I think I’ve heard about enough of experts predicting the future lately.
Well classic computers will always limited and power hungry. Quantum computer is the key to AI achieving next level
Quantum computers are only good at a very narrow subset of tasks. None of those tasks are related to Neural Networks, AGI, or the emulation of neurons.
The only people who say this know nothing about quantum or computers
I love the or in this sentence
Oh no!
Anyway…
I’ve been hearing about the imminent crash for the last two years. New money keeps getting injected into the system. The bubble can’t deflate while both the public and private sector have an unlimited lung capacity to keep puffing into it. FFS, bitcoin is on a tear right now, just because Trump won the election.
This bullshit isn’t going away. Its only going to get forced down our throats harder and harder, until we swallow or choke on it.
With the right level of Government support, bubbles can seemingly go on for literal decades. Case in point, Australian housing since the late 90s has been on an uninterrupted tear (yes, even in ‘08 and ‘20).
Fingers crossed.
AI was 99% a fad. Besides OpenAI and Nvidia, none of the other corporations bullshitting about AI have made anything remotely useful using it.
lalal AI has made some great innovations in taking songs and separating them into vocals and instrumentals. that’s a game changer for remix artists.
other than that niche utility and a handful of others, AI is largely bullshit.
Absolutely not true. Disclaimer, I do work for NVIDIA as a forward deployed AI Engineer/Solutions Architect—meaning I don’t build AI software internally for NVIDIA but I embed with their customers’ engineering teams to help them build their AI software and deploy and run their models on NVIDIA hardware and software.
To state this as simply as possible: I wouldn’t have a job if our customers weren’t seeing tremendous benefit from AI technology. The companies I work with typically are very sensitive to CapX and OpX costs of AI—they self-serve in private clouds. If it doesn’t help them make money (revenue growth) or save money (efficiency), then it’s gone—and so am I. I’ve seen it happen; entire engineering teams laid off because a technology just couldn’t be implemented in a cost-effective way.
LLMs are a small subset of AI and Accelerated-Compute workflows in general.
To state this as simply as possible: I wouldn’t have a job if our customers weren’t seeing tremendous benefit from AI technology.
Right because corporate management doesn’t ever blindly and stupidly overinvest in fads that blow up in their faces…
I work with typically are very sensitive to CapX and OpX costs of AI—they self-serve in private clouds. If it doesn’t help them make money (revenue growth) or save money (efficiency), then it’s gone—and so am I.
You clearly have no clue what you’re on about. As someone with a degrees and experience in both CS and Finance all I have to say is that’s not at all how these things work. Plenty of companies lose money on these things in the hopes that their FP&A projection fever dreams will come true. And they’re wrong much more often than you seem to think. FP&A is more art than science and you can get financial models to support any argument you want to make to convince management to keep investing in what you think they should. And plenty of CEOs and boards are stupid enough to buy it. A lot of the AI hype has been bought and sold that way in the hopes that it would be worthwhile eventually or that other alternatives can’t be just as good or better.
I’ve seen it happen; entire engineering teams laid off because a technology just couldn’t be implemented in a cost-effective way.
This is usually what happens once they finally realize spending money on hype doesn’t pay off and go back to more established business analytics, operations research, and conventional software which never makes mistakes if it’s programmed correctly.
LLMs are a small subset of AI and Accelerated-Compute workflows in general.
No one ever said otherwise. And we’re talking about AI only, no moving the goalposts to accelerated computing, which is a mechanism through which to implement a wide range of solutions and not a specific one in and of itself.
That’s fair. I see what I see at an engineering and architecture level. You see what you see at the business level.
That said. I stand by my statement because I and most of my colleagues in similar roles get continued, repeated and expanded-scope engagements. Definitely in LLMs and genAI in general especially over the last 3-5 years or so, but definitely not just in LLMs.
“AI” is an incredibly wide and deep field; much more so than the common perception of what it is and does.
Perhaps I’m just not as jaded in my tech career.
operations research, and conventional software which never makes mistakes if it’s programmed correctly.
Now this is where I push back. I spent the first decade of my tech career doing ops research/industrial engineering (in parallel with process engineering). You’d shit a brick if you knew how much “fudge-factoring” and “completely disconnected from reality—aka we have no fucking clue” assumptions go into the “conventional” models that inform supply-chain analytics, business process engineering, etc. To state that they “never make mistakes” is laughable.
That’s fair. I see what I see at an engineering and architecture level. You see what you see at the business level.
I respect that. Finance was my old career and I hated it. I liked coding more so I went back got my M.S. in CS and now do embedded software which I love. I left finance specifically because of what both of us have talked about. It’s all about using nunber to tell whatever story you want and it’s filled with corporate politics. I hated that world. It was disgusting and people were terrible two faced assholes.
That said. I stand by my statement because I and most of my colleagues in similar roles get continued, repeated and expanded-scope engagements. Definitely in LLMs and genAI in general especially over the last 3-5 years or so, but definitely not just in LLMs.
“AI” is an incredibly wide and deep field; much more so than the common perception of what it is and does.
So I think I need to amend what I said before. AI as a whole is definitely useful for various things but what makes it a fad is that companies are basically committing the hammer fallacy with it. They’re throwing it at everything even things where it may not be a good solution just to say hey look we used AI. What I respect about you guys at Nvidia is that you all make really awesome AI based tools and software that actually does solve problem that other types of software and tools either cannot do or cannot do well and that’s how it should be.
At the same time I’m also a gamer and I really hope Uncle Jensen doesn’t forget about us and how we literally were his core market for most of Nvidia’s history as a business.
Now this is where I push back. I spent the first decade of my tech career doing ops research/industrial engineering (in parallel with process engineering). You’d shit a brick if you knew how much “fudge-factoring” and “completely disconnected from reality—aka we have no fucking clue” assumptions go into the “conventional” models that inform supply-chain analytics, business process engineering, etc. To state that they “never make mistakes” is laughable.
What I said was that traditional software if programmed correctly doesn’t make mistakes. As for operations research and supply chain optimization and all the rest of it, it’s not different that what I said about finance. You can make the models tell any story you want and it’s not even hard but the flip side is that the decision makers in your organization should be grilling you as an analyst on how you came up with your assumptions and why they make sense. I actually think this is an area where AI could be useful because if trained right it has no biases unlike human analysts.
The other thing to sort of take away from what I said is the “if it is programmed correctly” part which is also a big if. Humans make mistakes and we see it a lot in embedded where in some cases we need to flash our code onto a product and deploy it in a place where we won’t be able to update it for a long time or maybe ever and so testing and making sure the code works right and is safe is a huge thing. Tool like Rust help to an extent but even then errors can leak through and I’ve actually wondered how useful AI based tools could eventually be in proving the correctness of traditional software code or finding potential bugs and sources of unsafety. I think a deep learning based tool could make formal verification of software a much cheaper and more commonplace practice and I think on the hardware side they already have that sort of thing. I know AMD/Xilinx use machine learning in their FPGA tools to synthesize designs so I don’t see why we couldn’t use such a thing for software that needs to be correct the first time as well.
So that’s really it. My only gripe at all with AI and DL in particular is when executive who have no CS or engineering background throw around the term AI like it’s the magic solution to everything or always the best option when the reality is that sometimes it is and other times it isn’t and they need to have a competent technology professional make that call.
I would say LLMs specifically are in that ball park. Things like machine vision have been boringly productive and relatively un hyped.
There’s certainly some utility to LLMs, but it’s hard to see through all the crazy over estimations and being shoved everywhere by grifters.
Nvidia made money, but I’ve not seen OpenAI do anything useful, and they are not even profitable.
ChatGPT is basically the best LLM of its kind. As for Nvidia I’m not talking about hardware I’m talking about all of the models it’s trained to do everything from DLSS and ACE to creating virtual characters that can converse and respond naturally to a human being.
Huh?
The smartphone improvements hit a rubber wall a few years ago (disregarding folding screens, that compose a small market share, improvement rate slowed down drastically), and the industry is doing fine. It’s not growing like it use to, but that just means people are keeping their smartphones for longer periods of time, not that people stopped using them.
Even if AI were to completely freeze right now, people will continue using it.
Why are people reacting like AI is going to get dropped?
People pay real money for smartphones.
People pay real Money for AIaaS as well…
Because in some eyes, infinite rapid growth is the only measure of success.
Hope?
Because novelty is all it has. As soon as it stops improving in a way that makes people say “oh that’s neat”, it has to stand on the practical merits of its capabilities, which is, well, not much.
I’m so baffled by this take. “Create a terraform module that implements two S3 buckets with cross-region bidirectional replication. Include standard module files like linting rules and enable precommit.” Could I write that? Yes. But does this provide an outstanding stub to start from? Also yes.
And beyond programming, it is otherwise having positive impact on science and medicine too. I mean, anybody who doesn’t see any merit has their head in the sand. That of course must be balanced with not falling for the hype, but the merits are very real.
The merits are real. I do understand the deep mistrust people have for tech companies, but there’s far too much throwing out of the baby with the bath water.
As a solo developer, LLMs are a game-changer. They’ve allowed me to make amazing progress on some of my own projects that I’ve been stuck on for ages.
But it’s not just technical subjects that benefit from LLMs. ChatGPT has been a great travel guide for me. I uploaded a pic of some architecture in Berlin and it went into the history of it, I asked it about some damage to an old church in Spain - turned out to be from the Spanish civil war, where revolutionaries had been mowed down by Franco’s firing squads.
Just today, I was getting help from an LLM for an email to a Portuguese removals company. I sent my message in English with a Portuguese translation, but the guy just replied back with a single sentence in broken English:
“Yes a can , need tho mow m3 you need delivery after e gif the price”
The first bit is pretty obviously “Yes I can” but I couldn’t really be sure what he was trying to say with the rest of it. So I asked ChatGPT who responded:
It seems he’s saying he can handle the delivery but needs to know the total volume (in cubic meters) of your items before he can provide a price. Here’s how I’d interpret it:
“Yes, I can [do the delivery]. I need to know the [volume] in m³ for delivery, and then I’ll give you the price.”
Thanks to LLMs, I’m able to accomplish so many things that would have previously taken multiple internet searches and way more effort.
Okay now justify the cost it took to create the tool.
There’s a pretty big difference between chatGPT and the science/medicine AIs.
And keep in mind that for LLMs and other chatbots, it’s not that they aren’t useful at all but that they aren’t useful enough to justify their costs. Microsoft is struggling to get significant uptake for Copilot addons in Microsoft 365, and this is when AI companies are still in their “sell below cost and light VC money on fire to survive long enough to gain market share” phase. What happens when the VC money dries up and AI companies have to double their prices (or more) in order to make enough revenue to cover their costs?
Nothing to argue with there. I agree. Many companies will go out of business. Fortunately we’ll still have the llama3’s and mistral’s laying around that I can run locally. On the other hand cost justification is a difficult equation with many variables, so maybe it is or will be in some cases worth the cost. I’m just saying there is some merit.
People are dumping billions of dollars into it, mostly power, but it cannot turn profit.
So the companies who, for example, revived a nuclear power facility in order to feed their machine with ever diminishing returns of quality output are going to shut everything down at massive losses and countless hours of human work and lifespan thrown down the drain.
This will have an economic impact quite large as many newly created jobs go up in smoke and businesses who structured around the assumption of continued availability of high end AI need to reorganize or go out of business.
Search up the Dot Com Bubble.
It’s absurdly unprofitable. OpenAI has billions of dollars in debt. It absolutely burns through energy and requires a lot of expensive hardware. People aren’t willing to pay enough to make it break even, let alone profit
Eh, if the investment dollars start drying up, they’ll likely start optimizing what they have to get more value for fewer resources. There is value in AI, I just don’t think it’s as high as they claim.
People differentiate AI (the technology) from AI (the product being peddled by big corporations) without making clear that nuance (Or they mean just LLMs, or they aren’t even aware the technology has a grassroots adoption outside of those big corporations). It will take time, and the bubble bursting might very well be a good thing for the technology into the future. If something is only know for it’s capitalistic exploits it’ll continue to be seen unfavorably even when it’s proven it’s value to those who care to look at it with an open mind. I read it mostly as those people rejoicing over those big corporations getting shafted for their greedy practices.
the bubble bursting might very well be a good thing for the technology into the future
I absolutely agree. It worked wonders for the Internet (dotcom boom in the 90s), and I imagine we’ll see the same w/ AI sometime in the next 10 years or so. I do believe we’re seeing a bubble here, and we’re also seeing a significant shift in how we interact w/ technology, but it’s neither as massive or as useless as proponents and opponents claim.
I’m excited for the future, but not as excited for the transition period.
I’m excited for the future, but not as excited for the transition period.
I have similar feelings.
I discovered LLMs before the hype ever began (used GPT-2 well before ChatGPT even existed) and the same with image generation models barely before the hype really took off. (I was an early closed beta tester of DALL-E)
And as my initial fascination grew, along with the interest of my peers, the hype began to take off, and suddenly, instead of being an interesting technology with some novel use cases, it became yet another technology for companies to show to investors (after slapping it in a product in a way no user would ever enjoy) to increase stock prices.
Just as you mentioned with the dotcom bubble, I think this will definitely do a lot of good. LLMs have been great for asking specialized questions about things where I need a better explanation, or rewording/reformatting my notes, but I’ve never once felt the need to have my email client generate every email for me, as Google seems to think I’d want.
If we can just get all the over-hyped corporate garbage out, and replace it with more common-sense development, maybe we’ll actually see it being used in a way that’s beneficial for us.
I initially started with natural language processing (small language models?) in school, which is a much simpler form of text generation that operates on words instead of whatever they call the symbols in modern LLMs. So when modern LLMs came out, I basically registered that as, “oh, better version of NLP,” with all its associated limitations and issues, and that seems to be what it is.
So yeah, I think it’s pretty neat, and I can certainly see some interesting use-cases, but it’s really not how I want to interface with computers. I like searching with keywords and I prefer the process of creation more than the product of creation, so image and text generation aren’t particularly interesting to me. I’ll certainly use them if I need to, but as a software engineer, I just find LLMs in all forms (so far) annoying to use. I don’t even like full text search in many cases and prefer regex searches, so I guess I’m old-school like that.
I’ll eventually give in and adopt it into my workflow and I’ll probably do so before the average person does, but what I see and what the media hypes it up to be really don’t match up. I’m planning to set up a llama model if only because I have the spare hardware for it and it’s an interesting novelty.
Apparently, there was only so much IP to steal
Oh nice, another Gary Marcus “AI hitting a wall post.”
Like his “Deep Learning Is Hitting a Wall” post on March 10th, 2022.
Indeed, not much has changed in the world of deep learning between spring 2022 and now.
No new model releases.
No leaps beyond what was expected.
\s
Gary Marcus is like a reverse Cassandra.
Consistently wrong, and yet regularly listened to, amplified, and believed.
I hope it all burns.
I work with people who work in this field. Everyone knows this, but there’s also an increased effort in improvements all across the stack, not just the final LLM. I personally suspect the current generation of LLMs is at its peak, but with each breakthrough the technology will climb again.
Put differently, I still suspect LLMs will be at least twice as good in 10 years.
supermicro’s accountants have just resigned 🤭
I believe this about as much as I believed the “We’re about to experience the AI singularity” morons.
I just want a portable self hosted LLM for specific tasks like programming or language learning.
You can install Ollama in a docker container and use that to install models to run locally. Some are really small and still pretty effective, like Llama 3.2 is only 3B and some are as little as 1B. It can be accessed through the terminal or you can use something like OpenWeb UI to have a more “ChatGPT” like interface.
I have a few LLMs running locally. I don’t have an array of 4090s to spare so I am limited to the smaller models 8B and whatnot.
They definitely aren’t as good as anything you get remotely. It’s more private and controlled but it’s much less useful (I’ve found) than any of the other models.
Welcome to the top of the sigmoid curve.
If you were wondering what 1999 felt like WRT to the internet, well, here we are. The Matrix was still fresh in everyone’s mind and a lot of online tech innovation kinda plateaued, followed by some “market adjustments.”
I think it’s more likely a compound sigmoid (don’t Google that). LLMs are composed of distinct technologies working together. As we’ve reached the inflection point of the scaling for one, we’ve pivoted implementations to get back on track. Notably, context windows are no longer an issue. But the most recent pivot came just this week, allowing for a huge jump in performance. There are more promising stepping stones coming into view. Is the exponential curve just a series of sigmoids stacked too close together? In any case, the article’s correct - just adding more compute to the same exact implementation hasn’t enabled scaling exponentially.
