in the hands of good researchers there is still a real possibility that they'll use the compute to find some kind of technical innovation to give them a bigger edge.
To serve their models on Azure and into phones. Co-Pilot is a joke. At least Google have rolled Gemini into Docs, Sheets, Android Studio, Gmail among other things.
Which is harder, to build a working AI, or to build a phone client? If OpenAI can do the former, it won't matter that Google is better at doing the latter.
It absolutely blows my mind that people swallowed "AGI is just around the corner" hook, line and sinker. We don't understand human cognition despite literally thousands of years of thought and study, but some nerds with a thousand GPUs are going to make a machine think? Ridiculous. People were misled by graphs that they were told pointed to a singularity right around the corner, despite obvious errors in the extant systems.
What's even more laughable is that OpenAI dumbed down the definition of AGI so much it doesn't align at all with general intelligence anymore, and people just accepted it.
Not only can they not reach AGI, they cannot reach their own definition of AGI. But people will still gobble whatever next lie Altman will sell them for $2,000 a month.
> Artificial general intelligence (AGI) is the ability of an intelligent agent to understand or learn __any__ intellectual task that a human being can.
LLMs are nowhere near this. Not even close.
But let's get even more restrictive by looking at the actual required characteristics to declare a system an AGI:
> There is wide agreement among artificial intelligence researchers that intelligence is required to do the following:
> - reason, use strategy, solve puzzles, and make judgments under uncertainty;
> - represent knowledge, including common sense knowledge;
> - plan;
> - learn;
> - communicate in natural language;
> - and integrate all these skills towards common goals.
> Other important capabilities include:
> - input as the ability to sense (e.g. see, hear, etc.), and
> - output as the ability to act (e.g. move and manipulate objects, change own location to explore, etc.)
Once again, LLMs are not even scratching the surface of AGI.
Of course if you look at the current version of the page, where the bar has been massively lowered, it might look like AGI is achieved, but that's because the definition has changed, not the technology.
> This is not AGI.
> Let's take an old version of the Wikipedia page about AGI, from end of 2022
sure, its not AGI by old wikipedia definition, which imo is human-centric and more definition of superintelligence (requires ability to exceed all existing humans in all tasks). But it is AGI by current wikipedia definition.
That is literally my point... The definition of AGI has been twisted so much by the ones that claim to be on the verge of achieving it that it is not anymore in line with what AGI has always meant.
Reading wikipedia about history of the term, it sounds like term was popularized by this book: https://www.amazon.com/Artificial-General-Intelligence-Cogni... which says that general means "ability so solve variety of tasks in variety of domains", not "all tasks in all domains". So "always meant" is easily challengeable here.
Anyone paying attention _DIDN'T_ buy it. It's only the AI-hype-bros who seriously considered AGI to be a real possibility within this decade, let alone THIS YEAR like a ton of people said (people who all suspiciously had a lot of financial gain to be had from believing that to be true.)
> We don't understand human cognition despite literally thousands of years of thought and study, but some nerds with a thousand GPUs are going to make a machine think? Ridiculous.
DNA doesn't understand intelligence.
"The question of whether a computer can think is no more interesting than the question of whether a submarine can swim." - Edsger Dijkstra.
And what we have now, with LLMs, met the standard I had for AGI just five years ago.
Only thing that changed for me since then is noticing that not only does everyone have a different definition of what AGI even is, also none of the three initials are even boolean-valued: "how intelligent" can be a number, or even a vector that varies by domain as linguistic intelligence doesn't have to match spatial intelligence; "how general" could be what percentage of human cognitive tasks the AI can do; "how artificial" is answered entirely differently by those who care differently about learning from first principles vs. programmed algorithms vs. learning from reading the internet.
> People were misled by graphs that they were told pointed to a singularity right around the corner, despite obvious errors in the extant systems.
Tautology.
If there weren't obvious errors in the extant systems… there wouldn't be anything left to do.
Agreed. Dogs & 4 year olds & LLM's all have general intelligence. Defining "AGI" to mean "superhuman intelligence" seems ridiculous to me. The Turing test was the standard benchmark for so long, but now that an LLM can pass the Turing test we've moved the goal posts.
The Turing test was never a standard benchmark. It was common in pop science coverage, but no AI researcher (including Turing himself!) ever proposed it as a useful test of intelligence. It was an illustration by Turing that we can meaningfully talk about a machine thinking in the first place. We've known for over half a century since ELIZA that the ability to carry on a plausible natural language conversation is much more narrow than it intuitively seems.
What study are you referring to here? The ones I've seen don't particularly resemble anything Turing described. The preprint from UCSD got heavy coverage in the popular press but its headline claim is an elementary statistical mistake, modifying the test so it's no longer a binary comparison but still treating 50% as a meaningful pass threshold. It hasn't yet passed peer review nine months later, and I'm cautiously optimistic that it never will.
I agree in my view the modern language models already processes general intelligence: they are better at some tasks and worse than others compared to human, but in principle they can tackle any problem (if given agency like tool access)
this is probably a religious question so people will not be convinced of intelligent machines even if they have an EQ and IQ of 200 because they don't work the same way human cells do I suppose.
In fairness, GPT-3.5 and GPT-4 were much closer to AGI than anything before. Many people also believed AGI was around the corner in the old LISP days, when there was much less capital.
I still believe GPT-4 has some general intelligence, just a very tiny amount. It can take what it was trained on and slightly modify it to answer a slightly novel question.
It seems to me that some nerds with a thousand GPUs have clearly made a machine think. State-of-the-art AI models are better than me at a significant number of intellectual tasks; what would it mean, precisely, to say that I'm thinking when I write a bad poem but GPT-4o is not thinking when it writes a better one?
The specific term "AGI" has always referred to a program that strictly matches to human intelligence, with no significant areas where it's worse than a typical human. I agree that we're not there and not obviously close to there. But the idea that it's all a parlor trick and LLMs have no cognition at all seems obviously false to me.
It was not that long ago that solving a maze was thought of as AI. Or a bot in a video game adapting to your play style was AI.
Today, LLMs being trained to output text, images and video equal to or better than any human is now what we consider to be AI.
There is a semantic game we play to move the goalposts at every new tick of these technologies so that humans are still on top.
Maybe the next big trick is to define AGI as having a system that was not trained on the entire corpus of human output available on the internet, and train it with nothing. Can an AI be “born” with no model like a baby and learn to speak, walk, function in society without being primed for success with a state of the art model? Could you drop this AI into any human society, regardless of culture, and have it seamlessly integrate?
What precisely we consider “cognition” feels like a philosophical debate, one where we have no single true answer. And doesn’t really matter?
"AGI" has always referred to something a lot closer to your "next big trick" than an LLM. An AGI is a program that can slot into anywhere a human fits and do as well as a human would; GPT-4 or Claude can't do that.
I do think it's true that anyone a decade ago would have predicted their capabilities are impossible without AGI, which suggests that the entire conceptual framework is less useful than we thought. The eventual debate about whether we've "reached AGI" will probably have a lot more to do with the contract between OpenAI and Microsoft than any real paradigm shift in capabilities.
I guess I'm not sure what your point is. If you understand "thinking" to be a highly advanced capability that millions or billions of people can't do, sure, perhaps it's true that GPT-4 is not "thinking" in that sense. Most people understand the term to be more broad than that. If a four year old tries really hard to compute 4 + 16, I'd definitely say they're "thinking" about it, even if they don't get the right answer or make some silly mistake I can easily point out.
I haven't tried 4.5 but it is being touted as having greater EQ - presumably meaning emotion quotient. That sounds like progress, or at least something that's worth releasing to find out what people can do with it.
Having said that, 4.5 is clearly a misstep, one that should realign the goals of the AI industry to focus more on utility and cost effectiveness.
Isn't that just fine-tuning? LLMs can understand emotions just fine, but ChatGPT was always neutered and fine-tuned to act as a "helpful assistant" and claim it was "just a language model" and not say anything controversial.
Unless you know what cognition is, it's not inherently ridiculous that a bunch of nerds with a thousand GPUs are going to make a machine think. What's ridiculous is that, when we don't understand human cognition, a bunch of nerds are sure they're going to make a machine think, and how close they are to doing it.
> We don't understand human cognition despite literally thousands of years of thought and study
Is this a requirement for achieving AGI? The history of progression of the ML field indicates that the answer is "no". We don't really understand how concepts are encoded in today's models, yet that doesn't stop them from being economically useful. So why would the special case of AGI be any different?
I mean, according to many of the same people about a decade ago, the world economy should be running on bitcoin by now. And we should all be living in the ‘metaverse’. Like, it’s largely the same people who fall for every fad falling for this.
You don't think LLMs are thinking? You don't think they are approaching human levels of intelligence? What part of human cognition don't we understand?
fdsjgfklsfd says>You don't think LLMs are thinking? You don't think they are approaching human levels of intelligence? What part of human cognition don't we understand?
Ans. No, no and the entirety of human cognition.
ChatGPT et al contain collections of words. When prompted they generate word sequences found in known word sources. That's all. They don't observe or reason about the world. They don't even reason about words. They merely append the most likely next word to a sequence of words.
The laymen's definition for AGI is utterly irrelevant to investors. If training needs to happen for an AI employee so there's an employee that can be copy/pasted and scaled up/down without overhead, so then put in the work to train it. You have to train a human employee as well anyway.
I don’t think I’d be this pessimistic. There’s still o3 that hasn’t been seen in its full form, and whatever improvements are made to it now that multi head latent attention is out in the wild.
Orion does seem to have been a failure, but I also find it a bit weird that they seemingly decided to release the full model rather than a distillation, which is the pattern we now usually see with foundation models.
So, did they simply decide that it wasn’t worth the effort and dedicate the compute to other, better things? Were they pushed by sama to release it anyway, to look like they were still making progress while developing something really next gen?
OpenAI releases one dud and suddenly people come out of the woodwork to trash them.
I agree that OpenAI's endless hyping about AGI seems pretty unrealistic, but let's take a breather here. There are no major research projects where you don't run into setbacks and failures before you reach your goals.
And even if they never really progress beyond where they are today with their current models, just bringing down the cost could open up a lot of doors for useful applications.
Yep, GPT 4.5 is not a breakthrough, but there will be more breakthroughs.
Someone will figure out how to make AIs understand their own ignorance and stop bullshitting when they don't know something, someone will figure out how to make AIs learn on the fly instead of fine-tuning new model versions, etc.
I used to be a skeptic. Chat GPT is awesome. Similar to Google, it's become a staple of my everyday life, like writing, or composing latex, or even doing math. I cannot imagine not having it without losing productivity. Is the business model sustainable? Probably not; I already see some signs of throttling in an attempt to cut costs, like if you overuse computation for math. Just the ability to render the latex from wolfram alpha is great. Writing latex used to be such a chore.
You can get pretty much all the benefits, if you know how to apply a little bit of elbow grease, from open source models at this point so even if the business model doesnt work, this is a technology that we all have now which is great.
Humans learn. LLM context windows are vastly larger than our short-term memory, but vastly smaller than our long-term recollection. LLMs can recall vastly more information than our long-term memory, but only from their static training data.
Also, even though LLMs can generate text much faster than humans, we may be internally thinking much faster. Each adult human brain has over 100 billion neurons and 100 trillion synapses, and each has been working every moment, for decades.
This is what separates human reasoning from LLM reasoning, and it can’t be solved by scaling the latter to anything feasible.
I wish AI companies would take a decent chunk of their billions, and split it into 1000+ million-dollar projects that each try a different idea to overcome these issues, and others (like emotion and alignment). Many of these projects would certainly fail, but some may produce breakthroughs. Meanwhile, spending the entire billion on scaling compute has failed and will continue to fail, because everyone else does that, so the resulting model has no practical advantages and makes less money than it cost to train before it becomes obsoleted by other people’s breakthroughs.
Gary Marcus. By all accounts, he doesn't understand how LLMs work, so usually he's wrong about technical matters.[a]
But here, I think he's right about business matters. The massive investment in computing capacity we've seen in recent years, by Open AI and others, can generate positive returns only if the technology continues to improve rapidly so it can overcome its limitations and failure modes in the short run.
If the rate of improvement has slowed down, even temporarily, OpenAI and others like Anthropic are likely to face financial difficulties.
This is going to sound disrespectful, but nobody cares. Both bloggers and CEOs will continue to argue they have the bestest AI. Our goal should simply be making sure AI is helping more than it harms. Much like nuclear research 75 years ago, we've setup this century where AI will be simultaneously massively weaponized and massively assistive technology. There's also an ultimate goal of having AGI solve all our problems like fusion was supposed to do. Let me know when your AI CEO is ready to discuss how all this scaling is helping the kids sleep better instead of making quarterly profits rise faster than the competition.
There are folks that don't like anything from Gary Marcus, because he's been a nay-sayer from the beginning. This article happens to be spot on though.
"Gary Marcus has been warning the field for years that pure statistical approaches like LLMs probably wouldn't suffice for AGI. Half a trillion dollars later, it looks like maybe he had a point."
112 comments
[ 3.7 ms ] story [ 176 ms ] threadA business that burns money at the rate OpenAI does, without any clear path to profitability, will eventually die.
If GPT 4.5 had nothing of value to offer other than a flex of opening eyes ability to scale, it’s still a signal that they have the chops to throw the most amount of compute at the upcoming reinforcement learning race.
If you’re actually selling enterprise solutions in this space, you’ll quickly learn that a large number of Enterprises have their own private deployment of open AI on Azure and are pushing their vendors to use that even if it means they get lower quality outputs for certain use cases.
Data and model quality aren’t the only moat-able features.
This is a clear signal that the people remaining to run the show at OpenAI may not actually know what they're doing. The big names building the technology have fled from Altman's power grab. They all have startups of their own.
Meanwhile, the people at Anthropic keep making Claude better and better. No bombast, just tech. And then there's the model not on the list of many of these shoot-outs: Grok 3. That one has scale, and it's actually good.
"We can sorta kinda also make a big AI!" is not the message you want to be sending. It has to clobber the other "main" models out there.
Manhattan Project: Started 1942, combat use 1945.
Solid state transistor: Invented 1954, in radios 1955.
Hybrid cars, iPhone, etc etc.
I think you are wrong that this is the limit for single shot though. We have reached a limit due to data, but I expect what will happen next is that we will transition into a slow growth phase where chain of thought models will be used to effectively create more training data, which will then be used to train a better single shot model, which will then be extended into a better chain of thought model, which will then produce higher quality training data for another single shot. And so on. Kind of like what happens as knowledge passes from teacher to student across successive generations. Effectively a continuing process of compression and growth in intelligence, but progressing rather slowly compared to what we have seen in the last 5 years.
https://en.m.wikipedia.org/wiki/Creative_destruction
You either accept that change happens and use your life experience to help shape it in a positive direction or, well… I dunno. Become a old curmudgeon and watch the world blow by you.
And your tech examples don't offload actual process of thinking to others the way AI does. The comparisons are surface level and ignore what I actually said.
The current state of LLMs _can_ be helpful for education. Millions of people use them as such and benefit from it.
A far bigger problem with the technology IMO is the generative aspect. We already have a large problem with disinformation and spam on the internet, and generative AI will increase this by many orders of magnitude. Discerning fact from fiction is already difficult today; it will be literally impossible to do in the future, _unless_ we invent more technology to save us from it. This is a problem we haven't even begun to address. The public is collectively blinded by the novelty of the technology, while entrepreneurs are jumping over themselves trying to profit from the latest gold rush. Very few people with the power to change anything are actually thinking about the long-term, or even mid-term, impacts.
Curriculum isn’t moving fast enough. Just a few years ago every teacher had to adapt for COVID and go 100% online/remote in many areas. Kids are still turning in every assignment online, even in classroom settings in my district.
So, yeah, kids can just paste the question into ChatGPT and copy the answer. Nobody learns anything.
This isn’t AI ruining education, it’s schools being under-resourced and unable to move as quickly as society is changing. Teachers are still buying their own supplies, how can they adapt their entire curriculum in the course of a couple years to work under this entirely new paradigm of LLMs?
Give it a bit and I am convinced schools will go back to oral reports, handwritten essays and whatever is needed to make sure children are not just pasting garbage back and forth
Honestly, I think this is what we need to kill toxic social media and phone addiction as well. If AI forces us to talk and interact as a community again, it’s a win. Leave the internet to the bots and AI.
I was talking with someone in a non-tech industry and we have such a long way to go for even decades-old information system improvements. They don't even have basic things like an effective system of record or analytics. They have no way to measure the success of any particular initiative. If revenue is down - they don't really know why, they just randomly change stuff until it hopefully goes back up.
That said, annoying people will move on to hyping something else.
Also, it’s inefficient allocation of capital. Every cent being spent on this is money that could be spent on something useful (of course, absent the AI bubble, not _all_ of it would be, but some of it would be).
This layer itself will inevitably see its cost come down per unit of use on a long road towards commoditization. It will probably get better and better and more sophisticated but again the value will be primarily up stack, not accrued primarily from a company like this. It's not to say they couldn't be a great company... even Google is a great company that has enabled countless other companies to bloom. The myopic way people look to these one size fit all companies is just so disconnected from our economy works.
If we are waiting for a new breakthrough architecture, it could be decades. Our brains send signals with very high concurrency to individual processors (neurons). Each neuron is massively more complex than a single ReLU function and we have billions of them. If we need that kind of parallel processing and data transfer to match human thought and flexibility, it could be another 70 years before we see AGI.
That said, I do think LLMs are one of the biggest AI breakthroughs since the inception of AI. And I am sure that it, or something very similar will be part of an eventual AGI.
I could not agree more. There is much value still to be gained from blockchain technology!
Right now, these models are built by the establishment to serve the establishment—and that’s a load of shit that needs to change. The real fun starts the day some random group of “anonymous” on 4chan can train one of these models to generate incredibly convincing deepfakes of world leaders, all using the compute power in their own homes.
Power to the people. Fuck the system, and all that jazz.
The compute for training is beginning to seem a poor investment since it is depreciating fast and isn't producing value in this case. That's a seriously big investment to make if it's not productive but since a lot of it actually belongs to Azure they could cut back here fast if they had to. I hope they won't because in the hands of good researchers there is still a real possibility that they'll use the compute to find some kind of technical innovation to give them a bigger edge.
So does Google. And Google can roll out their premium models into phones, household devices, cars and online platforms to add value.
OpenAI has a website.
It's not even close.
Concentrated capital is truly a wild thing.
Not only can they not reach AGI, they cannot reach their own definition of AGI. But people will still gobble whatever next lie Altman will sell them for $2,000 a month.
Absolutely pathetic.
that definition likely not their's but came from Microsoft when they dumped 10B into OAI and was a condition for revenue sharing clause.
One can argue that AGI is already achieved, LLMs are more proficient and general in knowledge tasks than any specific individual.
This is not AGI.
Let's take an old version of the Wikipedia page about AGI, from end of 2022, before the ChatGPT craze: https://en.wikipedia.org/w/index.php?title=Artificial_genera...
> Artificial general intelligence (AGI) is the ability of an intelligent agent to understand or learn __any__ intellectual task that a human being can.
LLMs are nowhere near this. Not even close.
But let's get even more restrictive by looking at the actual required characteristics to declare a system an AGI:
> There is wide agreement among artificial intelligence researchers that intelligence is required to do the following:
> - reason, use strategy, solve puzzles, and make judgments under uncertainty;
> - represent knowledge, including common sense knowledge;
> - plan;
> - learn;
> - communicate in natural language;
> - and integrate all these skills towards common goals.
> Other important capabilities include:
> - input as the ability to sense (e.g. see, hear, etc.), and
> - output as the ability to act (e.g. move and manipulate objects, change own location to explore, etc.)
Once again, LLMs are not even scratching the surface of AGI.
Of course if you look at the current version of the page, where the bar has been massively lowered, it might look like AGI is achieved, but that's because the definition has changed, not the technology.
sure, its not AGI by old wikipedia definition, which imo is human-centric and more definition of superintelligence (requires ability to exceed all existing humans in all tasks). But it is AGI by current wikipedia definition.
Reading wikipedia about history of the term, it sounds like term was popularized by this book: https://www.amazon.com/Artificial-General-Intelligence-Cogni... which says that general means "ability so solve variety of tasks in variety of domains", not "all tasks in all domains". So "always meant" is easily challengeable here.
DNA doesn't understand intelligence.
"The question of whether a computer can think is no more interesting than the question of whether a submarine can swim." - Edsger Dijkstra.
And what we have now, with LLMs, met the standard I had for AGI just five years ago.
Only thing that changed for me since then is noticing that not only does everyone have a different definition of what AGI even is, also none of the three initials are even boolean-valued: "how intelligent" can be a number, or even a vector that varies by domain as linguistic intelligence doesn't have to match spatial intelligence; "how general" could be what percentage of human cognitive tasks the AI can do; "how artificial" is answered entirely differently by those who care differently about learning from first principles vs. programmed algorithms vs. learning from reading the internet.
> People were misled by graphs that they were told pointed to a singularity right around the corner, despite obvious errors in the extant systems.
Tautology.
If there weren't obvious errors in the extant systems… there wouldn't be anything left to do.
What study are you referring to here? The ones I've seen don't particularly resemble anything Turing described. The preprint from UCSD got heavy coverage in the popular press but its headline claim is an elementary statistical mistake, modifying the test so it's no longer a binary comparison but still treating 50% as a meaningful pass threshold. It hasn't yet passed peer review nine months later, and I'm cautiously optimistic that it never will.
https://news.ycombinator.com/item?id=40386571
this is probably a religious question so people will not be convinced of intelligent machines even if they have an EQ and IQ of 200 because they don't work the same way human cells do I suppose.
I still believe GPT-4 has some general intelligence, just a very tiny amount. It can take what it was trained on and slightly modify it to answer a slightly novel question.
And me running is closer to the speed of light than me walking, yet neither are even remotely close to the speed of light.
The specific term "AGI" has always referred to a program that strictly matches to human intelligence, with no significant areas where it's worse than a typical human. I agree that we're not there and not obviously close to there. But the idea that it's all a parlor trick and LLMs have no cognition at all seems obviously false to me.
Today, LLMs being trained to output text, images and video equal to or better than any human is now what we consider to be AI.
There is a semantic game we play to move the goalposts at every new tick of these technologies so that humans are still on top.
Maybe the next big trick is to define AGI as having a system that was not trained on the entire corpus of human output available on the internet, and train it with nothing. Can an AI be “born” with no model like a baby and learn to speak, walk, function in society without being primed for success with a state of the art model? Could you drop this AI into any human society, regardless of culture, and have it seamlessly integrate?
What precisely we consider “cognition” feels like a philosophical debate, one where we have no single true answer. And doesn’t really matter?
I do think it's true that anyone a decade ago would have predicted their capabilities are impossible without AGI, which suggests that the entire conceptual framework is less useful than we thought. The eventual debate about whether we've "reached AGI" will probably have a lot more to do with the contract between OpenAI and Microsoft than any real paradigm shift in capabilities.
That is AI. AI ≠ AGI.
Literally millions or billions of people are better than me at a significant number of intellectual tasks
"Nerds with a thousand GPUs,
Nerds with a thousand streams,
with you only I experience,
the love, the chat of my dreams!"
"So prompt me, and list me,
index me, repeat me,
I'm yours till I elide,
so in love, so in love,
so in love with you, my GPT, am I. "
-With apologies to Cole Porter and anyone blinded by reading this post.
https://youtu.be/qdeM24FFpvM
Having said that, 4.5 is clearly a misstep, one that should realign the goals of the AI industry to focus more on utility and cost effectiveness.
Is this a requirement for achieving AGI? The history of progression of the ML field indicates that the answer is "no". We don't really understand how concepts are encoded in today's models, yet that doesn't stop them from being economically useful. So why would the special case of AGI be any different?
Ans. No, no and the entirety of human cognition.
ChatGPT et al contain collections of words. When prompted they generate word sequences found in known word sources. That's all. They don't observe or reason about the world. They don't even reason about words. They merely append the most likely next word to a sequence of words.
Orion does seem to have been a failure, but I also find it a bit weird that they seemingly decided to release the full model rather than a distillation, which is the pattern we now usually see with foundation models.
So, did they simply decide that it wasn’t worth the effort and dedicate the compute to other, better things? Were they pushed by sama to release it anyway, to look like they were still making progress while developing something really next gen?
I agree that OpenAI's endless hyping about AGI seems pretty unrealistic, but let's take a breather here. There are no major research projects where you don't run into setbacks and failures before you reach your goals.
And even if they never really progress beyond where they are today with their current models, just bringing down the cost could open up a lot of doors for useful applications.
Someone will figure out how to make AIs understand their own ignorance and stop bullshitting when they don't know something, someone will figure out how to make AIs learn on the fly instead of fine-tuning new model versions, etc.
A smart enough AI would summarize each of his posts as "I still hate the current AI boom".
There must be a term for such writers? He's certainly consistently on message.
Also, even though LLMs can generate text much faster than humans, we may be internally thinking much faster. Each adult human brain has over 100 billion neurons and 100 trillion synapses, and each has been working every moment, for decades.
This is what separates human reasoning from LLM reasoning, and it can’t be solved by scaling the latter to anything feasible.
I wish AI companies would take a decent chunk of their billions, and split it into 1000+ million-dollar projects that each try a different idea to overcome these issues, and others (like emotion and alignment). Many of these projects would certainly fail, but some may produce breakthroughs. Meanwhile, spending the entire billion on scaling compute has failed and will continue to fail, because everyone else does that, so the resulting model has no practical advantages and makes less money than it cost to train before it becomes obsoleted by other people’s breakthroughs.
But here, I think he's right about business matters. The massive investment in computing capacity we've seen in recent years, by Open AI and others, can generate positive returns only if the technology continues to improve rapidly so it can overcome its limitations and failure modes in the short run.
If the rate of improvement has slowed down, even temporarily, OpenAI and others like Anthropic are likely to face financial difficulties.
---
[a] In the words of Geoff Hinton: https://www.youtube.com/watch?v=d7ltNiRrDHQ
---
Note: At the moment, the OP is flagged. To the mods: It shouldn't be, because it conforms to the HN guidelines.
Disclosure - I am neither bearish or a mega bull on LLMs. LLMs useful in some cases.
D'oh!