I think the main problem with AI sustainability is that all this VC investment is burning through ungodly amounts of money to produce something that provides no moat.
Even if all AI investment froze tomorrow, I'd still have my 405B Llama 3.1 model, along with countless other smaller models, and I'd run them to do whatever the heck I felt like doing, with no commitment to any provider.
Writing code with AI? I could swap to a local model. Costs nothing. Provides no revenue to any VC-backed company.
Yes, bigger models will always command a premium for the highest end of reasoning. But you don't always need the best possible reasoning. GPT-3 and early GPT-4 were more than good enough for a ton of use cases last year.
And we've seen the pace of development in the open source world these past two years. The open source community (Meta in particular) has completely obliterated the commercial value of these models.
If there were no "weights available" models, OpenAI would have an incredible, unbeatable moat, and they would be worth truly astounding amounts of money. But as it stands, we all have free, unfettered access to local models far better and cheaper than models that were flagship 18 months ago, and close enough to the performance of current flagship models that it won't make a difference for a ton of use cases.
There is no way to justify the current level of investment, with local models being freely available.
The assumption I'm making, of course, is that this transformer technology won't ever lead to AGI - it will just be another tool in our ever-expanding tool-belt.
"OpenAI will have to continue to raise more money than any startup has ever raised in history, in perpetuity, to survive."
As of 2022, Uber had supposedly raised a total of $25.2B in funding over 32 rounds.
So they may be ahead in the money-drain business.
"There is no way to justify the current level of investment, with local models being freely available."
The business problem OpenAI faces is that their systems aren't good enough that users can trust the results. Slightly crappier results are also available at much lower cost. Unless OpenAI can definitively fix the "hallucination" problem, and at least return "Don't know" when appropriate, this isn't going to work at OpenAI's price point.
Chatbots are about as good as low-end outsourced call centers. They can sort of help with programming. The systems that generate pictures can do some impressive things. LLMs produce better blithering than most bloggers. It's really impressive. But, absent a major theoretical breakthrough, that's not enough to fund OpenAI.
One big difference is Uber only has a few peers and became instrumental before the realities of market size caught up to them.
But every hyperscale data center owner is investing for themselves, major companies capable of the R&D are investing in their own science, and the market for LLMs is slow to materialize relative to the hype. That's money that might have gone to OpenAI if their situation was similar to Uber's.
> The business problem OpenAI faces is that their systems aren't good enough that users can trust the results.
It depends what you're doing with it. Answering arbitrary questions is very different from document summarization or from mapping arbitrary questions onto a list of canned questions to then answer.
It also doesn't have to be perfect. Instead the cost of an overall system (AI + error reporting & handling) just has to be cheaper than that same overall system using whatever the AI is replacing.
If it makes errors 3% of the time, and people from Mechanical Turk make errors 2% of the time, its usefulness depends on whether or not those 50% more errors cost more than the cost savings from paying a model provider rather than paying humans.
Humans hallucinate too though. A lot. A good tip for system design is to assume the human operator goes off the rails at some point and does something absurd and nonsensical.
Hallucinations don't need to be fixed in one go as much as improved progressively. There is some magic threshold where they'll stop being an issue for specific tasks, or alternatively they simply become more reliable than humans. The problem will die with a whimper.
What's the misunderstanding? The hallucinations I've seen from AI tend to not be that dissimilar to when humans are guessing. They will guess that a function like that exists. The difference being that humans are capable of checking their work and see if this actually is true by referencing some existing source of truth.
> I'm not entirely cynical on the value of LLMs, but I've yet to see one say "I dont know", or "I'm not sure, but here's my best guess".
I've used LLMs to build form autofilling based on unstructured documents. It correctly does not answer fields that it doesn't know, and does not try to guess anything. It has been pretty much error-free.
It's all about your prompting. Without explicitly being given guidance on how not to answer, you're right, they will never say they don't know.
If you've never actually seen that happen, I encourage you to experiment more with LLMs; there's lots that can be achieved with the right prompting.
It seems like OpenAI is trying to build a moat now, each generation appears to be less and less open to the point that you can't even see LLM reasoning steps.
It's been a while since we've had a major tech company commodify/open source a major dependency (and not just layer some rent seeking pay/subscription service over existing assets). Meta seems to have the best long term plan for LLMs.
I like your comment, but it doesn't go far enough! The other issue here is the price of FLOPS is still dropping; so even if training the model weights was a moat it will probably evaporate quickly over the next few years if price trends hold. All this capital spend would have been much cheaper if deferred by a few years.
If they aren't making massive profits right now then there may well be a problem.
Not only that, but these models are at some point superior. Even if AI companies offer embedding services, which some do not or only recently added such capabilities, the best results are often a specifically tuned openly available model like llama.
I do think this moat can still exist though, as providing a product is different than running a model somewhere, even if you have the compute power to allow prompting by a large amount of people. And making a model ready for the business user is quite a bit of work.
Although true, the capabilities of commercial products will probably be eclipsed sooner rather than later. We saw that in image generation. Crowd sourcing the problem did produce better results. Generating videos is probably still locked behind enormous compute power (well, the 400+B model of Llama is too...)
I think the focus of these companies isn't the performance of the model itself, it is providing interfaces to all kinds of systems and maybe solve a specific task.
Sounds like a good plan until the US and EU declare open-weight models "unsafe" and ban them. "It can help somebody create a bioweapon," etc.
I would imagine lobbying efforts and scaremongering is already underway. This is a powerful technology that the general public (and small businesses) might not have access to in the near future.
I think the point is that anyone can enter the AI cloud market to compete with the fat cats if they get too pricy. NVIDIA's de-facto standards are the closest thing to a market moat. If AI ends up highly tied to NVIDIA's standards, then NVIDIA could become the Microsoft of AI. It's hard to know how tool chains will evolve to answer that though, it's new territory. (It could be all moot if the bots take over and eat us humans. I taste like chicken.)
You can get a LOT of compute credits, at the public rates the cloud providers charge (not OpenAI's discount) for the price of an NVIDIA 4xxx with 24GB or more of RAM. Enough to do what most people need AI to do for a long time.
AWS' moat, if they have one, is not in renting out generic compute. That's been done since before anyone ever heard the term "cloud".
I think the point was that the things LLMs are useful for can increasingly be done with freely available models you can run on your own hardware, hardware you rent from someone else, or via a 3rd party's hosted service. Want to run Meta's models on Google's server? Nothing stopping you.
As it has been mentioned numerous times in the past, we are not the average user. I don't know anyone outside our bubble who would bother running a local LLM, or going into all the trouble building an x8 GPU rig to run one of the biggest models.
Why would non-geeks want to run a local LLM? Electricity prices alone are quite high. Why not pay a subscription on a cloud service and be done with it.
interesting. I have not found Llama able to do any of the sort of 'reasoning' that the larger foundational models can do. It's too easily tricked up by small things.
The larger models aren't perfect, but worlds apart from Llama.
I always find it really strange when articles like this claim nobody is paying for generative AI. I can't find reliable stats on this but there are at least a million ChatGPT Plus subscribers. Does that not count?
It takes time for new advancements to get proliferated through the economy.
Yes, those subscribers do count, but the author writes of sustainability doubts in that respect, arguing that current hype and FOMO-style thinking may well fizzle out:
>I believe that a lot of businesses are "trying" AI at the moment, and once those trials end (Gartner predicts that 30% of generative AI projects will be abandoned after their proof of concepts by end of 2025), they'll likely stop paying for the extra features, or stop integrating generative AI into their companies' products.
I don't disagree it is unsustainable, I'm really trying to be more precise about whether anyone is getting value out of the tools. I'm just really skeptical that nobody is getting value.
I don't think anyone is arguing it's giving zero value, only that the total value is not keeping up with investments. Look at the recent smart-speaker bubble: people indeed use them, but not often enough for profitable activities, and so the speaker market popped. Trying to goad people into shopping with them failed.
Making AI pictures of Pee Wee Herman riding a shark blindfolded is indeed fun, but not profitable for Microsoft because it's too hardware intensive for ads to cover. I gotta make more goofy pics before the bottom falls out...
Objective info is currently hard to come by, but my horse sense is that Big AI is subsidizing it for many projects, including those in other companies, to gain both market share and investor excitement (deserved or not).
If one looks at most the bubbles of the past, the writing is on the wall. AI won't go away, but will probably take longer to make profitable than anticipated, just like dot-coms and smart-speakers. Force feeding it is causing indigestion, and it's likely to PukeGPT.
The article does mention that OpenAI has huge revenue.
> While The Information reported that OpenAI's revenue is $3.5 to $4.5 billion a year in July, The New York Times reported last week that OpenAI's annual revenues have "now topped $2 billion," which would mean that the end-of-year numbers will likely trend toward the lower end of the estimate.
But then the author claims that the business value is questionable.
> And even if they did, it isn't clear whether generative AI actually provides much business value at all. The Information reported last week that customers of Microsoft's 365 suite [snip]
I would have appreciated a deeper discussion of why OpenAI's revenue isn't a data point toward generative AI having some business value. Presumably if nobody was using generative AI in a way that gives them value, OpenAI wouldn't be using all those GPU hours. That's what I was missing from the article personally.
I though his point was fairly clear. When Microsoft tries to charge non-AI companies at or above cost for generative AI products basically no one takes the deal. OpenAI and Anthropic have a lot of revenue but:
> Based on how unprofitable they are, I hypothesize that if OpenAI or Anthropic charged prices closer to their actual costs, there would be a ten-to-a-hundred-times increase in the price of API calls, though it's impossible to say how much without the actual numbers.
Lots of revenue is flowing into generative AI but would that trend continue if they started needing to actually cover costs? And how much of that revenue is from AI companies that would pop right alongside them?
> I would have appreciated a deeper discussion of why OpenAI's revenue isn't a data point toward generative AI having some business value. Presumably if nobody was using generative AI in a way that gives them value, OpenAI wouldn't be using all those GPU hours. That's what I was missing from the article personally.
Firstly, the revenue numbers are rumors. I have no doubt OpenAI has significant revenue, but both reported numbers are imprecise suggesting they likely are estimates at best.
Furthermore, a non-trivial amount of OpenAI's revenue is certainly coming from other AI startups. They in turn are likely burning investor cash, which isnt an indication that OpenAI is providing business value, its an indication that they provide a tool to speculate on future value. Or, more cynically, the provide a tool for companies to convince investors to give them more money.
I agree with a lot of what's said here, but don't agree fully with the doom and gloom.
The job losses have already happened. Companies have laid off quite a bit of employees because they wanted to get ahead of the AI wave. They thought they could replace most of their engineers, and turn 1x into 10x. The only field that has benefited is the parasitical companies that have sprung up around these AI services trying to rentseek their way to profitability. So when the bubble bursts, laid off talent will be able to demand a premium to come back and fix the smoldering remains.
That said, it's still going to cause reasonably bad damage as a whole because so much of the tech industry is dependent on angel investors which behaves in almost cult-like ways when it comes to trying to find something to fund.
The job losses were entirely from post Covid earnings expectations falling and from anti-inflationary efforts like the fed increasing rates. Very little if it had anything to do with AI.
The two are intrinsically tied together because many departments eviscerated their engineering talent while leaving their massive AI investments and teams untouched. That's entirely my point and is a particularly hard one to refute if you look at who is doing the layoffs and in what departments.
Is that really a good example of LLM capabilities? LLMs don't even see those letters because of tokenization.
It's a bit like asking a Chinese speaker questions about imaginary Latin alphabet letters in Han characters. Sure, it demonstrates a limitation, but it's a bit of an edge case.
Its a good example of what LLMs arent capable of. If they think the word Mississippi has the letter A in it (per article), thats a strong indication that the transformer architecture may never be able to achieve AGI.
Here's another example of a simple question that a state of the art model (Claude 3.5) gets wrong, as tested just now:
Prompt: How many words are in this sentence?
Response: This sentence contains 7 words.
Interestingly, it seemed to count the number of words in my sentence correctly, but still answered incorrectly.
> Its a good example of what LLMs arent capable of. If they think the word Mississippi has the letter A in it (per article), thats a strong indication that the transformer architecture may never be able to achieve AGI.
It only indicates that tokenization is used. The LLM architecture can be used without tokenization. It would just mean that the available space in the context window would be used less efficiently. For example, a 10 letter word which could be represented by one token would instead take 10 slots.
Tokenization reorganizes information but doesn't remove it. It may be easier/harder to learn stuff like letter counting with different tokenization schemes, but the main reason it's hard is that there's not much text about letter counting in the training set. Ie, you could easily train any of the ChatGPT models to count letters in words by generating a bunch of training samples explicitly for this task, but it's not worth the bother.
Tens of thousands of people will lose their jobs? Do that many people even work on AI development? And why does OpenAI burn so much money? Maybe they'll have to stop offering their compute time at a discount, but generative AI is doing useful work and will never go away. There are many companies using it to productive ends and there will be demand for programmers to integrate generative AI in business processes.
> Do that many people even work on AI development?
That's not the relevant figure. It's more about how much capital has been wagered by large tech cos and the major startups & VCs on a near-complete takeover of white-collar work by AI (as that's the only level of payoff that could possibly justify the mind-boggling level of investment).
> Tens of thousands of people will lose their jobs? Do that many people even work on AI development?
The implication is if AI is a bust, Google, Microsoft and Meta alone would have to run leaner organisations to make sense. Each of them is massively betting on LLMs are their core growth engine.
None of those orgs is going to run leaner if AI is a bust. They'll invest as much, if not more, into trying to find other ways to grow - exactly the same reason they all invested in AI in the first place.
> They'll invest as much, if not more, into trying to find other ways to grow
Management will try to hold on to the capital for as long as shareholders let them. But if the growth prospects from AI significantly diminish, 20,000 jobs in tech is a pretty conservative estimate for lay-offs.
I’m 100% sure the rush to add AI bots of various sorts (often just natural-language document search—which often replaces such complete neglect of making things easier to find that I wonder how much good it’ll actually do vs any traditional approach applied well if they’d ever bothered to do that, or even vs nothing since that’s apparently been fine so far) across basically all of every industry as far as I can tell, is helping prop up developer employment rates until rates drop and stupid money inflates the AI thing into a real bubble on the investor side.
It's the opposite in my experience. These sort of things are actually resulting in lower developer rates. Even if you went with an out-of-the-box solution for customer support (as an example) you still needed engineers and designers on hand to customize and maintain the solution. But that's almost entirely gone away.
Good article but not sure if a 'subprime crisis' is the right description - Openai is suppose to be risky and extremely high reward. This is exactly what VCs want to invest in.
Maybe OpenAI doesn't make it and maybe it turns out way too early for AGI ... but there is way too much stuff is working right now in multiple domains to say the industry will flop.
These boom/bust cycles are a symptom of a top-heavy economy. Too much money is chasing too little fundamental innovation. Generative AI is cool, sure, but it’s more like a Starlink-sized idea imo. We need more of those types of ideas, not to pump trillions more dollars into making copycat LLMs.
We had some really interesting developments in machine learning before LLMs - they were slow but steady, every week something interesting came up. And them came the LLMs and basically most money started flowing in this direction, for better or worse. It's hard to say whether it makes sense or not but what has been achieved already has some value for some particular tasks.
I find this to be a bit exaggerated but there are a few fair points. I think people are starting to see that ai is not as good as the demos and flashy videos makes them believe. Since the day chatgpt came out, my biggest problem with it has been the fact that you will ask it a mildly difficult question on any subject and it will confidently give you a response that's complete bs. And I'm starting to see people coming to the same conclusion as time goes by, despite the advancements. Ai has a place in this world even with it's current limitations and shortcomings, no doubt about it - there are cases in which it does a nearly perfect job - I'm currently building an AI rig for my home lab because it has a ton of use cases even outside my daily work(where I do use llms for processing unstructured data).
The second issue I see is that data is becoming widely unavailable-everything is getting blocked, crawlers are getting cut off immediately, 1000 API calls cost as much as a mid-sized flat in a European capital - the complete opposite of what the internet was supposed to be.
The third issue is the fact that people are gladly using chatgpt to answer questions on stackoverflow for example. We know how badly llms start performing when you train them on their own data... Not to mention the considerable spike in critical bugs in open source projects over the last two years-there's a good chance it a lot to do with it.
The fourth is social media-there are already tons of examples where troll farms are no longer paid workers in some dump in siberia but are in fact powered by chatgpt out some other service. And to think that I laughed at the dead internet theory when I first heard about it...
The article seems to be saying that the possible returns on the investments doesn't seem to match the amounts the AI companies are asking of investors.
My problem with AI boom seems is that the insane valuations of these AI companies seems to be based on nothing more than the power of the better funded AI companies to outbid possible competitors for limited amounts of chips available from chip foundries, ie TSMC et al.
If there is a glut of chips or the models become more refined and efficient then what power do these companies then have?
The best usage of AI that I have seen is where people need something to show but quality doesn't matter much. AI is great to generate a lot of "something" (data, text), and if you pay attention you'll start to see where it fails. But for jobs/tasks where nobody pays attention (or nobody cares), AI is great.
Need to write a report to present at your company but nobody will care about it? Use AI.
Blog posts just for SEO? Use AI.
Illustration image as a header for a post where you need an image just to share the post on social? Use AI.
I am not saying people is in the right to do this, but this is where I saw AI being really useful at.
Ah no, I am not saynig they are useful from a bussiness point of view (I have no idea about that). I am saying more as a user who uses GPT for free, for example.
Though I agree with almost everything said in this article, I can't help but realise that this is all feelings driven and not data driven. I also _feel_ that the current direction of generative AI is unsustainable and this will all collapse horribly if companies and investors keep following the current trajectory. But unfortunately we have very little data or precedence, either from this article and in general historically, to support this bias.
I use ai-chat quite a bit and the better it gets, I feel more threatened but that is because it leaves me some free time to think like that. And every other month when ai-chat starts spewing garbage answers, I feel pissed at the AI for making me do my research, but it gives me heart warm knowing that this shit cannot replace me.
I have also come to realise that AI needs to be trained to give you correct answers and cannot simply innovate on its own, which is what it needs to be "revolutionary". Also, our entire tech industry is based on products that do deterministic information retrieval. Whether that is getting accurate numbers from a bank account, or computing medical parameters or the velocity of incoming missiles from a bunch of formulae. AI on the other hand seems like tech that will give out answers like "the sum of 1 and 2 is 3 with 99.9% probability".
In any case, these are all just feelings and though I find myself nodding along with the article, there is no information here that is concrete.
GPT-4 is genuinely useful for some stuff (summarizing documents, giving hints how to solve things in languages I do not speak, etc.)
Meta and others have released open weights with the claim that they compare to GPT-4; I imagine these are good enough for many of the similar tasks. There are bound to be at least a few more improvements in open weights before bust.
Apple is already building laptops with a mind towards local AI. As NVidia's and AMD's et al AI chips drop in price, they will be included in regular desktops and laptops.
While local-AI becomes more practical, the prices on remote-AI will go up, further driving local-AI improvements. Perhaps at some point will will have a subscription based weights service, where you get updated proprietary weights for your local model for $X/yr.
Local-AI will be fine for Microsoft. And for Google. So, I don't think AI is going to disappear; If anything, it will become more ubiquitous. Weights may start being released less frequently and the SaaS model may go, but that would likely be a net gain.
It's the nature of all cycles. I work in AI and I think the capacity for AI hasn't even been scratched as a utility. We'll see it continue to make breakthroughs and be applied across tons of business cases. We're in the "this is neat, how can I learn more and apply it more places" phase -- like a toy phase.
If the article is right, watching the dinosaurs fail will be a sad sight, but I see a hope for a next era of small and swift mammals - small, task-specific, often local run models running on my phone, PC, VPS ...
AGI to end the humanity will need to be financed by the Chinese Communist Party, though.
The obvious thing to do, if gen AI proves bust, is to funnel funding and other resources into miltech.
This will further alter the personality, intent, and product of Silicon Valley; refuseniks will be winnowed, and many people may find themselves working on projects they may find disturbing.
Silicon Valley, and computing technology in general, has its origins in the US military. It would not be surprising to see the industry return to those origins as great power competition heats up.
This blog post is so spot on it actually hurts. Big Tech is out of ideas and the MBA asshats running those orgs are looking to sell us on the product of their gullibility. The bubble is going to burst and a lot of people will eat shit when that happens.
While I love generative AI and support open source AI movements because of my enthusiasm for the technology, I have been extremely skeptical about all the efforts to monetize it in big time capitalist terms, both from big companies and several friends who thought AI was their big break to get into tech millions. There's just no full product there that isn't tacking an LLM or image generation into an existing product. So, at face value, I agree with this article, but I think it feels a little too gleeful, a little too "I told you so", even if in some ways it validates my own personal "I told you so". I worry the author overall does not like AI technology, which is popular to dislike these days, and that is coloring the overall tone of the piece, it's hard to tell.
I think businesses are more focused on getting AI to a point where it can do trillions of dollars of work by itself, and then selling those services. Charging for a chat bot in the short term is just a stop gap until they get there
Even if all AI investment froze tomorrow, I'd still have my 405B Llama 3.1 model, along with countless other smaller models, and I'd run them to do whatever the heck I felt like doing, with no commitment to any provider.
Writing code with AI? I could swap to a local model. Costs nothing. Provides no revenue to any VC-backed company.
Yes, bigger models will always command a premium for the highest end of reasoning. But you don't always need the best possible reasoning. GPT-3 and early GPT-4 were more than good enough for a ton of use cases last year.
And we've seen the pace of development in the open source world these past two years. The open source community (Meta in particular) has completely obliterated the commercial value of these models.
If there were no "weights available" models, OpenAI would have an incredible, unbeatable moat, and they would be worth truly astounding amounts of money. But as it stands, we all have free, unfettered access to local models far better and cheaper than models that were flagship 18 months ago, and close enough to the performance of current flagship models that it won't make a difference for a ton of use cases.
There is no way to justify the current level of investment, with local models being freely available.
The assumption I'm making, of course, is that this transformer technology won't ever lead to AGI - it will just be another tool in our ever-expanding tool-belt.