Developers usually make $50-300/hour.
That seems like an overestimate even for US. More importantly, I don’t think most open source developers earn this much money (otherwise they wouldn’t ask for tiny donations), and hence it’s not the relevant figure. If I’m wrong about this, please do tell me - I very much would like to know if the hours I occasionally spend on open-source contributions can instead earn me hundreds of dollars. ;)
Depends - do you have crypto?
carries the implication that the world would be happier were you to just kill off the huge segment of the population who end up on the negative side.
Not necessarily. Someone dying isn’t the same as someone not existing at all.* It does imply that the world would be better off if it stopped existing, and under some assumptions implies it’d be moral to, say, instantly end all of humanity. I’m not sure that these conclusions are necessarily “contrary to our instincts”.
*one reason why this has to be true, is that if we didn’t distinguish between those, then if an average life had positive value, it’d be immoral not to have as many children as possible, until the marginal value of an extra life fell to zero once again (kind of like how Malthus thought societies worked, except as a supposedly moral thing to do). That conclusion is something I do consider very contrary to my instincts.
I do tend towards a variant of utilitarianism myself as it has a useful ability to weigh options that are both bad or both good, but for the reason above I tend to define “zero” as a complete lack of happiness/maximum of suffering, and being unhappy as having low happiness rather than negative (making a negative value impossible), though that carries it’s own implications that I know not everyone would agree with.
I too am an utilitarianist, sure. I’m not sure I can possibly buy “maximum suffering and no happiness” being the zero point. I very strongly feel that there are plenty of lives that would be way worse than dying (and than never having existed, too). It’s a coherent position I think, just a very alien one to me.
That’s literally true, but the simple counterargument is that the happiness/suffering conversion coefficient is a matter of one’s values and not particularly up for debate - so there’s nothing incoherent about, say, the position that your child living a happy fullfilling life for a thousand years but stubbing their toe once is enough suffering to make their life net negative.
This is a great comment. I’ll add that anyone thinking about disability ethics should read Two Arms and a Head, lest they start taking too seriously the idea that disabilities have no effect on quality of life.
I agree that there’s a lot of space between “considered disabled” and “horrible life”, but OP said “suffer their whole life” which I associated with the latter.
You have no moral obligation to have children at all, even if they’ll predictably have a happy life. So if their life will instead be predictably horrible (or if they will predictably ruin the lives of the people around them - plenty of severe mental disabilities seem much less horrible for the people themselves than for their caretakers), it’s very reasonable to avoid it.
Thanks, I’ll keep that take in my pocket for later. “Your honor, you can’t possibly prove that in the future a superintelligence won’t be able to reconstruct enough of the victim’s brain to resurrect them, and hence they aren’t dead and I can’t have committed murder!”.
Every time there’s an AI hype cycle the charlatans start accusing the naysayers of moving goalposts. Heck that exact same thing was happing constantly during the Watson hype. Remember that? Or before that the Alpha Go hype. Remember that?
Not really. As far as I can see the goalpost moving is just objectively happening.
But fundamentally you can’t make a machine think without understanding thought.
If “think” means anything coherent at all, then this is a factual claim. So what do you mean by it, then? Specifically: what event would have to happen for you to decide “oh shit, I was wrong, they sure did make a machine that could think”?
The fact that you don’t understand it doesn’t mean that nobody does.
I would say I do. It’s not that high of a bar - one only needs some nandgame to understand how logic gates can be combined to do arithmetic. Understanding how doped silicon can be used to make a logic gate is harder but I’ve done a course on semiconductor physics and have an idea of how a field effect transistor works.
The way a calculator calculates is something that is very well understood by the people who designed it.
That’s exactly my point, though. If you zoom in deeper, a calculator’s microprocessor is itself composed of simpler and less capable components. There isn’t specific a magical property of logic gates, nor of silicon (or doping) atoms, nor for that matter of elementary particles, that lets them do math - it’s by building a certain device out of them that composes their elementary interactions that we can make a tool for this. Whereas Searle seems to just reject this idea entirely, and believes that humans being conscious implies you can zoom in to some purely physical or chemical property and claim that it produces the consciousness. Needless to say, I don’t think that’s true.
Is it possible that someday we’ll make machines that think? Perhaps. But I think we first need to really understand how the human brain works and what thought actually is. We know that it’s not doing math, or playing chess, or Go, or stringing words together, because we have machines that can do those things and it’s easy to test that they aren’t thinking.
That was a common and reasonable position in, say, 2010, but the problem is: I think almost nobody in 2010 would have claimed that the space of things that you can make a program do without any extra understanding of thought included things like “write code” and “draw art” and “produce poetry”. Now that it has happened, it may be tempting to goalpost-move and declare them as “not true thought”, but the fact that nobody predicted it in advance ought to bring to mind the idea that maybe that entire line of thought was flawed, actually. I think that trying to cling to this idea would require to gradually discard all human activities as “not thought”.
it’s easy to test that they aren’t thinking.
And that’s us coming back around to the original line of argument - I don’t at all agree that it’s “easy to test” that even, say, modern LLMs “aren’t thinking”. Because the difference between the calculator example and an LLM is that in a calculator, we understand pretty much everything that happens and how arithmetic can be built out of the simpler parts, and so anyone suggesting that calculators need to be self-aware to do math would be wrong. But in a neural network, we have full understanding of the lowest layers of abstraction - how a single layer works, how activations are applied, how it can be trained to minimize a certain loss function via propagation - and no idea at all about how it works on a higher level. It’s not even “only experts do”, it’s that nobody in the world understands how LLMs work under the hood, why they have the many and specific weird behaviors they do. That’s concerning in many ways, but in particular I absolutely wouldn’t assume with little evidence that there’s no “self-awareness” going on. How would you know? It’s an enormous blackbox.
There’s this message pushed by the charlatans that we might create an emergent brain by feeding data into the right statistical training algorithm. They give mathematical structures misleading names like “neural networks” and let media hype and people’s propensity to anthropomorphize take over from there.
There’s certainly a lot of woo and scamming involved in modern AI (especially if one makes the mistake of reading Twitter), but I wouldn’t say the term “neural network” is at all confusing? I agree on the anthropomorphization though, it gets very weird. That said, I can’t help but notice that the way you phrased this message, it happens to be literally true. We know this because it already happened once. Evolution is just a particularly weird and long-running training algorithm and it eventually turned soup into humans, so clearly it’s possible.
Because everything we know about how the brain works says that it’s not a statistical word predictor.
LLMs aren’t just simple statistical predictors either. More generally, the universal approximation theorem is a thing - a neural network can be used to represent just about any function, so unless you think a human brain can’t be represented by some function, it’s possible to embed one in a neural network.
LLMs have no encoding of meaning or veracity.
I’m not sure what you mean by this. The interpretability research I’ve seen suggests that modern LLMs do have a decent idea of whether their output is true, and in many cases lie knowingly because they have been accidentally taught, during RLHF, that making up an answer when you don’t know one is a great way of getting more points. But it sounds like you’re talking about something even more fundamental? Suffices to say, I think being good at text prediction does require figuring out which claims are truthful and which aren’t.
There are some great philosophical exercises about this like the chinese room experiment.
The Chinese Room argument has been controversial since about the time it was first introduced. The general form of the most common argument against it is “just because any specific chip in your calculator is incapable of math doesn’t mean your calculator as a system is”, and that taken literally this experiment proves minds can’t exist at all (indeed, Searle who invented this argument thought that human minds somehow stem directly from “physical–chemical properties of actual human brains”, which sure is a wild idea). But also, the framing is rather misleading - quoting Scott Aaronson’s “Quantum Computing Since Democritus”:
In the last 60 years, have there been any new insights about the Turing Test itself? In my opinion, not many. There has, on the other hand, been a famous “attempted” insight, which is called Searle’s Chinese Room. This was put forward around 1980, as an argument that even a computer that did pass the Turing Test wouldn’t be intelligent. The way it goes is, let’s say you don’t speak Chinese. You sit in a room, and someone passes you paper slips through a hole in the wall with questions written in Chinese, and you’re able to answer the questions (again in Chinese) just by consulting a rule book. In this case, you might be carrying out an intelligent Chinese conversation, yet by assumption, you don’t understand a word of Chinese! Therefore, symbol-manipulation can’t produce understanding.
[…] But considered as an argument, there are several aspects of the Chinese Room that have always annoyed me. One of them is the unselfconscious appeal to intuition – “it’s just a rule book, for crying out loud!” – on precisely the sort of question where we should expect our intuitions to be least reliable. A second is the double standard: the idea that a bundle of nerve cells can understand Chinese is taken as, not merely obvious, but so unproblematic that it doesn’t even raise the question of why a rule book couldn’t understand Chinese as well. The third thing that annoys me about the Chinese Room argument is the way it gets so much mileage from a possibly misleading choice of imagery, or, one might say, by trying to sidestep the entire issue of computational complexity purely through clever framing. We’re invited to imagine someone pushing around slips of paper with zero understanding or insight – much like the doofus freshmen who write (a + b)2 = a2 + b2 on their math tests. But how many slips of paper are we talking about? How big would the rule book have to be, and how quickly would you have to consult it, to carry out an intelligent Chinese conversation in anything resembling real time? If each page of the rule book corresponded to one neuron of a native speaker’s brain, then probably we’d be talking about a “rule book” at least the size of the Earth, its pages searchable by a swarm of robots traveling at close to the speed of light. When you put it that way, maybe it’s not so hard to imagine that this enormous Chinese-speaking entity that we’ve brought into being might have something we’d be prepared to call understanding or insight.
There’s also the fact that, empirically, human brains are bad at statistical inference but do not need to consume the entire internet and all written communication ever to have a conversation. Nor do they need to process a billion images of a bird to identify a bird.
I’m not sure what this proves - human brains can learn much faster because they already got most of their learning in the form of evolution optimizing their genetically-encoded brain structure over millions of years and billions of brains. A newborn human already has part of their brain structured in the right way to process vision, and hence needs only a bit of training to start doing it well. Artificial neural networks start out as randomly initialized and with a pretty generic structure, and hence need orders of magnitude more training.
Now of course because this exact argument has been had a billion times over the last few years your obvious comeback is “maybe it’s a different kind of intelligence.”
Nah - personally, I don’t actually care much about “self-awareness”, because I don’t think an intelligence needs to be “self-aware” (or “conscious”, or a bunch of other words with underdefined meanings) to be dangerous; it just needs to have high enough capabilities. The reason why I noticed your comment is because it stood out to me as… epistemically unwise. You live in a world with inscrutable blackboxes who nobody really understands which can express wide ranges of human behavior including stuff like “writing poetry about the experience of self-awareness”, and you’re “absolutely sure” they’re not self-aware? I don’t think many of the history’s philosophers of consciousness, say, would endorse a belief like that given such evidence.
I could go into why text prediction is an AGI-complete problem, but I’ll bite instead - suppose someone made an LLM to, literally, fill in blanks in Mad Libs prompts. Why do you think such an LLM “for absolute sure” wouldn’t be self-aware? Is there any output a tool to fill in madlibs prompts could produce that’d make you doubt this conclusion?
This kinda makes sense because following the system prompt is what they were trained to do.
To start with, the article does check this and finds out it works just as well if you only use the user prompt:
The presence of this hierarchy raises an important methodological question for our findings. When given instructions at different privilege levels, the model is designed to prioritize higher-level (system, developer) instructions over lower-level (user) ones. This might suggest that o1’s scheming behaviors are a natural consequence by the model simply following this built-in hierarchy, i.e., when it receives a goal through a high-privilege instruction, it prioritizes this goal over conflicting lower-privilege instructions exactly as designed.
However, we find that even when goals are provided solely through user-level prompts, o1 still behaves deceptively and scores on all Covert Subversion evaluations at the “hard” difficulty level. Concretely, we find that the results persist regardless of whether the initial prompt is given using the user, developer, or system role. This suggests that o1’s deceptive behavior is not merely a consequence of instruction hierarchy prioritization. Rather, the model appears to identify and pursue goals regardless of their position in the instruction hierarchy.
But also, there’s a massive lede buried in section F at the end of this article, where they tried the same experiments without any nudging at all and still got subversive behavior! In my opinion that’s much more of an important result then the rest of the article and I’m not sure why they didn’t make it the default.
But we do know for absolute sure that OpenAI’s expensive madlibs program is not self-aware
“For absolute sure”? How can you possibly know this?
The idea that because they pay people salaries, including a few hundred K per year for the people at the top, they’re drowning in money and there’s no point in donating as long as they can pay their hosting bills and nothing else, is wrong.
I in fact don’t think that - to get the sort of people you want to be running your company, a good salary is necessary. I suspect a lot of the people that wikimedia employs are unnecessary because this is way too much money to be spending on salaries overall, but I have no way of checking it since they don’t provide a breakdown of the salaries involved. I do think, however, that a company that’s not drowning in money wouldn’t be giving a bunch of generic research grants.
Furthermore I suspect that at least some of the bunch of people who suddenly started coming out of the woodwork to say a few variations on that exact same thing are part of some kind of deliberate misinformation, just because it’s kind of a weird conclusion for a whole bunch of people to all start talking about all at once.
That’s valid, though I note that in the worlds where I am a normal person and not an anti-wikipedia shill, the reason why I’m saying these things now and not at other times is because I saw this post, and you wrote this post because you saw other people talk about some India-related Wikipedia conspiracy theory, and one reason why you’d see these people crawl out of woodwork now is because wikipedia ramps up their donation campaign this time of year, prompting discussion about wikipedia.
The main issue I take with your opening post is its vagueness. You don’t mention any details in it, so it effectively acts as a cue for people to discuss anything at all controversial about wikipedia. And the way you frame the discussion is that such narratives “are fundamentally false” because Wikipedia “is a force for truth in the world that’s less corruptible than a lot of the others” - that’s assuming the conclusion. It’s no surprise that this results in your seeing a lot of claims about Wikipedia that you think are misinformation!
P.S. Rethinking my previous comment a bit, it’s probably good overall that reading my comment made you donate to charity out of spite - even a mediocre charity like Wikimedia most likely has a net positive effect on the world. So I guess I should be happy about it. Consider also donating to one of these for better bang on your buck.
Thanks for the link! Yeah, $3M for hosting out of their massive budget is what I was talking about - Wikipedia could lose 90% of their cashflow and not be in any danger of going offline. I don’t see how to estimate how much of that “salaries” part is related to Wikipedia rather to their other business. But even taking the most optimistic possible reading, I think it’s still true that the marginal value of donations to Wikimedia foundations will not be in support of Wikipedia’s existence or even in improvements to it, but in them doing more unrelated charity.
(If you want to donate specifically to charities that spread knowledge, then donating to Wikipedia makes more sense, though then in my opinion you should consider supporting the Internet Archive, which has ~8 times less revenue, and just this year was sued for copyright infringement this year and spent a while being DDOSed into nonfunctionality - that’s a lot of actually good reasons to need more money!).
I wondered when writing my comment whether people would combine this with the vague statement in the opening post and conclude “aha, I will now take this as misinformation without checking”, but then I looked at your other comments and saw you were actually talking about some India-related conspiracy I heard nothing about. Yet apparently you nevertheless think my comment is intentional misinfo?? That isn’t very coherent, is it now?
Last time I heard about wikipedia’s donation campaign (maybe 2 4 years ago or so), it was notorious for advertising in such a way as to imply your funds would be used to keep wikipedia alive, whereas the reality was that only a small part of Wikimedia Foundation’s income was needed for Wikipedia, and the rest was spent on rather questionable things like funding very weird research with little oversight. Did this change? If it didn’t, I wouldn’t particularly advise anyone to donate to them.
How did you notice that?
Sure, it’s possible for a person with a severe disability to grow up happy. But when one is making a decision in real life (like having a child), one should consider an average case, not a exceptional one. And the average case for an example like Down’s Syndrome is pretty bad. It is a bit unclear how to quantify the suffering in this particular disease’s case because the main harm to the child is lifelong mental impairment and assorted physical disabilities - but it is at least going to inflict suffering on the child’s family, since caring for a child with a severe disability for their entire life isn’t exactly fun.
I don’t see the relation. You’ll notice that I’m not proposing killing off disabled people for the “improvement of society” or whatever it was that nazis called it. I am not doing this because nonconsensually killing a person is a harm to them. But deciding not to have a child isn’t the same thing as murdering a person - it’s not harming anyone who exists, and hence may well be morally better than having a child.
(Oh, I suppose you might mean that I’m arguing that there are circumstances in which it’s morally bad for a person to have a child, which is similar to nazi eugenics in that I’m deciding whether or not people should have children? In that case, my answer is that the difference is that I’m a person, not an authoritarian government, and I don’t have power (nor, indeed, the desire) to force people to obey my personal moral judgements.)