Physicist and dabbler in wiring fantasy/science fiction.
Ben
I am reminded of a great section in “The open Society and its enemies”. Most of this book is a criticism of Plato’s “fascist” (an anachronistic label) politics. Plato defines “justice” as something like “that which is good for society, IE good for the city.”
It is intriguing how strongly Plato’s philosophy reckoned that the city should come before the individual, to the point of advocating a society where literally everyone was a selectively bred cog in the nation’s machinery and valued exactly according to their effectiveness. The question “but what is the point of an efficiently run city if everyone is miserable” doesn’t seem to even occur to Plato. Its like it never crossed his mind that utility was a thing to pursue on the macroscale. The connection is that this post (I think) argues that abstracting the people into “society” too strongly might lead to accidentally “Plato”-ing, that is lead to policies which are net-negative for people but appear net-positive when the people are abstracted into a homogenous whole.
Its not about the objective randomness, to me its about the fact that the frequency is by necessity hypothetical. Yes their will only be one tomorrow, and rain might be pre-determined. But we can make arguments about “in a sample of days like tomorrow we know that some % will see rain.” But their can only be one universe, even in principle, so the idea of generalizing to a class of universes and taking our universe as a member of that class I think can cause problems.
I always find anthropic arguments a bit mysterious. I completely agree that if all people live in one of two houses, one with 1 resident and the other with 100, then if we are a randomly chosen person we will most-likely have a lot of people sharing our home.
But when you substitute the word “house” for “universe” something important is different. Only one of the two universes actually exists. The other one is hypothetical. Its in fact more hypothetical than a normal “could have been”, because we don’t even know if chemistry/star formation or whatever could have been different, even in principle. So when we count frequencies to estimate the probability in the “universe” case we are counting people who are by-definition hypothetical.
In the situation where God flicks a coin, and on heads creates a house with 100 people, and on tails creates a house with 1 person. Then I am not sure either way about an anthropic argument that says “I have just found out I exist, which means I am very likely in the crowded house timeline”. But I think that without the guarantee that such a coin toss ever happened, that is without any reason to think the 100 person house was ever a possibility—then my own feeling is that anthropic arguments don’t work. Although I find them enough of a logical minefield that I would not put my certainty super high.
You are right, probably closer to simulacrum. Although the two do overlap.
I find it hard to imagine that you are right when you imply that politics hasn’t got anything to do with policy. ”...has anything at all to do...”
To take a recent US example, pulling out of Afghanistan. No matter how much we focus on the Meta level politics, (eg. “Trump started it but Biden finished it, so who gets the blame?”, “Will it embolden the US’s enemies my creating an image of weakness?”) it is still clear that an actual physical thing really happened, and that it would not have happened had the political landscape looked different in some key ways.
Here in the UK we left the European Union. That was discussed and voted on primarily through the lens of political issues (Typical conversation starters might have been: “will it help Boris unseat Cameron so he can launch his own leadership bid?”, “are we the kind of country that is welcoming to foreigners?”, “But surely we identify collectively much more on the level of a nation state than some kind of united states of Europe?”). But it happened, we have legally left. And the practicalities are slowly fitting into place.
Politics is not the same as policy, there are steps in between. But the suggestion that they have literally zero causal connection seems to be an obvious nonsense. I think this illusion possibly forms when the actual policy never touches you. The median voter is not a migrant, so migration policy feels fictional, only read about. The median voter is not a soldier (or a foreigner), so foreign wars also feel abstract. Most people have not been accused of a crime, so criminal policy again follows the pattern. Same with so many other things. When the COVID lockdown kicked in and for the first time in my life I actually felt a policy change “hit me” it was a bit of a shock. So, in typical times, it feels like politics doesn’t effect policy—because almost all policy is targeted at a minority that probably doesn’t include you.
Very interesting. The lateness of knitting is so strange.
To throw my theory of what they are “actually” for on the heap: gambling.
My first thought was “weird dice”, but wikipedia says they never have inscriptions or numbers. But it also says they are often found with coins and the holes are all different sizes. So you could imagine a gambling game along the following lines: I roll the thing, and the size of the circle I roll tells me the size of the coin you owe me. Then its your turn to roll for the coin I owe you. (I would hope this would just be a component of a more interesting game.)
The “levels of mazedom” is an interesting insight. The cleanest examples I can see are in politics:
Level 1: “Policy X is good. I will try and persuade people to implement policy X, or get elected to do it.”
Level 2: “I want to get elected. Policy Y is seen as good. I will advocate policy Y and implement it.”
Level 3: “I want to get elected. My supporters don’t actually think that policy Z is good or even possible, but advocating for it sends a signal that will help me win.” [eg. ‘… and Mexico will pay for it.’]
Level 4: “I want to get elected. I will set a theme—a feeling for my campaign. I will avoid specifics of actual policies to avoid alienating people who might disagree with them.”
Yes, you could. That is a clever idea. Weirdly the equivalent thought had occurred to me for murder (redact them back into some kind of pre-fetus or something) but not for this much cleaner-cut case with a hard drive.
Oh, thanks for trying it. Its a good effort in a way, it is definitely it trying to put the table in the footnote. So has the right target. Unfortunately from my test it doesn’t seem to actually work—no error message, but the table doesn’t render. It is so much less frustrating to see it fail at the right goal than provide a lot of information about a distinct goal that happens to share the same keywords.
Reading the article, and painting it with my own experiences with Google failing me, it looks like Google is best when you want to understand a specific, named (possibly niche) object, but that chat GPT potentially has an advantage when you want to understand two things with a specific relationship, or if you can’t name the specific object you want but can only describe it in relation to other objects.
An example from my own life. I few years ago I wanted to know if it was possible to put a latex table in a footnote. I still don’t know if it is possible, because Google buried me in websties telling me how to attach a footnote to a table—a completely different formatting problem that shares the same two key words. I suspect chat GPT would be good for this.
I may be interpreting Zvi’s GPT scale incorrectly, but I think it is mostly a measure of novelty and not quality. Higher GPTs will (presumably) take more and more inferential steps away from the training data and have more ability to seem novel. So a GPT that does a specific comedians style exactly, and is hilariously funny, is probably a lower hanging fruit than one that invents its own comedic style that is meh-to-okay.
So, if I understand it right, the 6′s are the comedians who are very weird. For example, back when I was a teenager the UK TV/radio comedy scene I would say Milton Jones would have a very high GPT value. This is because every other comedian told long interconnected stories with all the jokes woven together, while pacing around the stage and trying to interact with the audience. Typical content would be fairly dense in references to sex and overall be a bit rude and dirty. The stories were all calibrated to seem kind-of-beleiveable, like the story would start with the comedian talking about how badly their first gig went and then become harder to credit as it continued. Milton Jones just stood up straight and machine gunned one liners. The topics were always clean, nothing dirty, racial or rude. Nothing remotely intended for you to think was real/true. He wasn’t necessarily the funniest (depending on taste), but the fact that he had very deliberately looked at the expected style and intentionally leaned 100% against it is presumably something GPT-X doesn’t do for low X.
Humour is already a powerful weapon of human manipulation. Our ex prime minister Boris Johnson wouldn’t have lasted nearly as long as he did (probably would never have got in) if he weren’t very good at getting us all to laugh.
Interesting. I didn’t know that anyone ever took just one paraceatamol. In the UK the packets all say that the dosage is two tablets, unless you are a small child. So people just kind of follow the instructions (at least I do). I have migraines infrequently (once a month, slightly less) but when I do realise I am having a migraine their is no way I am taking any paraceatamol less than what the packet stipulates. Firstly because I am in pain, and secondly because migraines seem to fog up by basic cognition in weird ways—I get weirdly stubborn about following instructions. Imagine me there, my vision swimming, can’t read because my sight is blury and the room is spinning. But the box says to “Always read the instructions in full before use” so I have to read this tiny text on this little paper before I can have my paracetamols. So when the instructions tell me 2, that is what it shall be. (I now throw the instructions away while sane).
Last year there was an “Around the World in 80 days” tv series. In that the information booth/clock and grand central station is a plot-relevant meet up point used by the characters more or less as a Schelling point. That is strong data that this is the accepted norm, it is also a push to make it the accepted norm. (Although the Empire State building hadn’t been built yet so the characters had no chance to pick it).
Another problem with the empire state building is that I thought I knew what the empire state building looked like because it is famous. But, when I actually went to New York it didn’t look like I thought it should—it turns out that (at least for me) the famous image in my head labelled “Empire State building” was actually an image of the Chrysler building. My suspicion is that this is really common.
I don’t know if GR or some cosmological thing (inflation) breaks reversibility. But classical and quantum mechanics are both reversible. So I would say that all of the lowest-level processes used by human beings are reversible. (Although of course thermodynamics does the normal counter-intuitive thing where the reversibility of the underlying steps is the reason why the overall process is, for all practical purposes, irreversible.)
This paper looks at mutual information (which I think relates to the cross entropy you mention), and how it connects to reversibility and entropy. https://bayes.wustl.edu/etj/articles/gibbs.vs.boltzmann.pdf
(Aside, their is no way that whoever maintains the website hosting that paper and the LW community don’t overlap. The mutual information is too high.)
The reversibility seems especially important to me. In some fundamental sense our universe doesn’t actually allow an AI (or human) no matter how intelligent to bring the universe into a controlled state. The reversibility gives us a thermodynamics such that in order to bring any part of the world from an unknown state to a known state we have to scramble something we did know back to a state of unknowing.
So, in our universe, the AI needs access to fuel (negative entropy) at least up to the task it is set. (Of course it can find fuel out their in its environment, but everything it finds can either be fuel, or can be canvas for its creation. But at least usually it cannot be both. Because the fuel needs to be randomised (essentially serve as a dump for entropy), while the canvas needs to be un-randomised.
Yes, I agree that it doesn’t solve all morality and provide the one true moral code. However, at least in my mind, it can be useful for working out which actions are acceptable within a certain narrow-ish group.
I think the categorical imperative is a nice framework. I don’t think your counterexample quite works for me. The babyeater applies the imperative, and is a morally upstanding babyeater who eats lots of its own children. Meanwhile the human applies the imperative and is a morally upstanding human who doesn’t kill any humans (baby or otherwise).
Both actors are acting morally, according to the imperative. That they are not acting the same way just shows that they are different agents with different values. Conversely, a babyeater and a human could both fail to live up to this imperative (eg. the babyeater thinks that child eating is good for the wider world, and wants everyone else to eat their own children to ensure the world stays nice, but it makes an exception for itself.). For both humans and babyeaters adopting the imperative might change their policy. It changes it in different ways for the two of them because they are different.
As a random aside: In Scotland court cases can have 3 possible outcomes, “Guilty”, “Not Guilty” and “Not Proven”. Where the last one (as I understand it) basically translates to “The jury are confident that you are guilty, but reluctantly admit that their isn’t enough legally admissible evidence to get a fair conviction.” I think that, legally “Not proven” is equivalent to “Not guilty”. Although while “Not guilty” should (ideally) undo any reputational damage from the accusation, “Not proven” will not.
Maybe Popper isn’t accurate. Their is this thing Popper does where he criticises Plato very heavily, but takes pains to make it clear he is not criticising Socrates, which comes across as really weird because (from what I understand) our modern knowledge of Socrates largely comes from Plato—so splitting hairs between which of the two said what with any precision is not really possible. So that is a flag that something fishy is going on.
I may be mistaken that almost everyone is selectively bred in Plato’s ideal city. Wikipeida’s summary seems to suggest that it was everyone: “The rulers assemble couples for reproduction, based on breeding criteria.” https://en.wikipedia.org/wiki/Republic_(Plato) , but I have only done a low-effort search on this.
The wider narrative of Popper’s book is that Fascism and Communism are both enemies of what he calls “The open society” (read: liberal, democratic, pluralist), and in that sense are similar to one another. The “communist” interpretation of Plato’s republic makes a lot of sense. Popper’s interpretation of it as fascist makes some sense as well though. It was apparently partly modelled of of ancient Sparta (the archetypical fascist state) - which also banned its citizens from owning personal property (especially gold), [the state provided a home, goods and slaves to every citizen], they also had a eugenics thing going on (babies inspected at birth for defects), and to some extent did the Plato thing where all children were raised collectively and not by their parents. (Maybe only males were raised collectively, not sure).