Based on what ? The gap between the release of GPT-3 and 4 is still much bigger than the time that has elapsed since 4 was already released so really, Based on what ?
I don't think we can continue simply tweaking the transformer architecture to achieve meaningful gains. We will need new architectures, hopefully ones that more closely align with biological intelligence.
In theory, the simplest way to real superhuman AGI would be to start by modeling a real human brain as a physical system at the neural level; a real neural network. What the AI community calls "neural networks" are only very loose approximations of biological neural networks. Real neurons are subject to complex interactions between many different neurotransmitters and neuromodulators and they grow and shift in ways that look nothing like backpropagation. There already exist decently accurate physical models for single neurons, but accurately modeling even C. elegans (as part of the OpenWorm project) is still a way's off. Modeling a full human brain may not be possible within our lifetime, but I also wouldn't rule that out.
And once we can accurately model a real human brain, we can speed it up and make it bigger and apply evolutionary processes to it much faster than natural evolution. To me, that's still the only plausible path to real AGI, and we're really not even close.
At least for the use cases I've been directly exposed to, I don't think that is the case. They need to keep being priced about where they are right now. It wouldn't take very much of a rate hike for their end users to largely decide that not using the product makes more financial sense.
OpenAI went from GPT-4, which was mind blowing, to 4o, which was okay, to o1 which was basically built in chain-of-thought.
No new Whisper models (granted, advanced voice chat is pretty cool). No new Dalle models. And nobody is sure what happened to Sora.
We see little progress in "soft" skills like creative writing. EQBench is a benchmark that tests LLM ability to write stories, narratives, and poems. The winning models are mostly tiny Gemma finetunes with single-digit parameter counts. Huge foundation models with hundreds of billions of parameters (Claude 3 Opus, Llama 3.1 405B, GPT4) are nowhere near the top. (Yes, I know Gemma is a pruned Gemini). Fine-tuning > model size, which implies we don't have a path to "superhuman" creative writing (if that even exists). Unlike model size, fine-tuning can't be scaled indefinitely: once you've squeezed all the juice out of a model, what then?
OpenAI's new o1 model exhibits amazing progress in reasoning, math, and coding. Yet its writing is worse than GPT4-o's (as backed by EQBench and OpenAI's own research).
I'd also mention political persuasion (since people seem concerned about LLM-generated propaganda). In June, some researchers tested LLM ability to change the minds of human subjects on issues like privatization and assisted suicide. Tiny models are unpersuasive, as expected. But once a model is large enough to generate coherent sentences, persuasiveness kinda...stops. All large models are about equally persuasive. No runaway scaling laws are evident here.
This picture is uncertain due to instruction tuning. We don't really know what abilities LLMs "truly" possess, because they've been crippled to act as harmless, helpful chatbots. But we now have an open-source GPT-4-sized pretrained model to play with (Llama-3.1 405B base). People are doing interesting things with it, but it's not setting the world on fire.
Isn't that kind of obvious? Even human speakers and writers have problems changing people's minds, let alone reliably.
People are persuaded to change their opinions based on social proof, so this isn’t surprising.
Many people have anecdotally, when I tell them what I do for a living, have told me that any other profession would have the common sense/street smarts to not make their scarce skill redundant. It goes further than that; many professions have license requirements, unions, professional bodies, etc to enforce this scarcity on the behalf on their members. After all a scarce career in most economies is one not just of wealth but higher social standing.
If all it does is allow us to churn more high level software, which let's be honest is demand inelastic due to mostly large margins on software products (i.e. they would of paid a person anyway due to ROI) it doesn't seem it will add much to society other than shifting profit in tech from Labor to Capital/owners. May replace call centre jobs too I guess and some low level writing jobs/marketing. Haven't seen any real new use cases that change my life yet positively other than an odd picture/ai app, fake social posts,annoying AI assistants in apps, maybe some teaching resources that would of been made/easy to acquire anyway by other means etc. I could easily live without these things.
If this is all it is seems Ai will do or mostly do it seems like a bit of a disappointment. Especially for the massive amount of money going into it.
If creating software becomes cheaper then that means I can transform all the ideas I’ve had into software cheaply. Currently I simply don’t have enough hours in the day, a couple hours per weekend is not enough to roll out a tech startup.
Imagine all the open source projects that don’t have enough people to work on them. With LLM code generation we could have a huge jump in the quality of our software.
Well, that's good for them, but bad for humanity in general.
If we had a choice between a system where doctors get high salary and lot of social status, or a system where everyone can get perfect health by using a cheap device, and someone would choose the former, it would make perfect sense to me to call such person evil. The financial needs of doctors should not outweigh the health needs of humanity.
On a smarter planet we would have a nice system to compensate people for losing their privilege, so that they won't oppose progress. For example, every doctor would get a generous unconditional basic income for the rest of their life, and then they would be all replaced by cheap devices that would give us perfect health. Everyone would benefit, no reason to complain.
On a nicer planet sure, we would have a system like that. But most of the planet is not like that - the great advantage of the status quo is that even people who are naturally not altruistic somewhat co-operate with each other due to mutual need. Besides there is ways to mitigate that and still give the required services especially if they are commonly required. The doctors example - certain countries have worked it out without resorting to AI risks. I'm not against AI ironically in this case either, there is a massive shortage of doctors services that can absorb the increased abundance Imv - most people don't put software in the same category. There is bad sides to humanity with regards to losing our mutual dependence on each other as well (community, valuing the life of others, etc) - I think sadly AI allows for many more negatives than simply withholding skills for money if not managed right, even that doesn't happen everywhere today and is a easier problem to solve. The loss of any safe intelligent jobs for climbing and evening out social mobility due to mutual dependence of skills (even the rich can't learn everything and so need to outsource) is one of them.
Though maybe you are just using exponential figuratively in place of meaning rapid and significant development and investment.
I'm still a big critic of AI generally but they're definitely not as bad as crypto which is shocking.
I am not ocd-riddled insecure man trying to subconsiously immitate much of the crowd, in any form of fasion. If that will make me an outlier, so be it, a happier one.
I suspect new branch of artisanal human-mind-made trademark is just behind the corner, maybe niche but it will find its audience. Beautiful imperfections, clear clunky biases and all that.
I agree 100% with this sentiment, but, it also is a decent description of individual humans.
This is what processes and control systems/controls are for. These are evolving at a slower pace than the LLMs themselves at the moment so we're looking to the LLM to be its own control. I don't think it will be any better than the average human is at being their own control, but by no means does that mean it's not a solvable problem.
But you can understand individual humans and learn which are trustworthy for what. If I want a specific piece of information, I have people in my life that I know I can consult to get an answer that will most likely be correct and that person will be able to give me an accurate assessment of their certainty and they know how to accurately confirm their knowledge and they’ll let me know later if it turns out they were wrong or the information changed and…
None of that is true with LLMs. I never know if I can trust the output, unless I’m already an expert on the subject. Which kind of defeats the purpose. Which isn’t to say they’re never helpful, but in my experience they waste my time more often than they save it, and at an environmental/energy cost I don’t personally find acceptable.
Another example from just yesterday is I needed to solve a complex recurrence relation. A friend of mine who is good at math (math PhD) helped me for about 30 minutes still without a solution and a couple of false starts. Then he said try ChatGPT and we got the answer in 30s and we spent about 2 minutes verifying it.
Is anything googleable in 30s? It feels like finding the right combination of keywords that bypasses the personalization and poor quality content takes more than one attempt these days.
Wow. Nice of you to see a coworker go through a traumatic life event, and the best you can drudge up is to bitch about lost productivity and decrease in selfless output of quality to someone else's benefit when they are at the time trying to stitch their life back together.
SMH. Goddamn.
Hope your recurrence relation was low bloody stakes. If you spent only two minutes verifying something coming out of a bullshit machine, I'd hazard you didn't do much in the way of boundary condition verification.
I feel sorry for your family.
A friend of mine was moving from software development into managing devs. He told me: "They often don't do things the way or to the quality I'd like, but 10 of them just get so much more done than I could on my own." This was him coming to terms with letting go of some control, and switching to "guiding the results" rather than direct control.
The LLMs are a lot like this.
The LLMS are a lot like this.
But humans can be held accountable, LLMs cannot.
If I pay a human expert to compile a report on something and they decide to randomly make up facts, that's malpractice and there could be serious consequences for them.
If I pay OpenAI to do the same thing and the model hallucinates nonsense, OpenAI can just shrug it and say "oh that's just a limitation of current LLMs".
Humans may be wrong a lot but at least the vast majority will have the decency to say “I don’t know”, “I’m not sure”, “give me some time to think”, “my best guess is”. In contrast to most LLMs today that in full confidence just spews out more hallucinations.
Why would that be a good thing? The big thing with computers is that they are reliable in ways that humans simply can't ever be. Why is it suddenly a success to make them just as unreliable as humans?
If we are evaluating LLM suitability for tasks typically performed by humans, we should judge them by the same standards we judge humans. That means it's OK to make mistakes sometimes.
Agreed. If everyone understood that and operated under that assumption, it wouldn’t be that much of an issue. Alas, these guessing machines are marketed as all-knowing oracles that can already solve half of humanity’s problems and a significant number of people treat them as being right every time, even in instances where they’re provably wrong.
If you can't see how this tech is useful Idk what to tell you, you have no imagination AND aren't looking around you at the products, marketing, etc. that already exists. These takes remind me of the luddites of ~2012 who were still doubting the Internet in general.
You're confused at best.
But hype also exists. How closely they are matched is not yet clear.
But your comment seems to indicate that the "pro-tech" position is automatically the best. This is _not_ true, as cryptocurrency has already demonstrated.
His whole argument against AI is basically the anti-tech stance: "Well crypto failed that means AI will too" It's coming from a place of disdain for technology. That's your typical Hacker News commenter. This site is like Fox News in 2008 - some of the dumbest people alive
What I am against is the “if it’s tech it’s good” mindset that seems to have infected far too many. I mention crypto because it’s the largest example of tech that is not good for the world.
The only thing that LLMs are at risk of subverting is the livelihood of millions of people. AI is a capital intensifier, so the rich will get richer as it sees more uptake.
About copyright - yeah, I’m quite concerned for my friends who are writers and artists.
> You'll get left behind with these takes, it's not smart. If you don't care about advancing technology or society then have fun being a luddite, but you're on the wrong side of history.
FWIW I work in applied research on LLMs.
> I work at a LLM startup (appeal to authority)
Has nothing to do with the points made lol you guys are desperate
Between that and the personally abusive emails you've been sending, it's clear that you don't want to use HN as intended, so I've banned the accounts.
Is any of it adding value though? I can see that AI has made it easier to do SEO spam and make an excuse for your lack of customer support, just like IVR systems before it. But I don't believe those added any real value (they may have generated profits for their makers, but I think that was a zero- or negative-sum trade). Put it this way: is AI being used to generate anything that people are actually happy to receive?
Okay, so some people are making money with it, but no true value was added, eh?
What comes to mind is - I consider myself an intelligent being capable of recognising my limits - but if you put my brain in a vat and taught me a new field of science, I could quite easily make claims about it that were completely incorrect if your teaching was incorrect because I have no actual real world experience to match it up to.
This is not the case with current model - they are forever stuck at junior level, and they won't improve no matter how much you correct them.
I know humans like that too. I don't ask them questions that I need good answers too.
This is conceptually essentially Moore’s law, but about every 5.5 months. That’s the only thing that matters at this stage.
I watched everyone make the same arguments about the general Internet, and then the Web, then mobile, then Bitcoin. It’s just a toy. It’s not that useful. Is this supposed to be the revolution? It uses too much power. It won’t scale. The technology is a dead end.
The general pattern of improvement to technology has been radically to the upside at an increasing pace parabolically for decades and there’s nothing indicating that this is a break in the pattern. In fact it’s setting up to be an order of magnitude greater impact than the Internet was. At a minimum, I don’t expect it to be smaller.
Looking at early telegraphs doesn’t predict the iPhone, etc.
Optimism is warranted here until it isn’t.
As soon as I saw integrated voice+text LLM demos, my first thought was that this was precisely the technology needed to make assistants like Siri not total garbage.
Sure, Apple's version 1.0 will have a lot of rough edges, but they'll be smoothed out.
In a few versions it'll be like something out of Star Trek.
"Computer, schedule an appointment with my Doctor. No, not that one, the other one... yeah... for the foot thing. Any time tomorrow. Oh thanks, I forgot about that, make that for 2pm."
Try that with Siri now.
In a few years, this will be how you talk to your phone.
Or... maybe next month. We're about to find out.
I could see something like Uber or Uber Eats trying to be early on something like this, since they already standardized the ordering for all the restaurants in their app. Scheduling systems are all over the place.
Granted, any third-party integrations would be a significant step up from my simple scenario of "voice and text comprehension" and local device state manipulation.
With some irony, the Hacker News user interface is essentially all just simple text.
A theme in current computer design seems to be: Assume the user doesn't use a text editor and, instead, needs an 'app' for every computer interaction. Like cars for people who can't drive, and a car app for each use of the car -- find a new BBQ restaurant, need a new car app.
Sorry, Silicon Valley, with text anyone who used a typewriter or pocket calculator can do more and have fewer apps and more flexibility, versatility, generality.
Show me an average person who has bought something real w bitcoin (who couldn’t have bought it for less complexity/transaction cost using a bank) and I’ll change my mind
That all seemed to go out the window when companies developed wallets to simplify the process for the average user, and when the prices surged, some started requiring account verification to tie it to a real identity. At that point, it’s just a bank with a currency that isn’t broadly accepted. The idea of digital cash was effectively dead, at least for the masses who aren’t going to take the time to figure out how to use Bitcoin without a 3rd party involved. Cash is simple.
I couldn’t fathom what use Bitcoin could possibly have, but completely overlooked the bad actors that would benefit.
The problem with this line of argument is that LLMs are not new technology, rather they are the latest evolution of statistical language modelling, a technology that we've had at least since Shannon's time [1]. We are way, way past the telegraph era, and well into the age of large telephony switches handling millions of calls a second.
Does that mean we've reached the end of the curve? Personally, I have no idea, but if you're going to argue we're at the beginning of things, that's just not right.
________________
[1] In "A Mathematical Theory of Communication", where he introduces what we today know as information theory, Shannon gives as an example of an application a process that generates a string of words in natural English according to the probability of the next letter in a word, or the next word in a sentence. See Section 3 "The Series of Approximations to English":
https://people.math.harvard.edu/~ctm/home/text/others/shanno...
Note: Published 1948.
You’re conveniently forgetting all the things that followed the same trajectory as LLMs and then died out.
Btc is currently like 60k... What does died out mean to you?
Additionally, there's a good reason we moved off deflationary hard currencies and onto inflationary fiat currencies. Bitcoin acts more like a commodity than a medium of exchange. People tend to buy it, hold it, and then eventually cash out. If I am given a bunch of bitcoin, the incentive is for me not to spend it, but rather keep it close and wait for it to appreciate — what good is a currency that people don't spend?
Also I find it weird when I read that due to its mathematically proven finite supply it is basic economics that gives it value. Value in modern economics is defined as what people are willing to give up to obtain that thing. Right now, people are willing to give up a lot for bitcoin, but mainly because other people are also willing to give up a lot for bitcoin, which gives it value.
It's a remarkable piece of engineering that has enabled this (solving the double spending problem especially), but it doesn't have inherent value in and if itself. There are many finite things in the world that are not valued as highly as bitcoin is. There's a finite number of beanie babies, a finite number is cassette tapes, a finite number of blockbuster coupons...
Gold is similar — should we all agree tomorrow that gold sucks and should never be regarded as a precious metal, then it won't lose its value completely (there's only a finite amount of it, and some people will still want it, e.g. for making connectors). But its current valuation is far higher than it would be for its scarcity alone — people mainly want gold, because other people want gold.
Well, they're not wrong. They are not that useful toys.
(Yes, the "Web" included.)
blockchain and no-trust decentralization has so much promise, but grifters all go for what got done first and can be squeezed money out of. same is happening with LLMs, as a lot of current AI work started with text first.
they might still lowkey be necessary evils because without them there would not have been so much money or attention flowing in this way.
I've been hearing this for the past 5 years, yet nothing of practical use based on blockchains has materialized yet.
your westernness is showing
go ask SA or Africa how useful it is that they arent restricted by insane dictatorial capital controls anymore
AI, given its requirements for large computation and money, and its ability to make easily available intelligence to certain groups, IMO has a real potential to do the exact opposite - take away power from individuals especially if they are middle class or below. In the wrong hands it can definitely destroy openness and freedom.
Even if it is "Open" AI, for most of society their ability to offer labor and intelligence/brain power is the only thing they can offer to gain wealth and sustenance - making it a commodity tilts the power scales. If it changes even a slice of what it is marketed at; there are real risks for current society. Even if it increases production of certain goods, it won't increase production of the goods the ultra wealthy tend to hold (physical capital, land, etc) making them as a proportion even more wealthy. This is especially true if AI doesn't end up working in the physical realm quick enough. The benefits seem more like novelties to most individuals that they could do without where to large corps and ultra wealthy individuals the the benefits IMO are much more obvious with AI (e.g. we finally don't need workers). Surveillance, control, persuasion, propaganda, mass uselessness of most of the population, medical advances for the ultra wealthy, weapons, etc can now be done at almost infinite scale and with great detail. If it ever gets to the point of obsoleting human intelligence would be a very interesting adjustment period for humanity.
The flaw isn't the technology; its the likely use of it by humans and their nature. Not saying LLMs are there yet or even if they are the architecture to do this, but agentic behaviour and running corporations (as OpenAI makes its goal on their presentation slides to be) seem to be a way to rid many of the need for other people in general (to help produce, manage, invent and control). That could be a good or bad thing, depending on how we manage it but one thing it wouldn't be would be simple.
Enough billions of dollars have been spent on LLMs that a reasonably good picture of what they can and can't do has emerged. They're really good at some things, terrible at others, and prone to doing something totally wrong some fraction of the time. That last limits their usefulness. They can't safely be in charge of anything important.
If someone doesn't soon figure out how to get a confidence metric out of an LLM, we're headed for another "AI Winter". Although at a much higher level than last time. It will still be a billion dollar industry, but not a trillion dollar one.
At some point, the market for LLM-generated blithering should be saturated. Somebody has to read the stuff. Although you can task another system to summarize and rank it. How much of "AI" is generating content to be read by Google's search engine? This may be a bigger energy drain than Bitcoin mining.