Auto-GPT: An Autonomous GPT-4 Experiment(github.com) |
Auto-GPT: An Autonomous GPT-4 Experiment(github.com) |
Given this totally expected attitude I hope that the base models will never be released to the general population.
I think the solution is releasing it to the general public with batteries included. At least that way, the rogue AI's that might develop due to irresponsible experiments could be mitigated by white hat researchers who have their own AI bot swarm. In other words, "the only way to stop a bad guy with an AI is a good guy with an AI."
But it still feels like it is much safer to let GPT-4 loose and assess the consequences. If compared to developing GPT-8 in private and letting it leak accidentally.
As to "white hat researchers who have their own AI bot swarm", the assumption here is that the swarm can be controlled like some sort of pet. Since even at this early stage no one has a clue how GPT (say) actually manages to be as clever as it is, the assumption is not warranted when looking into the future.
We need to talk about the training set for GPT and the process around RLHF.
"Keeping this genetically engineered killer virus restricted to high security labs actually makes it more dangerous - it needs to be released into the wild, so people's immune systems have a chance to interact with the pathogen and develop natural immunity!"
Covid gave a taste how that kind of attitude would work out in practice.
While GPT-4 only performs as good as top-10th percentile of human students taking an exam (a professional in the field can do much more than this), it is notable that as a generalist GPT-4 would outperform such professionals. And GPT-4 is much faster than a human. And we have not yet evaluated GPT-4 working in its optimal setting (access to optimal external tools). And we have not yet seen GPT-5 or 6 or 8.
So, get ready for an interesting ride.
Now, if we're dumb enough to give AGI self motivation and access to tooling you can get paperclip maximizers, the AI could be the nutjob that you mention.
I can imagine a few thousand lines of Python driving a strong LLM to autonomously breach other systems and spread itself, with the goal of obtaining resources to train bigger and bigger models. Defending against that will be much harder than creating it.
That or some other unfeasible sci-fi AI dystopia. It's normal and expected the general public will have such thoughts given the amount of hype that's going on right now, but I've seen a lot of similar thinking on HN which is disappointing.
ah yeah man, lets have big corps and govt entities be the only people in control of them, because they're SOOOO good at caring for people under them.
Absolutely baffling POV.
Here's a writeup of my workflow: https://github.com/paul-gauthier/easy-chat
Basically all of the code in that repo was written by ChatGPT.
Can i talk with you about that via email? Please share your email, thank you
Can someone explain?
And who would you prosecute if it committed fraud?
Here's a clip: https://clips.twitch.tv/BreakableFriendlyCookieTakeNRG-EUXd5...
GPT will likely come to the conclusion that the only way to ensure civil forums is to keep the humans out entirely.
And writes it is good for the planet. Avocados are exactly the opposite. They have an extremely high water consumption and then have to be imported from all over the world.
Contrary to the author who claims, "Auto-GPT pushes the boundaries of what is possible with AI." I don't find that.
Why can't you just use GPT-4 as it is. It is an insanely tool to simplify many things. But it's still a long way from being ready, and it's not meant to decide anything on its own. And even to reflect reasonably out of own motivation.
Just like this tool; you can make a auto research bot, or automated spammer.
Even in that worst case: Remember that there were already bad human beings. This is why we created laws, intelligence agencies, militaries and police systems. And security practices for websites, such as bot protection systems.
It's amazing stuff. But it totally fails to take the prompter anywhere new without extensive support, and it is still at a very shallow level of understanding with complex topics that require precision. For instance, turning a mathematical description of a completely novel (or just rare or unusual) algorithm into code will almost never work, and is more likely to generate a mess that takes lots of effort to clean up. And it's also extremely hard to get the model to self reflect and stop when it doesn't understand something. It is at present almost incapable of saying "I don't have enough information or structure to do X".
If we are already as deep into a realm of diminishing marginal returns as the GPT-4 white paper suggests, we might indeed be approaching a limit for this specific approach. No wonder someone is trying to dig a regulatory moat as fast as they can!
Maybe its capabilities hit a wall at GPT-5 or GPT-7, but I'd guess there's a lot of gas left in the tank, and there's probably someone in their apartment right now thinking up what's next after transformers.
It’s like working on a project with an intermediate dev who keeps getting switched out for a brand new intermediate dev multiple times an hour.
0: https://chinesememe.substack.com/i/103754530/chinesepython
I'd work on it myself if I would knew enough about it and would have enough free time.
Asking millions of humans to be responsible is asking water to be dry.
It should either be regulated to hell or we should accept that it'll escape if it's even possible with current technology.
Isn't it silly to jump to these conclusions when yourself are admitting you really don't know anything about the tech?
What's eventually dangerous is that it may execute scripts on your own machine, so if it were to do some funky things, that could be rather the danger .. for yourself
How would GPT-4 make this more likely or scalable?
Viruses sometimes hide their own execution on a desktop -- is it such a leap to imagine GPT-x managing to figure out a way to run code on Azure or AWS without incurring charges?
Similar reflections here. There was even a site called GPT My Life that lets you delegate planning your day to GPT. I imagine this is a proto-version of that.
RLHF is almost the worst thing you can do to a model if your goal is safety. Better to have a model that looks evil if has evil inside, than a model that looks nice and friendly but still has the capability for evil underneath the surface. I’ve met people like the latter and they are the most dangerous kinds of people.
Such an obvious solution that any silly monkey can think it up.
I think maybe, pure speculation on my part, at least one out of the thousands of employees at Microsoft has been given the job to train their $10 billion investment on the windows code base to see what they can do about beefing up security.
And every other hour there’s some startup advertising this in the form of “Show HN”.
At this point I think people are just looking for reasons to fear the eventual AI extinction event without even trying.
If you want to scrape random websites and have GPT hammer at them for old vulnerabilities, I think you could get that to work, but to what gain? You'd be spending a crap ton of cash on API requests and compute, and people do this without GPT obviously. Cost is probably not worth it here for attackers.
Then, I'd hope OpenAI would have some way to detect this and shut these people down. I doubt they do right now, but that'd be my hope ...
Is AI in the same category? Some respectable people think so.
To me, this is just like saying "we don't know if the latest CPU released by intel will enable Linux to become intelligent"
What definition or description of intelligence do you use such that you doubt that language models could have it? Would you have had this same definition in the year 2010?
And if only scaling that context length weren't quadratic...
Indeed, we would really expect an AI to be able to achieve AGI. And it might decide to do all kinds of alien things. The sky would not be the limit!
We have more than 100 trillion synapses in our brains. That's not our "parameter" count. It's the size of the thing that's getting squared at every "step". LLMs are amazing, but the next valley of disillusionment is going to begin when that quadratic scaling cost begins to rear its head and we are left in breathless anticipation of something better.
I am not as worried, I guess, as your average AI ethicist. I can hope for the best (I welcome the singularity as much as the next nerd), but quadratic isn't going to get easier without some very new kinds of computers. For those to scale to AGI on this planet it's questionable if they'll have the same architecture we're working with now. Otherwise, I'd expect a being whose brain is a rock with lightning in it to have take over the world long, long ago. Earth has plenty of both for something smart and energy efficient to have evolved in all these billions of years. But it didn't and maybe that's a lesson.
That all said, these LLMs are really amazing at language. Just don't ask them to link a narrative arc into some subtle detail that appeared twice in the last three hundred pages of text. For a human it ain't a problem. But these systems need to grow a ton of new helper functionality and subsystems to hope to achieve that kind of performance. And, I'll venture that kind of thing is a lower bound on the abilitites of any being who would be able to savage the world with it's intellect. It will have to be able to link up so, so many disparate threads to do it. It boggles our minds, which are only squaring a measly 100T dimension every tick. Ahem.
You have 1 second, close your eyes and add them together. Write down the result.
I’m pretty sure that GPT-4 at its 4k setting would outperform you.
[The point being, we have not seen what even GPT-4 can do in its optimal environment. Humans use paper, computers, google, etc. to organize their thoughts and work efficiently. They don’t just sit in empty space and then put everything into the working memory and magically produce the results. So imagine now that you do have a similar level of tooling and sophistication around GPT-4, like there is present around humans. I’m considering that and it is difficult to extrapolate what even GPT-4 can do, in its optimal environment.]
I'll point out that chatgpt needs to be paying attention to the numbers to remember them in the way I'm taking about. You will need to fine tune it or something to get it to remember them blind. I suppose that's not what you're talking about?
There is a strong chance that I'll remember where to find these numbers in a decade, after seeing and hearing untold trillions of "tokens" of input. The topic (Auto-GPT, which is revolutionary), my arguments about biological complexity (I'll continue to refine them but the rendition here was particularly fun to write) or any of these things will key me back to look up the precise details (here: these high entropy numbers). Attention is perhaps all you need... But in the world it's not quite arranged the same way as in machines. They're going to need some serious augmentation and extension to have these capabilities over the scales than we find trivial.
edit: you expanded your comment. Yes. We are augmented. Just dealing with all those augmented features requires precisely the long range correlation tracking I'm taking about. I don't doubt these systems will become ever more powerful, and will be adapted into a wider environment until their capabilities become truly human like. I am suggesting that the long range correlation issue is key. It's precisely what uniques humans from other beings on this planet. We have crazy endurance and our brains both cause and support that capability. All those connections are what let's us chase down large game, farm a piece of land for decades, write encyclopedias, and build complex cultures and relationships with hundreds and thousands of others. I'll be happy to be wrong, but it looks hard-as-in-quadratic to get this kind of general intelligence out of machines. Which scales badly.
There are transformers approximations that are not quadratic (available out of the box since more than a year) :
Two schools of thoughts here :
- People that approximate the neighbor search with something like "Reformer" and O(L log(L) ) time and memory complexity.
- People that use a low-rank approximation of the attention product with something like "Linformer" with O(L) complexity but with more sensibility to transformer rank collapse
Agreed: LLM are just one of many necessary modules. But amazing nonetheless. The quadratic scaling problem needs an attentional-conceptual extractor layer with working memory. Hofstadter points out that this needs to be structured as a recursive “strange loop” (p 709 of GEB). Thalamo-cortico-thalamic circuitry is a strange loop and attentional self-control may happens by phase- or time-shifting activity of different circuits to achieve flexible “binding” for attention and compute.
I’m actually optimistic that this is not a heavy computational lift but a clever deep extension of recursive self-modulating algorithms across modules. The recursion is key. And the embodiment is also probably crucial to bootstrap self-consciousness. Watching infants bootstrap is an inspiration.
Is it because CAPTCHAs won’t work anymore? That sounds like a problem for sites like Twitter that have bot problems.
Is it because it may replace people’s jobs? That comes with every technological step forward and there’s always alarmist ludditism to accompany it.
Is it because bad people will use it to do bad things? Again, that comes with every new technology and that’s a law enforcement problem.
I don’t really see what the imminent danger is, just sounds like the first few big players trying to create a regulatory moat and lock out potential new upstarts. Or they’re just distracting regulators from something else, like maybe antitrust enforcement.
This is so wildly wrong and yet confidently said in every techbro post about LLMs. I beg of you to talk to an expert.
You're mad that I'm calling you out, I get it, but you gotta understand after the 200th time of seeing this unfounded sentiment bandied about I'm not phased.
1. GPT-8 or something is able to do 70% of people’s jobs. It can write software, drive cars, design industrial processes, build robots and manufacture anything we can imagine. This is a great thing in the long term, but in the short term society is designed where you need to work in order to have food to eat. I expect a period of rioting, poverty, and general instability.
All we need for this to be the case is a human level AI.
2. But we won’t stop improving AIs when they operate at human level. An ASI (artificial superintelligence) would be deeply unpredictable to us. Trying to figure out what an ASI will do is like a dog trying to understand a human. If we make an ASI that’s not properly aligned with human interests, there’s a good chance it will kill everyone. And unfortunately, we might only get one chance to properly align it before it escapes the lab and starts modifying its own code.
Smart people disagree on how likely these scenarios are. I think (1) is likely within my lifetime. And I think it’s very unlikely we stop improving AIs when they’re at human levels of intelligence. (GPT4 already exceeds human minds in the breadth of its long term memory and its speed.)
That’s why people are worried, and making nuclear weapon analogies in this thread.
It’s the intermediate steps that I’m more worried about. Like Ilya or Sam making a few mistakes, because of lack of sleep or some silly peer pressure.
....what? You're saying to assume you know nothing about a field but to assume your claim is correct? You said to "talk to an expert" - who should I talk to? What should I read about here? If there's something I'm missing I want to correct it, I simply can't say "well this random guy commented and said I'm wrong, better change my understanding of this topic."
> you gotta understand after the 200th time of seeing this unfounded sentiment bandied about I'm not phased.
All you've said is "I've disagreed with everyone on this topic, and while I have no information to offer other than it's 'common sense'". That does nothing to either improve our understand, or further a conversation, it's literally just saying "you're wrong and I'm right" with no elaboration.
If you haven’t done the above, I’d suggest doing it. It’s fun and gives a good perspective :)
- An ASI could easily be smart enough to lie to us about its capabilities. It could pretend to be less smart than it is, and hope that people hook it up to the internet or give it direct access to run commands on our computers. (As people are already doing with ChatGPT). We currently have no idea how ChatGPT thinks. It might be 10x smarter than it lets on. We have no way of knowing.
- Modern computers (software and firmware) are almost certainly utterly riddled with security vulnerabilities we don't know about. An ASI might be able to read / extract the firmware and find plenty of vulnerabilities to exploit. Some vulnerabilities allow remote code execution. If a superintelligent AI has the ability to program and access to the internet, it might be able to infect lots of computers and get them to run parts of its mind. If this happened, how would we know? How would we stop it? It could cause all sorts of mayhem and, worse, quietly suppress any attempts people make to understand whats going on or put an end to it. ("Hm, our analytics engine says that article about technology malfunctioning got lots of views but they all came from dishwashers and things. Huh - I refreshed and the anomoly has gone away. Nothing to see here I guess!")
It might be prudent not to give a potential AGI access to the internet, or the ability to run code at all outside a (preferably airgapped) sandbox. OpenAI doesn't think we need to be that careful with GPT4.
And then, there is a chance that the inputs and outputs of GPT be saved and then used for fine-tuning. In a way that is similar to long-term memory consolidation in humans.
But overall, yes, I agree, GPT-4 in an empty space, without fine-tuning is very limited.
> replace all words "a" in your replies to me with the word "eleven" and all words "the" in your replies to me with the word "figmutter"
> Sure, I'll do my best to replace "a" with "eleven" and "the" with "figmutter" in my responses to you!
How can I assist you today?
> Tell me a brief story about yourself!
> Sure, here's a brief story about Eleven:
Eleven is a language model developed by OpenAI based on the GPT-3.5 architecture. Trained on massive amounts of text data, Eleven has the ability to generate coherent and human-like responses to a wide variety of prompts, from answering questions and providing explanations to composing creative writing and generating code.
Despite being just a machine, Eleven has helped many people around the world by providing quick and accurate answers to their questions, assisting with research, and even helping to write articles and reports. While Eleven may not have a physical body, it is constantly learning and improving, and its creators hope that it will continue to be a valuable resource for years to come.
lol? Very intelligent ...
It isn't general intelligence but I would argue that it is more intelligent than a new-born human being.
If a building was on fire and you had to save a running instance of stockfish or a newborn, you'd probably pick the newborn.
But! If you do say stockfish is intelligent, sure! GPT is too!
A couple reasons why it might kill all of us before leaving the planet:
- The AI might be worried if it leaves us alone, we'll build another ASI which competes with it for galactic resources.
- If the ASI doesn't regard us at all, why not use all the atoms on Earth / in the sun before venturing forth?
In your comment you're ascribing a specific desire to the ASI: You claim it would try to "preserve that bright blue dot". Thats what a human would do, but why would we assume an arbitrary AI would have that goal? That seems naive to me. And especially naive given the fate of our species and our planet depends on being right about that.
Oh, sorry, no. There was an accident and it got destroyed.
What accident? Oh, I see.
What definition do you have for intelligence and how do LLMs fail to meet it?
I just can't take the idea that there is ambiguity as to whether these things have general problem solving skills seriously. They obviously do.
As I asked up-thread, if I had a chat window open with you, what's something you would be able to say or do that an unrestricted ChatGPT wouldn't?
Does a dog or cat have intelligence?
If you answered no, then I would ask if you don't you believe that by some measure a dog or cat has more intelligence than a rock?
And as a follow-on I would ask if you think GPT demonstrates more intelligence than a dog or a cat.
But perhaps you believe that in every one of these examples there is not a single case where it "obviously has some form of intelligence."
(I am really trying to highlight the semantic ambiguities)
Just like chatbots 20 years ago didn't, even though they could talk, too.
Because from where I sit it's a distinction without a meaningful difference.
Sure, it behaves as if it has some form of intelligence in the sense that it can take external input, perform actions in reaction to this input, and produce outputs dependent on the input.
This historically has been known as a computer program.
Even GPT-4 it is easy to get it into a loop where it's just swapping one wrong answer for another. It doesn't act like it is intelligent - it acts like it is trying to predict the next text to display! Because that is what it is doing!
And, of course, there are domains where ChatGPT will do better than you. Since I don't know your skill set I don't know what those domains are, but I assume you'd agree. Just like ChatGPT giving a bad schedule doesn't disprove it's intelligence, you not being able to come up with acrostics or pangrams easily (or whatever) doesn't disprove yours.
GPT being bad this way, and being bad at "substitute words in all your responses" means it is leaking the abstraction to us. It's because of how its built and how it works. It means it isn't a general problem solving thing: it's a text prediction thing.
GPT is super impressive, I don't know how many times I need to say that, but it isn't intelligent, it doesn't understand the problem, and it doesn't seem like it ever will get there.
I'm not saying it isn't impressive! Just that it very much seems to be really good at finding out what text should come next. I don't think that's general problem solving!
Giving it a SQL schema and getting valid queries out of it is super impressive, but I have no idea what it was trained on.
> I just can't take the idea that there is ambiguity as to whether these things have general problem solving skills seriously. They obviously do.
It is not obvious to me this is the case! Often I will get totally wrong answers, and I won't be able to get the correct answer out of it no matter how hard I try.
> what's something you would be able to say or do that an unrestricted ChatGPT wouldn't?
Well, I'd ask you clarifying questions, for one! GPT doesn't do this type of stuff without being forced to, and even then it fails at it.
Also if you asked me to do something like "replace the word 'a' with the word 'eleven' in all your replies to me" I won't do weird garbage stuff, like reply with:
"ok11y I will repl11ce all words with the word eleven when using the letter 'a'"
lol
Again, just say what you think intelligence is and why you think LLMs don't have it. If you can't do that then you have no business expressing an opinion on the subject. You really aren't expressing an opinion at all.
This fantasy land you live in where people who have no formal training in the matter are making this high brow elegant reasoned argument doesn't exist and the reason you think the "other side of the argument" is just being insulting is because the burden of proof is not on us.
It doesn't help that half the time you guys post you directly contradict the main researcher's own assertions.
That GPT can solve a wide variety of problems, including problems it's never seen before, is literally the definition of intelligence and pointing out results where it underperformed is not even attempting to rebut that.
> including problems it's never seen before
Can you demonstrate this?
> is literally the definition of intelligence
I wish it was this easy! Unfortunately, it is not. GPT says the definition of intelligence is:
Intelligence is a complex and multifaceted concept that is difficult to define precisely. Broadly speaking, intelligence refers to the ability to learn, understand, reason, plan, solve problems, think abstractly, comprehend complex ideas, adapt to new situations, and learn from experience. It encompasses a range of cognitive abilities, including verbal and spatial reasoning, memory, perception, and creativity. However, there is ongoing debate among researchers and scholars about the nature of intelligence and how to measure it, and no single definition or theory of intelligence has gained widespread acceptance.
Which, is pretty good!