The name of this CPU is bordering on securities fraud. When people see the term "AGI" now, they are assuming "Artificial General Intelligence", not "Agentic AI Infrastructure".
Of course people don't realize that, and people will buy ARM stock thinking they've cracked AGI. The people running Arm absolutely know this, so this name is what we in the industry call a "lie".
The marketers did this for 5G also, calling their product 5G before it was actually deployed, only because theirs came after 4G but wanted to ride the upcoming 5G buzz.
It seems marketing /depends/ on conflating terms and misleading consumers. Shakespeare might have gotten it wrong with his quip about lawyers.
If you showed someone what our computers can do with the latest LLMs now to someone 5 years ago they would probably say it sure looks a lot like AGI.
We have to keep defining AGI upwards or nitpick it to show that we haven't achieved it.
I would argue that LLMs are actually smarter than the majority of humans right now. LLMs do not have quite the agency that humans have, but their intelligence is pretty decent.
We don't have clear ASI yet, but we definitely are in a AGI-era.
I think we are missing an ego/motiviations in the AGI and them having self-sufficiency independent of us, but that is just a bit of engineering that would actually make them more dangerous, it isn't really a significant scientific hurdle.
Ok, but it's not AGI. People five years ago would have been wrong. People who don't have all the information are often wrong about things.
ETA:
You updated your comment, which is fine but I wanted to reply to your points.
> I would argue that LLMs are actually smarter than the majority of humans right now. LLMs do not have quite the agency that humans have, but their intelligence is pretty decent.
I would actually argue that they are decidedly not smarter than even dumb humans right now. They're useful but they are glorified text predictors. Yes, they have more individual facts memorized than the average person but that's not the same thing; Wikipedia, even before LLMs also had many more facts than the average person but you wouldn't say that Wikipedia is "smarter" than a human because that doesn't make sense.
Intelligence isn't just about memorizing facts, it's about reasoning. The recent Esolang benchmarks indicate that these LLMs are actually pretty bad at that.
> We don't have clear ASI yet, but we definitely are in a AGI-era.
> Intelligence isn't just about memorizing facts, it's about reasoning.
Initially, LLMs were basically intuitive predictors, but with chain of thought and more recently agentic experimentation, we do have reasoning in our LLMs that is quite human like.
That said, there is definitely a biased towards training set material, but that is also the case with the large majority of humans.
For the Esoland benchmarks, I would be curious how much adding a SKILLS.md file for each language would boost performance?
I am pretty confidence that we are in the AGI era. It is unsettling and I think it gives people cognitive dissonance so we want to deny it and nitpick it, etc.
> Initially, LLMs were basically intuitive predictors, but with chain of thought and more recently agentic experimentation, we do have reasoning in our LLMs that is quite human like.
If you handwave the details away, then sure it's very human like, though the reasoning models just kind of feed the dialog back to itself to get something more accurate. I use Claude code like everyone else, and it will get stuck on the strangest details that humans actively wouldn't.
> For the Esoland benchmarks, I would be curious how much adding a SKILLS.md file for each language would boost performance?
Tough to say since I haven't done it, though I suspect it wouldn't help much, since there's still basically no training data for advanced programs in these languages.
> I am pretty confidence that we are in the AGI era. It is unsettling and I think it gives people cognitive dissonance so we want to deny it and nitpick it, etc.
Even if you're right about this being the AGI era, that doesn't mean that current models are AGI, at least not yet. It feels like you're actively trying to handwave away details.
> though the reasoning models just kind of feed the dialog back to itself to get something more accurate.
Much of our reasoning is based on stimulating our sensory organs, either via imagination (self-stimulation of our visual system) or via subvocalization (self-stimulation of our auditory system), etc.
> it will get stuck on the strangest details that humans actively wouldn't.
It isn't a human. It is AGI, not HGI.
> It feels like you're actively trying to handwave away details.
Personally, I've used LLMs to debug hard-to-track code issues and AWS issues among other things.
Regardless of whether that was done via next-token prediction or not, it definitely looked like AGI, or at least very close to it.
Is it infallible? Not by a long shot. I always have to double-check everything, but at least it gave me solid starting points to figure out said issues.
It would've taken me probably weeks to find out without LLMd instead of the 1 or 2 hours it did.
In that context, I have a hard time thinking how would a "real" AGI system look like, that it's not the current one.
Not saying current LLMs are unequivocally AGI, but they are darn close for sure IMO.
Being able to actually reason about things without exabytes of training data would be one thing. Hell, even with exabytes of training data, doing actual reasoning for novel things that aren't just regurgitating things from Github would be cool.
Being able to learn new things would be another. LLMs don't learn; they're a pretrained model (it's in the name of GPT), that send in inputs and get an output. RAGs are cool but they're not really "learning", they're just eating a bit more context in order to kind of give a facsimile of learning.
Going to the extreme of what you're saying, then `grep` would be "darn close to AGI". If I couldn't grep through logs, it might have taken me years to go through and find my errors or understand a problem.
I think that they're ultimately very neat, but ultimately pretty straightforward input-output functions.
My definition of AGI hasn't changed - it's something that can perform, or learn to perform, any intellectual task that a human can.
5 years ago we thought that language is the be-all and end-all of intelligence and treated it as the most impressive thing humans do. We were wrong. We now have these models that are very good at language, but still very bad at tasks that we wrongly considered prerequisites for language.
A human can think logically with reason, not to say they are smart or smarter. But LLMs cannot. You can convince a LLM anything is correct and it will believe you. You can't convince a human anything is correct.
I can't argue that LLMs do not know an absolute insane amount of information about everything. But you can't just say LLMs are smarter then most humans. We've already decided that smartness is not about how much data you know, but thinking about that data with logical reasoning. Including the fact it may or may not be true.
I can run a LLM through absolutely incorrect data, and tell it that data is 100% true. Then ask it questions about that data and get those incorrect results as answers. That's not easy to do with humans.
> If you showed someone what our computers can do with the latest LLMs now to someone 5 years ago they would probably say it sure looks a lot like AGI.
Would they? Perhaps if you only showed them glossy demos that obscure all the ways in which LLMs fail catastrophically and are very obviously nowhere even close to AGI.
Certainly, they wouldn't expect that an AI able to score 150 on an IQ test is unable to play a casual game of chess because it isn't coherent enough to play without making illegal moves.
> Certainly, they wouldn't expect that an AI able to score 150 on an IQ test is unable to play a casual game of chess because it isn't coherent enough to play without making illegal moves.
To be fair, I am pretty sure Claude Code will download and run stockfish, if you task it to play chess with you. It's not like a human who read 100 books about chess, but never played, would be able to play well with their eyes closed, and someone whispering board position into their ear
> LLMs are actually smarter than the majority of humans right now
I consider myself a bit of a misanthrope but this makes me an optimist by comparison.
Even stupid people are waaaaaay smarter than any LLM.
The problem is the continued habit humans have of anthropomorphizing computers that spit out pretty words. It’s like Eliza only prettier. More useful for sure. Still just a computer.
I don't believe in a separation of mind and spirit. So I do think fundamentally, outside of a reliance on quantum effects in cognition (some of theorized but it isn't proven), its processes can be replicated in a fashion in computers. So I think that intelligence likely can be "just a computer" in theory and I think we are in the era where this is now true.
I don't believe in "spirits" from the get go. I think it's certainly theoretically possible that we could mimic human thought with a computer (quantum or otherwise) but I do not think that the LLMs we have now are doing that. I'd say that what we have right now is "just a computer".
This doesn't mean they aren't useful, I like Claude a lot, but I don't buy that it's AGI.
Idk about the health story, but in my use, ChatGPT has dramatically improved my understanding of my health issues and given sound and careful advice.
The second question sounds like a useless and artificial metric to judge on. The average person might miss such a “gotcha” logical quiz too, for the same reason - because they expect to be asked “is it walking distance.”
No one has ever relied on anyone else’s judgment, nor an AI, to answer “should I bring my car to the carwash.” Same for the ol’ “how many rocks shall I eat?” that people got the AI Overview tricked with.
I’m not saying anything categorically “is AGI” but by relying on jokes like this you’re lying to yourself about what’s relevant.
I would accuse you of nitpicking. My experience is that LLMs are generally as smart as the average human +90% of the time. A lack of perfect to me doesn't mean it isn't AGI.
>> My experience is that LLMs are generally as smart as the average human +90% of the time. A lack of perfect to me doesn't mean it isn't AGI.
In my experience, they contain more information than any human but they are actually quite stupid. Reasoning is not something they do well at all. But even if I skip that, they can not learn. Inference is separate from training, so they can not learn new things other than trying to work with words in a context window, and even then they will only be able to mimic rather than extrapolate anything new.
It's not the lack of perfect, it's the lack of reasoning and learning.
I 100% agree that learning is missing. We make up for it in SKILLS.md and README.md files and RAGs of various types. And we train the LLMs to deal with these structures.
I've seen a lot of reasoning in the latest models while engaging in agentic coding. It is often decent at debugging and experimentational, but around 30% it goes does wrong paths and just adds unnecessary complexity via misdiagnoses.
Marketing is marketing, nothing about it was ever about being factual when there is a total addressable market to go after and dollars to be made! This is inline with much of the other marketing that exists in the AI space as it stands now, not mention the use of AGI within the space as it stands currently.
I'm not saying anything is going to happen, ARM holdings has a lot more money and lawyers than Long Blockchain did, but I'm just saying that it's not weird to think that a deceptive name could be considered false advertising.
That would not hold up considering that they consistently use 'agentic' in their press release and make no mention of 'artificial general intelligence'. Just because two things have the same acronym does not mean that they stand for the same thing. Marketing being cheeky is not a crime.
It's not "being cheeky". They know that the holy grail for AI is AGI. They know that people are going to see the acronym AGI and assume Artificial General Intelligence. They know that people aren't going to read the full article.
This isn't just a crass joke or a pun, it's outright deception. I'm not a lawyer, maybe it wouldn't hold up in court, but you cannot convince me that they aren't doing this on purpose.
Yes, that's how fraud works a lot of the time. It removes you from the market but not until after it's removed your money. And there's an endless supply of new people ready to make the same mistake after you've learned your lesson.
Does an iced tea company changing their name to Long Blockchain make any sense? No, not really, it's pretty stupid actually, but it managed to bump the stock by apparently 380%.
The stock market can be pretty dumb sometimes. Let's not forget the weird GME bubble.
This is just a Neoverse CPU that Arm will manufacture themselves at TSMC and then sell directly to customers.
It isn't an "AI" CPU. There is nothing AI about it. There is nothing about it that makes it more AI than Graviton, Epyc, Xeon, etc.
This was already revealed in the Qualcomm vs Arm lawsuit a few years ago. Qualcomm accused Arm of planning to sell their CPUs directly instead of just licensing. Arm's CEO at the time denied it. Qualcomm ends up being right.
This was exactly my first thought when I saw the title. And after reading the contents of the blog, it's pretty clear that ARM is laser focused on getting a piece of their customer's cake by competing with them. This is likely why they are riding the AI hype train hard with their ill-suited name (AGI).
Unfortunately for them, I think hardware vendors will see past the hype. They'll only buy the platform if it is very competitively priced (i.e., much cheaper) since fortune favours long-lived platforms and organizations like Apple and Qualcomm.
I can imagine a lot of ARM engineers being frustrated at seeing their cores being used in stupid ways for decades to finally flex what they can do (outside of Apple).
Agreed, it will be _very_ interesting to see what waves this causes. It would be like TSMC deciding to make and sell their own CPUs, now ARM is directly competing with some of their clients.
Eh, I'm not so sure it'll be that big a deal. The whole supply chain is so twisted and tangled all the way up and down. Shuffling out one piece doesn't seem like it will, on its own, be so major. Samsung made the chips for the iPhone, then made their own phone, then Apple designed their own chips made by TSMC, now Apple is exploring the possibility of having Samsung make those chips again.
Also, it takes a willful ignorance of history for ARM to claim this is the first time they've manufactured hardware. I mean, maaaaybe, teeeeechnically that's true, but ARM was the Acorn RISC Machine, and Acorn was in the hardware business...at least as much as Apple was for the first iPhone.
> I work at ARM, we're launching a new CPU optimized for LLM usage. We're thinking of calling it "Arm Agentic AI Infrastructure CPU", or "Arm AGI CPU" for short. Do you think this is a good idea?
> No. I would not use it as the product name. “AGI CPU” will be read as artificial general intelligence, not “agentic AI infrastructure,” so it invites confusion and sounds hypey.
To bad these executives seemingly don't have access to ChatGPT.
They did ask AI if AGI what a great name.
It said that it was the greatest name possible. It's bold, aspirational, and ... polarizing?!
Oh god! Mistral tell me it's highly polarizing, will make the buzz and it's risky but anyway people will know that ARM is doing CPU again now (maybe I did put too many context).
Seems more likely this falls under the replication crisis umbrella. My wife's favorite numbers are my birthday (mm-dd), which is a small reason she fell in love with me. Neither of those numbers are related to her birthday. My favorite number(s) do not overlap with my birthday. Maybe my mm-dd values just aren't low enough, like 02-02?
> Studies 1-5 showed that people are disproportionately likely to live in places whose names resemble their own first or last names (e.g., people named Louis are disproportionately likely to live in St. Louis).
When I lived in Austin, it seemed like a third of boys born were being named Austin. I presume many of them will end up living there as adults but not because of this particular bias, because they were raised there and have family’s there seems to be a more likely driver.
How fun would it be if due to improved chips handling more model state RAM needs are reduced and Sama cannot make all those RAM purchases he booked?
VC without a degree who has no grasp of hardware engineering failed up when all he had to do was noodle numbers in an Excel sheet.
He is so far behind the hardware scene he thinks its sitting still and RAM requirements will be a nice linear path to AGI. Not if new chips optimized for model streaming crater RAM needs.
Hilarious how last decades software geniuses are being revealed as incompetent finance engineers whose success was all due to ZIRP offering endless runway.
The thing they are good at is bullshitting and selling hype. Which we see here doesn't mean they are actually going to be good at running a business. Smart leaders understand they are not omnipotent and omniscient so they surround themselves who know how to get things done. Weak, narcissist leaders think they're the smartest one in the room and fail.
Unfortunately failing upwards is still somehow common, probably because the skill of parting fools from their money is still valuable.
No, he is also good at networking. When OpenAI was mission-driven and Sam was more respected, he could convince the most talented people to work for him.
Now the talent is going to other places for a variety of reasons, not all due to Sam (one of which is little room for options to grow). However it’s hard to believe his tanking reputation is not badly hurting the company. Other than Jakub and Greg, I believe there are not many top tier people left, those in top positions are there because they are yes-men to Sam.
What RAM? OpenAI booked the silicon wafers, they can print anything they want on them. I wouldn't call them "far behind" on hardware when OpenAI are actively buying Cerebras chips.
Yes exactly; he is behind in that he has to buy others chips with little say on how they work.
Apple and Google control their own designs.
Sama is 100% an outsider, merely a customer. The chip insiders are onto his effort to pivot out of meme-stock hyping, into owning a chunk of their fiefdom. They laughed off his claims a couple years ago as insane VC gibberish (third hand paraphrase from social network in chip and hardware land).
No way he can pivot and print whatever. Relative to hardware industry he is one of those programmers who can say just enough to get an interview but whiffs the code challenge.
He has no idea where the bleeding edge is so he will just release dated designs. Chip IP is a moat.
Plus a bunch of RAM companies would be left hanging; no orders, no wafers. Sama risks being Jimmy Hoffa'd imploding the asset values of other billionaires.
If you read past the marketing talk, this is basically a massively multicore system (136) with significantly reduced power usage (300W).
Where does Agentic come into this? ARMs explanation is that future Agentic workloads will be both CPU and GPU bound thus the need for significant CPU efficiency.
Poor TSMC (and ASML)! They were already struggling with capacity to fulfill orders from their established customers. With ARM now joining the them, I don't know they're going to cope.
Edit: They new CPU will be build with the soon-to-be-former leading edge process of 3nm lithography.
TSMC has multiple fabs being constructed, they'll be okay. The biggest losers here are AMD, Intel and Apple who will be forced to pay AI-hype prices to mass-produce boring consumer hardware.
It isn't obvious to me that they intended to give this as the maximum single-core performance, or just the proportional share of 844GB/s across 136 cores. Implementations of Neoverse V2 by Nvidia and Amazon hit 20-30GB/s in single-threaded work.
is this a cpu that's meant for AI training or is it more for serving inference? I don't quite get why I would want to buy an arm CPU over a nvidia GPU for ai applications.
If I try to cut through the hype, it seems the main features of this processor, or rather processor + memory controller + system architecture, is < 100 ns for accessing anything in system memory and 6 GB/sec for each of a large-ish number of cores, so a (much?) higher overall bandwidth than what we would see in a comparable Intel x86_64 machine.
Call this an “AGI CPU” just feels like the most out of touch, terrible marketing possible. Maybe this is unfair but it makes me think ARM as a whole is incompetent just because it is so tasteless.
> Arm has additionally partnered with Supermicro on a liquid-cooled 200kW design capable of housing 336 Arm AGI CPUs for over 45,000 cores.
Also just bad timing on trying to brag about a partnership with Supermicro, after a founder was just indicted on charges of smuggling Nvidia GPUs. Just bizarre to mention them at all.
Meta are heavily invested in building their own chips with ARM to reduce their reliance on Nvidia as everyone is going after their (Nvidia) data center revenues.
This is why Meta acquired a chip startup for this reason [0] months ago.
Huh, many companies use TSMC, in fact, probably all of them use TSMC, including Intel, yet there are only a few who dominates in performance. There are much more in designing chips than what you just listed.
Intel uses its own fabs for certain IP, tsmc for others yeah. As far as I've seen the latest greatest Panther Lake that stuff is made in intel's arizona fabs.
There's a big difference between just providing IP and actually doing the physical design, manufacturing and packaging. You can't just send your RTL to TSMC and magically get packaged chips back.
I haven't ever ordered an ARM SoC but I also wouldn't be surprised if there were significant parts that they left up to integrators before - PLLs, pads, SRAM etc.
Many of these words are unexplained. "Memory and I/O on the same die". Oh? What does this mean? All of the DRAM in the photo/render is still on sticks. Do they mean the memory controller? Or is there an embedded DRAM component?
I found this article extremely frustrating to read. Maybe I lack some required prior knowledge and I am not the target audience for this.
> built on the Arm Neoverse platform
What the heck is "Arm Neoverse"? No explanation given, link leads to website in Chinese. Using Firefox translating tool doesn't help much:
> Arm Neoverse delivers the best performance from the cloud to the edge
What? This is just a pile of buzzwords, it doesn't mean anything.
The article doesn't seem to contain any information on how much it costs or any performance benchmarks to compare it with other CPUs. It's all just marketing slop, basically.
> The ARM Neoverse is a group of 64-bit ARM processor cores licensed by Arm Holdings. The cores are intended for datacenter, edge computing, and high-performance computing use. The group consists of ARM Neoverse V-Series, ARM Neoverse N-Series, and ARM Neoverse E-Series.
The name of this CPU is bordering on securities fraud. When people see the term "AGI" now, they are assuming "Artificial General Intelligence", not "Agentic AI Infrastructure".
Of course people don't realize that, and people will buy ARM stock thinking they've cracked AGI. The people running Arm absolutely know this, so this name is what we in the industry call a "lie".
The marketers did this for 5G also, calling their product 5G before it was actually deployed, only because theirs came after 4G but wanted to ride the upcoming 5G buzz.
It seems marketing /depends/ on conflating terms and misleading consumers. Shakespeare might have gotten it wrong with his quip about lawyers.
https://www.pbs.org/newshour/economy/att-to-drop-misleading-...
Considering AGI has been degraded into a generic feelgood marketing word, I can't wait to get my AGI-scented deodorant.
You can already drink AGI! Oh sorry, AG1. The resemblance must be a complete coincidence.
> I can't wait to get my AGI-scented deodorant.
Old spice for me, thanks!
If you showed someone what our computers can do with the latest LLMs now to someone 5 years ago they would probably say it sure looks a lot like AGI.
We have to keep defining AGI upwards or nitpick it to show that we haven't achieved it.
I would argue that LLMs are actually smarter than the majority of humans right now. LLMs do not have quite the agency that humans have, but their intelligence is pretty decent.
We don't have clear ASI yet, but we definitely are in a AGI-era.
I think we are missing an ego/motiviations in the AGI and them having self-sufficiency independent of us, but that is just a bit of engineering that would actually make them more dangerous, it isn't really a significant scientific hurdle.
Ok, but it's not AGI. People five years ago would have been wrong. People who don't have all the information are often wrong about things.
ETA:
You updated your comment, which is fine but I wanted to reply to your points.
> I would argue that LLMs are actually smarter than the majority of humans right now. LLMs do not have quite the agency that humans have, but their intelligence is pretty decent.
I would actually argue that they are decidedly not smarter than even dumb humans right now. They're useful but they are glorified text predictors. Yes, they have more individual facts memorized than the average person but that's not the same thing; Wikipedia, even before LLMs also had many more facts than the average person but you wouldn't say that Wikipedia is "smarter" than a human because that doesn't make sense.
Intelligence isn't just about memorizing facts, it's about reasoning. The recent Esolang benchmarks indicate that these LLMs are actually pretty bad at that.
> We don't have clear ASI yet, but we definitely are in a AGI-era.
Nah, not really.
> They're useful but they are glorified text predictors.
There is a long history of people arguing that intelligence is actually the ability to predict accurately.
https://www.explainablestartup.com/2017/06/why-prediction-is...
> Intelligence isn't just about memorizing facts, it's about reasoning.
Initially, LLMs were basically intuitive predictors, but with chain of thought and more recently agentic experimentation, we do have reasoning in our LLMs that is quite human like.
That said, there is definitely a biased towards training set material, but that is also the case with the large majority of humans.
For the Esoland benchmarks, I would be curious how much adding a SKILLS.md file for each language would boost performance?
I am pretty confidence that we are in the AGI era. It is unsettling and I think it gives people cognitive dissonance so we want to deny it and nitpick it, etc.
> There is a long history of people arguing that intelligence is actually the ability to predict accurately.
There sure is, and in psychological circles that it appears that there's an argument that that is not the case.
https://gwern.net/doc/psychology/linguistics/2024-fedorenko....
> Initially, LLMs were basically intuitive predictors, but with chain of thought and more recently agentic experimentation, we do have reasoning in our LLMs that is quite human like.
If you handwave the details away, then sure it's very human like, though the reasoning models just kind of feed the dialog back to itself to get something more accurate. I use Claude code like everyone else, and it will get stuck on the strangest details that humans actively wouldn't.
> For the Esoland benchmarks, I would be curious how much adding a SKILLS.md file for each language would boost performance?
Tough to say since I haven't done it, though I suspect it wouldn't help much, since there's still basically no training data for advanced programs in these languages.
> I am pretty confidence that we are in the AGI era. It is unsettling and I think it gives people cognitive dissonance so we want to deny it and nitpick it, etc.
Even if you're right about this being the AGI era, that doesn't mean that current models are AGI, at least not yet. It feels like you're actively trying to handwave away details.
> though the reasoning models just kind of feed the dialog back to itself to get something more accurate.
Much of our reasoning is based on stimulating our sensory organs, either via imagination (self-stimulation of our visual system) or via subvocalization (self-stimulation of our auditory system), etc.
> it will get stuck on the strangest details that humans actively wouldn't.
It isn't a human. It is AGI, not HGI.
> It feels like you're actively trying to handwave away details.
Maybe. I don't think so though.
What does AGI look like in your opinion?
Personally, I've used LLMs to debug hard-to-track code issues and AWS issues among other things.
Regardless of whether that was done via next-token prediction or not, it definitely looked like AGI, or at least very close to it.
Is it infallible? Not by a long shot. I always have to double-check everything, but at least it gave me solid starting points to figure out said issues.
It would've taken me probably weeks to find out without LLMd instead of the 1 or 2 hours it did.
In that context, I have a hard time thinking how would a "real" AGI system look like, that it's not the current one.
Not saying current LLMs are unequivocally AGI, but they are darn close for sure IMO.
> What does AGI look like in your opinion?
Being able to actually reason about things without exabytes of training data would be one thing. Hell, even with exabytes of training data, doing actual reasoning for novel things that aren't just regurgitating things from Github would be cool.
Being able to learn new things would be another. LLMs don't learn; they're a pretrained model (it's in the name of GPT), that send in inputs and get an output. RAGs are cool but they're not really "learning", they're just eating a bit more context in order to kind of give a facsimile of learning.
Going to the extreme of what you're saying, then `grep` would be "darn close to AGI". If I couldn't grep through logs, it might have taken me years to go through and find my errors or understand a problem.
I think that they're ultimately very neat, but ultimately pretty straightforward input-output functions.
My definition of AGI hasn't changed - it's something that can perform, or learn to perform, any intellectual task that a human can.
5 years ago we thought that language is the be-all and end-all of intelligence and treated it as the most impressive thing humans do. We were wrong. We now have these models that are very good at language, but still very bad at tasks that we wrongly considered prerequisites for language.
A human can think logically with reason, not to say they are smart or smarter. But LLMs cannot. You can convince a LLM anything is correct and it will believe you. You can't convince a human anything is correct.
I can't argue that LLMs do not know an absolute insane amount of information about everything. But you can't just say LLMs are smarter then most humans. We've already decided that smartness is not about how much data you know, but thinking about that data with logical reasoning. Including the fact it may or may not be true.
I can run a LLM through absolutely incorrect data, and tell it that data is 100% true. Then ask it questions about that data and get those incorrect results as answers. That's not easy to do with humans.
> If you showed someone what our computers can do with the latest LLMs now to someone 5 years ago they would probably say it sure looks a lot like AGI.
Would they? Perhaps if you only showed them glossy demos that obscure all the ways in which LLMs fail catastrophically and are very obviously nowhere even close to AGI.
Certainly, they wouldn't expect that an AI able to score 150 on an IQ test is unable to play a casual game of chess because it isn't coherent enough to play without making illegal moves.
> Certainly, they wouldn't expect that an AI able to score 150 on an IQ test is unable to play a casual game of chess because it isn't coherent enough to play without making illegal moves.
To be fair, I am pretty sure Claude Code will download and run stockfish, if you task it to play chess with you. It's not like a human who read 100 books about chess, but never played, would be able to play well with their eyes closed, and someone whispering board position into their ear
> LLMs are actually smarter than the majority of humans right now
I consider myself a bit of a misanthrope but this makes me an optimist by comparison.
Even stupid people are waaaaaay smarter than any LLM.
The problem is the continued habit humans have of anthropomorphizing computers that spit out pretty words. It’s like Eliza only prettier. More useful for sure. Still just a computer.
> Still just a computer.
I don't believe in a separation of mind and spirit. So I do think fundamentally, outside of a reliance on quantum effects in cognition (some of theorized but it isn't proven), its processes can be replicated in a fashion in computers. So I think that intelligence likely can be "just a computer" in theory and I think we are in the era where this is now true.
I don't believe in "spirits" from the get go. I think it's certainly theoretically possible that we could mimic human thought with a computer (quantum or otherwise) but I do not think that the LLMs we have now are doing that. I'd say that what we have right now is "just a computer".
This doesn't mean they aren't useful, I like Claude a lot, but I don't buy that it's AGI.
No they aren't
ChatGPT Health failed hilariously bad at just spotting emergencies.
A few weeks ago most of them failed hilariously bad at the question if you should drive or walk to the service station if you want to wash your car
Idk about the health story, but in my use, ChatGPT has dramatically improved my understanding of my health issues and given sound and careful advice.
The second question sounds like a useless and artificial metric to judge on. The average person might miss such a “gotcha” logical quiz too, for the same reason - because they expect to be asked “is it walking distance.”
No one has ever relied on anyone else’s judgment, nor an AI, to answer “should I bring my car to the carwash.” Same for the ol’ “how many rocks shall I eat?” that people got the AI Overview tricked with.
I’m not saying anything categorically “is AGI” but by relying on jokes like this you’re lying to yourself about what’s relevant.
I would accuse you of nitpicking. My experience is that LLMs are generally as smart as the average human +90% of the time. A lack of perfect to me doesn't mean it isn't AGI.
>> My experience is that LLMs are generally as smart as the average human +90% of the time. A lack of perfect to me doesn't mean it isn't AGI.
In my experience, they contain more information than any human but they are actually quite stupid. Reasoning is not something they do well at all. But even if I skip that, they can not learn. Inference is separate from training, so they can not learn new things other than trying to work with words in a context window, and even then they will only be able to mimic rather than extrapolate anything new.
It's not the lack of perfect, it's the lack of reasoning and learning.
I 100% agree that learning is missing. We make up for it in SKILLS.md and README.md files and RAGs of various types. And we train the LLMs to deal with these structures.
I've seen a lot of reasoning in the latest models while engaging in agentic coding. It is often decent at debugging and experimentational, but around 30% it goes does wrong paths and just adds unnecessary complexity via misdiagnoses.
"look, it completely lied about params that don't exist in a CLI!"
AGI doesn't mean perfect. It means human like and the latest models are pretty human like in terms of their fallibility and capabilities.
AGI is a poorly-defined concept anyway. It’s just vibes, nothing descriptive.
Marketing is marketing, nothing about it was ever about being factual when there is a total addressable market to go after and dollars to be made! This is inline with much of the other marketing that exists in the AI space as it stands now, not mention the use of AGI within the space as it stands currently.
Sure, but there are plenty of cases where a deceptive name has been considered enough to at least warrant an investigation: https://en.wikipedia.org/wiki/Long_Blockchain_Corp.
I'm not saying anything is going to happen, ARM holdings has a lot more money and lawyers than Long Blockchain did, but I'm just saying that it's not weird to think that a deceptive name could be considered false advertising.
That would not hold up considering that they consistently use 'agentic' in their press release and make no mention of 'artificial general intelligence'. Just because two things have the same acronym does not mean that they stand for the same thing. Marketing being cheeky is not a crime.
It's not "being cheeky". They know that the holy grail for AI is AGI. They know that people are going to see the acronym AGI and assume Artificial General Intelligence. They know that people aren't going to read the full article.
This isn't just a crass joke or a pun, it's outright deception. I'm not a lawyer, maybe it wouldn't hold up in court, but you cannot convince me that they aren't doing this on purpose.
> Of course people don't realize that, and people will buy ARM stock thinking they've cracked AGI.
Doesn't seem like a very credible assertion. Picking stocks in this way would remove you from the market pretty quickly.
Yes, that's how fraud works a lot of the time. It removes you from the market but not until after it's removed your money. And there's an endless supply of new people ready to make the same mistake after you've learned your lesson.
Didn't random companies add block chain to their names only just a few years ago and get 30+% jumps in stock price immediately?
I didn't say it would be a wise decision to pick stocks that way, but this has already happened: https://en.wikipedia.org/wiki/Long_Blockchain_Corp.
Does an iced tea company changing their name to Long Blockchain make any sense? No, not really, it's pretty stupid actually, but it managed to bump the stock by apparently 380%.
The stock market can be pretty dumb sometimes. Let's not forget the weird GME bubble.
Those in the industry don't call it a lie, they call it "marketing".
It's those out of the industry who call them lies.
Touché. I guess I should have said "I call it a lie".
In case you haven't noticed, this whole thing has been a grift since 2022. It's kind of amazing that nobody thought of making AGI processors before
the whole AI space is rife with much worse example of what could be considered securities fraud tbh
This is just a Neoverse CPU that Arm will manufacture themselves at TSMC and then sell directly to customers.
It isn't an "AI" CPU. There is nothing AI about it. There is nothing about it that makes it more AI than Graviton, Epyc, Xeon, etc.
This was already revealed in the Qualcomm vs Arm lawsuit a few years ago. Qualcomm accused Arm of planning to sell their CPUs directly instead of just licensing. Arm's CEO at the time denied it. Qualcomm ends up being right.
I wrote a post here on why Arm is doing this and why now: https://news.ycombinator.com/item?id=47032932
This was exactly my first thought when I saw the title. And after reading the contents of the blog, it's pretty clear that ARM is laser focused on getting a piece of their customer's cake by competing with them. This is likely why they are riding the AI hype train hard with their ill-suited name (AGI).
Unfortunately for them, I think hardware vendors will see past the hype. They'll only buy the platform if it is very competitively priced (i.e., much cheaper) since fortune favours long-lived platforms and organizations like Apple and Qualcomm.
This reminds me of Intel talking about faster web browsing with the new Pentium
I think the interesting bit is actually this:
For the first time in our more than 35-year history, Arm is delivering its own silicon products
I can imagine a lot of ARM engineers being frustrated at seeing their cores being used in stupid ways for decades to finally flex what they can do (outside of Apple).
Agreed, it will be _very_ interesting to see what waves this causes. It would be like TSMC deciding to make and sell their own CPUs, now ARM is directly competing with some of their clients.
Eh, I'm not so sure it'll be that big a deal. The whole supply chain is so twisted and tangled all the way up and down. Shuffling out one piece doesn't seem like it will, on its own, be so major. Samsung made the chips for the iPhone, then made their own phone, then Apple designed their own chips made by TSMC, now Apple is exploring the possibility of having Samsung make those chips again.
Also, it takes a willful ignorance of history for ARM to claim this is the first time they've manufactured hardware. I mean, maaaaybe, teeeeechnically that's true, but ARM was the Acorn RISC Machine, and Acorn was in the hardware business...at least as much as Apple was for the first iPhone.
Do they need to higher Design Verification engineers for this?
Thats a huge cost compared to the average RTL jockey
What would be the real advantage of doing that?
AGI (Agentic AI Infrastructure) is joining CSS (Compute Subsystems) in their lineup, apparently. Who’s naming this stuff?
The same people who abbreviate "generative" AI in a way that misleadingly conflates it with "general" AI.
Fraud is just the default lifestyle of marketers.
So Artificial General Intelligence and Cascading Style Sheets are not joining forces?
If there's ever a singularity as a result of AGI, it will likely look at CSS and decide that extermination is simply too good for the human race.
Always have been :)
AGI = Agentic AI Infrastructure
In case you were thinking about some other abbreviation...
Missed opportunity to call it AAII and market it as twice as powerful as regular AI.
We put AI in our AI so the AI is already baked in.
A^2I^2 or (AI)^2
I think this is a poetic encapsulation of the AI industry at this point. A beautifully poignant vignette.
It feels more like "blockchain" to me: https://www.cnbc.com/2017/12/21/long-island-iced-tea-micro-c...
It's... really something. Not good. Something.
It’s like they decided to moon all the onlookers while jumping the shark…
I don’t know if it was intentional or they were so far out over their skis that they got their bathing suit caught, but it’s impressive either way.
Should have called it A^3I^2 - Arm Agentic Artificial Intelligence Infrastructure.
what lenghts are they going to, just to say we have achieved AGI... now who's moving the goalpost?
AGI stands for Artificial General Intelligence.
Pretty sure it stands for "Artificial abbreviation & hype GeneratIon" nowadays
Are you sure it doesn't stand for Advanced Guessing Instrument? That's what the result often seem to indicate after all.
What a terrible, terrible name.
The coast is clear to come up with your own expansion for AI!
I mean, they could at least use AI to figure out how to name their AI product.
> I work at ARM, we're launching a new CPU optimized for LLM usage. We're thinking of calling it "Arm Agentic AI Infrastructure CPU", or "Arm AGI CPU" for short. Do you think this is a good idea?
> No. I would not use it as the product name. “AGI CPU” will be read as artificial general intelligence, not “agentic AI infrastructure,” so it invites confusion and sounds hypey.
To bad these executives seemingly don't have access to ChatGPT.
They did ask AI if AGI what a great name. It said that it was the greatest name possible. It's bold, aspirational, and ... polarizing?!
Oh god! Mistral tell me it's highly polarizing, will make the buzz and it's risky but anyway people will know that ARM is doing CPU again now (maybe I did put too many context).
maybe they did and why they got this slop?
Not bait at all
They pathetically don’t mention what it stands for anywhere in this press release. Deceptive marketing at worst, shameless AI-washing at best.
I would've went for Agentic Neural Infrastructure personally
ARMANI for short /s
This is like naming your kid World President Smith.
This could work. Right? https://psycnet.apa.org/record/2002-12744-001
My realtor's last name is House
Seems more likely this falls under the replication crisis umbrella. My wife's favorite numbers are my birthday (mm-dd), which is a small reason she fell in love with me. Neither of those numbers are related to her birthday. My favorite number(s) do not overlap with my birthday. Maybe my mm-dd values just aren't low enough, like 02-02?
> Studies 1-5 showed that people are disproportionately likely to live in places whose names resemble their own first or last names (e.g., people named Louis are disproportionately likely to live in St. Louis).
When I lived in Austin, it seemed like a third of boys born were being named Austin. I presume many of them will end up living there as adults but not because of this particular bias, because they were raised there and have family’s there seems to be a more likely driver.
"Nominative determinism" is everywhere once you look for it. My vet's last name is McStay.
I just listened to an interview with Carl Trueman about his new book which criticizes transhumanism.
> Studies 1-5 showed that people are disproportionately likely to live in places whose names resemble their own first or last names
There are several cities in the US that share my last name. I don't live near any of them.
> Study 6 extended this finding to birthday number preferences.
D'oh!
My urologist, and I swear I'm not making this up, has the last name "Wiener".
Reporting bias.
I miss the all-capitals ARM spelling.
Seeing "Arm AGI" spelled out on a page with an "arm" logo looks slightly cheesy.
But maybe it's actually a good fit for the societal revolution driven by AGI, comparable to the one driven by the DOT.com RevoLut.Ion. (dot com).
Anyways, it sounds like an A.R.M. branded version of the AppleSilicon revolution?
But maybe that's just my shallow categorization.
Seems like hubris to use this name.
Arm apparently now sells their own CPU's.
How fun would it be if due to improved chips handling more model state RAM needs are reduced and Sama cannot make all those RAM purchases he booked?
VC without a degree who has no grasp of hardware engineering failed up when all he had to do was noodle numbers in an Excel sheet.
He is so far behind the hardware scene he thinks its sitting still and RAM requirements will be a nice linear path to AGI. Not if new chips optimized for model streaming crater RAM needs.
Hilarious how last decades software geniuses are being revealed as incompetent finance engineers whose success was all due to ZIRP offering endless runway.
The thing they are good at is bullshitting and selling hype. Which we see here doesn't mean they are actually going to be good at running a business. Smart leaders understand they are not omnipotent and omniscient so they surround themselves who know how to get things done. Weak, narcissist leaders think they're the smartest one in the room and fail.
Unfortunately failing upwards is still somehow common, probably because the skill of parting fools from their money is still valuable.
No, he is also good at networking. When OpenAI was mission-driven and Sam was more respected, he could convince the most talented people to work for him.
Now the talent is going to other places for a variety of reasons, not all due to Sam (one of which is little room for options to grow). However it’s hard to believe his tanking reputation is not badly hurting the company. Other than Jakub and Greg, I believe there are not many top tier people left, those in top positions are there because they are yes-men to Sam.
What RAM? OpenAI booked the silicon wafers, they can print anything they want on them. I wouldn't call them "far behind" on hardware when OpenAI are actively buying Cerebras chips.
Yes exactly; he is behind in that he has to buy others chips with little say on how they work.
Apple and Google control their own designs.
Sama is 100% an outsider, merely a customer. The chip insiders are onto his effort to pivot out of meme-stock hyping, into owning a chunk of their fiefdom. They laughed off his claims a couple years ago as insane VC gibberish (third hand paraphrase from social network in chip and hardware land).
No way he can pivot and print whatever. Relative to hardware industry he is one of those programmers who can say just enough to get an interview but whiffs the code challenge.
He has no idea where the bleeding edge is so he will just release dated designs. Chip IP is a moat.
Plus a bunch of RAM companies would be left hanging; no orders, no wafers. Sama risks being Jimmy Hoffa'd imploding the asset values of other billionaires.
Oh wow already in use by Meta, OpenAI, and more ?? https://www.arm.com/products/cloud-datacenter/arm-agi-cpu/ec...
The TDP to memory bandwidth& capacity ratio form these blades is in a class of its own, yes?
What does "Built for rack-scale agentic efficiency" even means?
If you read past the marketing talk, this is basically a massively multicore system (136) with significantly reduced power usage (300W).
Where does Agentic come into this? ARMs explanation is that future Agentic workloads will be both CPU and GPU bound thus the need for significant CPU efficiency.
It's volume of tokens consumed x number of agents x rack space. Basically agentic computation density.
We just say words now that sound good for marketing but have no real meaning.
Big "but mongodb is web scale" vibes
It's a code sentence for let's go to the utility room to cross pollinate ideas.
I was gonna say just big DCs in marketing yap but really wtf does that mean?
It's when LLM agents are inefficient that you need a whole rack of servers to get shit done.
Translation: “Can you give us some money pretty please?”
Poor TSMC (and ASML)! They were already struggling with capacity to fulfill orders from their established customers. With ARM now joining the them, I don't know they're going to cope.
Edit: They new CPU will be build with the soon-to-be-former leading edge process of 3nm lithography.
TSMC has multiple fabs being constructed, they'll be okay. The biggest losers here are AMD, Intel and Apple who will be forced to pay AI-hype prices to mass-produce boring consumer hardware.
6GB/s/core
That's...not much right? Maybe it's a lot times N-cores? But I really hope each individual core isn't limited to that.
Edit: 17 minutes to sum RAM?
It isn't obvious to me that they intended to give this as the maximum single-core performance, or just the proportional share of 844GB/s across 136 cores. Implementations of Neoverse V2 by Nvidia and Amazon hit 20-30GB/s in single-threaded work.
What a product name choice! I wasn’t expecting ARM to pivot to selling snake oil.
is this a cpu that's meant for AI training or is it more for serving inference? I don't quite get why I would want to buy an arm CPU over a nvidia GPU for ai applications.
Why not ASI? They aim too low.
Interesting that Jensen Huang joined in the congratulations for this new product!
And the stock is down >2% today
If I try to cut through the hype, it seems the main features of this processor, or rather processor + memory controller + system architecture, is < 100 ns for accessing anything in system memory and 6 GB/sec for each of a large-ish number of cores, so a (much?) higher overall bandwidth than what we would see in a comparable Intel x86_64 machine.
Am I right or am I misunderstanding?
Call this an “AGI CPU” just feels like the most out of touch, terrible marketing possible. Maybe this is unfair but it makes me think ARM as a whole is incompetent just because it is so tasteless.
> Arm has additionally partnered with Supermicro on a liquid-cooled 200kW design capable of housing 336 Arm AGI CPUs for over 45,000 cores.
Also just bad timing on trying to brag about a partnership with Supermicro, after a founder was just indicted on charges of smuggling Nvidia GPUs. Just bizarre to mention them at all.
ARM riding the "everything is AI" train.
So sad.
finally, a CPU capable of making API calls to cloud providers
Meta are heavily invested in building their own chips with ARM to reduce their reliance on Nvidia as everyone is going after their (Nvidia) data center revenues.
This is why Meta acquired a chip startup for this reason [0] months ago.
[0] https://www.reuters.com/business/meta-buy-chip-startup-rivos...
I was wondering who convinced ARM to manufacture hardware. Turns out it was Meta.
Now if only they would go back to being "Acorn RISC Machines" and make a nice desktop home computer again...
One can dream.
DGX Spark is pretty nice. It could be cheaper if they removed the NIC though.
I have the ASUS variant. I like it well enough.
I see the NIC as a form of future proofing, but we'll see.
My Ryzen 9 mini-PC from 2 years ago outperforms this thing in raw CPU Though.
Nuvia/Qualcomm lawsuit and Softbank.
Fabless. Like AMD and Nvidia. So I would think about it more as branding and packaging than Manufacturing
Huh, many companies use TSMC, in fact, probably all of them use TSMC, including Intel, yet there are only a few who dominates in performance. There are much more in designing chips than what you just listed.
Intel uses its own fabs for certain IP, tsmc for others yeah. As far as I've seen the latest greatest Panther Lake that stuff is made in intel's arizona fabs.
There's a big difference between just providing IP and actually doing the physical design, manufacturing and packaging. You can't just send your RTL to TSMC and magically get packaged chips back.
I haven't ever ordered an ARM SoC but I also wouldn't be surprised if there were significant parts that they left up to integrators before - PLLs, pads, SRAM etc.
Many of these words are unexplained. "Memory and I/O on the same die". Oh? What does this mean? All of the DRAM in the photo/render is still on sticks. Do they mean the memory controller? Or is there an embedded DRAM component?
All processors have memory on the same die.
How much, what kind, and what is your source?
All mainstream server CPUs have a megabyte or two of SRAM on a core, of course.
Now every product will have the AI buzzword in it's name, just like 25 years ago product names started with letter e, from electronic.
So we will see AI Toilet Paper launching in the next months.
I found this article extremely frustrating to read. Maybe I lack some required prior knowledge and I am not the target audience for this.
> built on the Arm Neoverse platform
What the heck is "Arm Neoverse"? No explanation given, link leads to website in Chinese. Using Firefox translating tool doesn't help much:
> Arm Neoverse delivers the best performance from the cloud to the edge
What? This is just a pile of buzzwords, it doesn't mean anything.
The article doesn't seem to contain any information on how much it costs or any performance benchmarks to compare it with other CPUs. It's all just marketing slop, basically.
> The ARM Neoverse is a group of 64-bit ARM processor cores licensed by Arm Holdings. The cores are intended for datacenter, edge computing, and high-performance computing use. The group consists of ARM Neoverse V-Series, ARM Neoverse N-Series, and ARM Neoverse E-Series.
https://en.wikipedia.org/wiki/ARM_Neoverse
I feel like this is most products in the AI space lately. More marketing fuzz than substance. Hard to figure out what thing even does.