Posts
Comments
The ancients considered everything to be the work of spirits. The medievals considered the cosmos to be a kingdom. Early moderns likened the universe to a machine. Every age has its dominant metaphors. All of them are oversimplifications of a more complex truth.
Suppose you had an identical twin with identical genes and, until very recently, an identical history. From the perspective of anyone else, you're similar enough to be interchangeable with each other. But from your perspective, the twin would be a different person.
The brain is you, full stop. It isn't running a computer program; its hardware and software are inseparable and developed together over the course of your life. In other words, the hardware/software distinction doesn't apply to brains.
To call something an "uploaded brain" is to make two claims. First, that it is a (stable) mind. Second, that it is in some important sense equivalent to a particular meat brain (e.g., that its output is the same as the meat brain, or that its experiences are the same as the meat brain's). The sorts of methods you're talking about to stabilize the mind help with the first claim, but not with the second.
I've always struggled to make sense of the idea of brain uploading because it seems to rely on some sort of dualism. As a materialist, it seems obvious to me that a brain is a brain, a program that replicates the brain's output is a program (and will perform its task more or less well but probably not perfectly), and the two are not the same.
What if there would never be someone I trusted who could tell me I was Good Enough, that things were in some sense Okay?
The internalized feeling that you're not okay is a huge part of what motivates you to become better. If you lost it, you would be much more likely to become complacent and stagnate. Both inner peace and relentless drive are profoundly valuable, but they are mutually exclusive.
I certainly agree that brains are complicated.
I think part of the difference is that I'm considering the uploading process; it seems to me that you're skipping past it, which amounts to assuming it works perfectly.
Consider the upload of Bob the volunteer. The idea that software = Bob is based on the idea that Bob's connectome of roughly 100 trillion synapses is accurately captured by the upload process. It seems fairly obvious to me that this process will not capture every single synapse with no errors (at least in early versions). It will miss a percentage and probably also invent some that meat-Bob doesn't have.
This raises the question of how good a copy is good enough. If brains are chaotic, and I would expect them to be, even small error rates would have large consequences for the output of the simulation. In short, I would expect that for semi-realistic upload accuracy (whatever that means in this context), simulated Bob wouldn't think or behave much like actual Bob.
Surely both (1) and (2) are true, each to a certain extent.
Are the random thermal fluctuations pushing me around somehow better than the equally random measurement errors pushing my soft-copy around?
It depends. We know from experience how meat brains change over time. We have no idea how software brains change over time; it surely depends on the details of the technology used. The changes might be comparable, but they might be bizarre. The longer you run the program, the more extreme the changes are likely to be.
I can't rule it out either. Nor can I rule it in. It's conceivable, but there are enough issues that I'm highly skeptical.
Let's try again. Chaotic systems usually don't do exactly what you want them to, and they almost never do the right thing 1000 times in a row. If you model a system using ordinary modeling techniques, chaos theory can tell you whether the system is going to be finicky and unreliable (in a specific way). This saves you the trouble of actually building a system that won't work reliably. Basically, it marks off certain areas of solution space as not viable.
Also, there's Lavarand. It turns out that lava lamps are chaotic.
That wasn't well phrased. Oops.
Any physical system has a finite amount of mass and energy that limit its possible behaviors. If you take the log of (one variable of) the system, its full range of behaviors will use fewer numbers, but that's all that will happen. For example, the wind is usually between 0.001 m/s (quite still) and 100 m/s (unprecedented hurricane). If you take the base-10 log, it's usually between -3 and 2. A change of 2 can mean a change from .001 to .1 m/s (quite still to barely noticeable breeze) or a change from 1 m/s to 100 m/s (modest breeze to everything's gone). For lots of common phenomena, log scales are too imprecise to be useful.
Chaotic systems can't be predicted in detail, but physics and common sense still apply. Chaotic weather is just ordinary weather.
That's the point. Nobody thought such tiny variations would matter. The fact that they can matter, a lot, was the discovery that led to chaos theory.
Consider - A typical human brain has ~100 trillion synapses. Any attempt to map it would have some error rate. Is it still "you" if the error rate is .1%? 1%? 10%? Do positive vs. negative errors make a difference (i.e. missing connections vs. spurious connections)?
Is this a way to get new and exciting psychiatric disorders?
I don't know the answers, or even how we'd try to figure out the answers, but I don't want to spend eternity as this guy.
Hastings clearly has more experience with chaos theory than I do (I upvoted his comment). I'm hoping that my rather simplistic grasp of the field might result in a simpler-to-understand answer (that's still basically correct).
Chaos theory is a branch of math; it characterizes models (equations). If your model is terrible, it can't help you. What the theory tells you is how wildly your model will react to small perturbations.
AFAIK the only invention made possible by chaos theory is random number generators. If a system you're modeling is extremely chaotic, you can use its output for random numbers with confidence that nobody will ever be able to model or replicate your system with sufficient precision to reproduce its output.
an exponential decrease in measurement error will only buy you a linear increase in how long that simulation is good for.
True, and in the real world attempts to measure with extreme precision eventually hit limits imposed by quantum mechanics. Quantum systems are unpredictable in a way that has nothing to do with chaos theory, but that cashes out to injecting tiny amounts of randomness in basically every physical system. In a chaotic system those tiny perturbations would eventually have macroscopic effects, even in the absence of any other sources of error.
The seminal result for chaos theory came from weather modeling. An atmospheric model was migrated to a more powerful computer, but it didn't give the same results as it had on the old computer. It turned out that, in the process of migration, the initial condition data had been rounded to the eighth decimal place. The tiny errors compounded into larger errors, and over the course of an in-model month the predictions completely diverged. An error in the eighth decimal place is roughly comparable to the flap of a butterfly's wing, which led to the cliche about butterflies and hurricanes.
If you're trying to model a system, and the results of your model are extremely sensitive to miniscule data errors (i.e. the system is chaotic), and there is no practical way to obtain extremely accurate data, then chaos theory limits the usefulness of the model. It may still have some value; using standard models and available data it's possible to predict the weather rather accurately for a few days and semi-accurately for a few days more, but it may not be able to predict what you need.
This is one reason I've always been skeptical of the "uploaded brain" idea. My intuition is that inevitable minor errors in the model of the brain would cause the model to diverge from the source in a fairly short time.
I don't want to live in a world where there's only the final survivors of selection processes who shrug indifferently when asked why we don't revive all the beings who were killed in the process which created the final survivors.
If you could revive all the victims of the selection process that brought us to the current state, all the crusaders and monarchists and vikings and Maoists and so, so many illiterate peasant farmers (on much too little land because you've got hundreds of generations of them at once, mostly with ideas that make Putin look like Sonia Sotomayor), would you? They'd probably make quite the mess. Bringing them back would probably restart the selection process and we probably wouldn't be selected again. It just seems like a terrible idea to me.
In that case, it's non-farmers by a good margin. Our ancestry goes back well over a billion years, mostly in species with short generation times. Farming goes back roughly ten thousand years in a single species with a ~25 year generation time.
Almost everyone believes one and only one of the following statements:
- We have, or can soon develop, the technology to power society with clean energy at a modern standard of living, or
- Global warming is not a major problem (e.g. it won't damage agricultural productivity to famine-inducing levels).
Logically, these statements have nothing to do with each other. Either, neither, or both could be correct (I suspect neither). They are, in the words of the post, questionable assumptions.
See also Beyond the Peak – Ecosophia and Germany's economy struggles with an energy shock that's exposing longtime flaws | AP News.
FYI, the green energy field is much, much worse in this regard than the AI field.
I should have added - Determine whether this is a modeling problem or a manufacturing problem. If the model was sound but the physical screw was faulty, you'll need an entirely different response.
Actually ideal:
- Reinforce that screw by the end of the day.
- Fix the modeling error by the end of the week.
- Develop a more robust modeling methodology over the next few months.
- Brainstorm ideas to improve the institutional culture (without sacrificing flexibility, because you're aware that these values require a tradeoff). Have a proposal ready for the next board meeting.
Now you're getting it. The world can't be fixed. It can't even be survived. But it can be a nice place to live.
The worst people in the world, the Stalins and the Osama bin Ladens, try to be heroes but they're as flawed as anyone else. If they start to succeed those flaws can manifest in horrifying ways. They often destroy imperfect but necessary things in attempts to build perfect things that can't exist, like true Communism or functional political Islam. Humility and temperance are called heavenly virtues for a reason.
That even though I decided that my morality would never demand that I be a hero... there nonetheless just isn't a coherent, enduring shape that fits my soul that doesn't make that the thing I ultimately want for myself.
Reading that, I'm not sure whether you're grieving because you've given up on that belief or because it's true. I hope the former. The desire to be a hero is dangerous - a hero needs villains. As Nietzsche said, he who fights with monsters might take care lest he thereby become a monster.
I'm not saying that you shouldn't do what you can, but we all fall short of perfection. We all burn fossil fuels, for one thing. A bit of humility is in order.
It's not your job to save the world. That's God's job, and if that job has driven God crazy it would explain quite a bit.
I think you're onto something. I think, for this purpose, "child" means anyone who doesn't know enough about the topic to have any realistic chance at successful innovation. A talented 16 year old might successfully innovate in a field like music or cooking, having had enough time to learn the basics. When I was that age kids occasionally came up with useful new ideas in computer programming, but modern coding seems much more sophisticated. In a very developed field, one might not be ready to innovate until several years into graduate school.
A 16-year-old Salafi will be strongly influenced by his Salafi upbringing. Even if he* rebels, he'll be rebelling against that specific strain of Islam. It would take a very long and very specific journey to take him toward California-style liberalism; given the opportunity to explore he'd likely end up somewhere very different.
*My understanding of this particular Islamic school is hazy, but I doubt our student is female.
A child who's educated in a Salafi school has two choices - become a Salafi or become a failed Salafi. One of those is clearly better than the other. Salafis, like almost every adult, know how to navigate their environment semi-successfully and the first job of education is to pass on that knowledge. It would better if the kid could be given a better education, but the kid won't have much control over that (and wouldn't have the understanding to choose well). Kids are ignorant and powerless; that's not a function of any particular political or philosophical system.
I think in general it's best for children to learn from adults mostly by rote. Children should certainly ask questions of the adults, but independent inquiry will be at best inefficient and usually a wrong turn. The lecture-and-test method works, and AFAIK we don't have anything else that teaches nearly as well.
Later, when they have some understanding, they can look around for better examples.
Yeah, that's the idea I was going for.
I propose a test - if apologizing for or clarifying a controversial position is obviously a bad move, you're dealing with Level 4 actors. In such cases, your critics don't care about what you believe. Their narrative calls for a villain, and they've chosen you.
I think there are more limits than that because plausibility matters. The set of positions Ibram X Kendi could plausibly take is very different from the positions available to Donald Trump. Too big a reach and you'll look insincere, opportunistic, or weak. It's easy to alienate your social coalition and much harder to gain acceptance in a new one.
You seem to be steering in the direction of postmodernism, which starts with the realization that there are many internally consistent yet mutually exclusive ways of modeling the world. Humility won't solve that problem, but neither will a questioning mindset.
Every intellectual dead-end was once the product of a questioning mind. Questioning is much more likely to iterate toward a dead end than to generate useful results. This isn't to say that it's never useful (it obviously can be), but it rarely succeeds and is only the optimal path if you're near the frontiers of current understanding (which schoolchildren obviously aren't).
The best way to get out of a local maximum that I've found is to incorporate elements of a different, but clearly functional, intellectual tradition.
I don't know about Moldova, but it seems obvious that the creation of modern Israel depended on the idea that the Palestinians could be managed and equally obvious that it hasn't worked out that way. The only real endgames are genocide or leaving and personally I'd vote for leaving.
In 14 centuries of Islamic history from Spain to Indonesia, with limited travel and much regional variation for most of it, there will be many opportunities to find examples that match our own culture's Current Thing. Some Muslims are hypocrites; some Westerners look for homosexual subtext where none was intended. Many Muslim empires have risen in vigor and fallen in decadence. Still, the orthodox position is clear - homosexuality is both sinful and illegal. I've seen a Jew eat pork and laugh it off; it would be a mistake to make pork a key component of an appeal to militant Jews.
More to the point, in this era gay rights are associated with the West at its most liberal, which is exactly what Islamists oppose. Activity that might have been tolerated a thousand years ago is now perceived as a Western obsession and its practitioners as enemy sympathizers.
Trying to replace enemy leadership with more congenial leadership never works. You reliably get a corrupt puppet government and an insurgency. The only exception is after a comprehensive defeat of the enemy (i.e. post-WWII Germany and Japan), which begs the question (in the sense that this tactic would allow us to win only in cases where we've already won).
One thing to keep in mind is that a lot of the food supply in the Middle East and North Africa used to come from Russia and Ukraine (link). Actions targeted at increasing food security in the region won't solve the political issues, but probably would turn down the temperature of the region. It would also help with the immigration crisis in Europe. It's not an easy or quick solution (growing food takes time, and the scale of the problem is staggering), but it's hard to see how peace would come without food.
Sun Tzu says that the keys to victory lie in knowing yourself and your enemy. When I got to #4, it became obvious that you know very little about Islam. There are no LGBTQ+ safe spaces in Islam. A relevant wikipedia page says "Homosexual acts were forbidden (haram) in traditional Islamic jurisprudence and therefore were subject to punishment. The types of punishment prescribed for non-heterosexual activities include flogging, stoning, and the death penalty, depending on the particular situation and the school of thought." The major Muslim countries are signatories to a UN counter-statement opposing gay rights. See here for a consensus statement of the Australian National Imam's Council, whom I would expect to be much more liberal than the average Palestinian, saying, "From the time of the Prophet until now, all scholars of every time and era, have agreed that the practice of homosexuality is a forbidden act and a sin in Islam." It's just a complete non-starter.
Me - Ph.D. in solid state materials chemistry. Been out of the game for a while. Less understanding of physics than some other commenters but have a different perspective that might be useful.
My first thought is that they have a minority phase; the samples are likely ~99% LK99 and ~1% unknown phase with weird properties. You can see it in the video; part of the specimen is levitating but a corner of it isn't.
The second thing I would do is try to make a bunch of variants with slightly different compositions to identify the minority phase.
The first thing I would do is try to make versions with different amounts of hydrogen. Hydrogen is ubiquitous, diffuses readily into and out of most materials, and is invisible to most materials analysis techniques, but it can have a profound effect on a material's properties. If you get different properties by annealing the sample under high-pressure hydrogen*, you're on the right track.
*For safety's sake you would typically use forming gas, a non-flammable mixture of hydrogen with nitrogen (or sometimes argon). Ammonia is also sometimes used but is more dangerous.
Strongly upvoted. A few comments:
I think of a human being as a process, rather than a stable entity. We begin as embryos, grow up, get old, and die. Each step of the process follows inevitably from the steps before. The way I see it, there's no way an unchanging upload could possibly be human. An upload that evolves even less so, given the environment it's evolving in.
On a more practical level, the question of whether a software entity is identical to a person depends on your relationship to that person. Let's take Elizer Yudkowski for example:
- I personally have never met the guy but have read some of the stuff he wrote. If you told me that he'd been replaced with a LLM model six months ago, I wouldn't be able to prove you wrong or have much reason to care.
- His friends as family would feel very differently, because they have deeper relationships to him and many of the things they need from him cannot be delivered by an LLM.
- To Elizer himself, the chatbot would obviously not be him. Elizer is himself, the chatbot is something else. Uniquely, Elizer doesn't have a demand for Elizer's services; he has a supply of those services that he attempts to find demand for (with considerable success so far). He might consider the chatbot a useful tool or an unbeatable competitor, but he definitely wouldn't consider it himself.
- To Elizer's bank it's a legal question. When the chatbot orders a new server, does Elizer have to pay the bill? If it signs a contract, is Elizer bound?
- Does the answer change if there's evidence that it was hacked? What sorts of evidence would be sufficient?
- If asked, AI-lizer would claim to perceive itself as Elizer. Whether it actually has qualia, and what those qualia are like, we will not know.
A lot of the nonprofit boards that I've seen use a "consent agenda" to manage the meeting. The way it works is:
- The staff create the consent agenda and provide it to the board members perhaps a week in advance.
- Any single board member can take any item off the consent agenda and onto the regular agenda.
- The consent agenda is passed in a single motion. It always passes unanimously, because anything that any member thinks merits attention has been moved onto the regular agenda (where it is separately discussed and voted on).
It doesn't do much for governance directly, but fewer time-wasting consent votes can make room for more discussion of issues that matter.
In the US, parties still aren't recognized by the Constitution. Every election is a choice between all of the people who qualify for the ballot for each office. Several groups of like-minded politicians quickly emerged, and over time these became our major parties.
It's not uncommon for an American candidate to run as an independent (i.e. not affiliated with a party), although they hardly ever win.
To the extent that I understand what you're saying, you seem to be arguing for curiosity as a means of developing a detailed, mechanistic ("gears-level" in your term) model of reality. I totally support this, especially for the smart kids. I'm just trying to balance it out with some realism and humility. I've known too many people who know that their own area of expertise is incredibly complicated but assume that everything they don't understand is much simpler. In my experience, a lot of projects fail because a problem that was assumed to be simple turned out not to be.
I get your point, and I totally agree that answering a child's questions can help the kid connect the dots while maintaining the kid's curiosity. As a pedagogical tool, questions are great.
Having said that, most people's knowledge of most everything outside their specialties is shallow and brittle. The plastic in my toothbrush is probably the subject of more than 10 Ph.D. dissertations, and the forming processes of another 20. This computer I'm typing on is probably north of 10,000. I personally know a fair amount about how the silicon crystals are grown and refined, have a basic understanding of how the chips are fabricated (I've done some fabrication myself), know very little about the packaging, assembly, or software, and know how to use the end product at a decent level. I suspect that worldwide my overall knowledge of computers might be in the top 1% (of some hypothetical reasonable measure). I know very little about medicine, agriculture, nuclear physics, meteorology, or any of a thousand other fields.
Realistically, a very smart* person can learn anything but not everything (or even 1% of everything). They can learn anything given enough time, but literally nobody is given enough time. In practice, we have to take a lot of things on faith, and any reasonable education system will have to work within this limit. Ideally, it would also teach kids that experts in other fields are often right even when it would take them several years to learn why.
*There are also average people who can learn anything that isn't too complicated and below-average people who can't learn all that much. Don't blame me; I didn't do it.
Being honest, for nearly all people nearly all of the time questioning firmly established ideas is a waste of time at best. If you show a child, say, the periodic table (common versions of which have hundreds of facts), the probability that the child's questioning will lead to a significant new discovery are less that 1 in a billion* and the probability that they will lead to a useless distraction approach 100%. There are large bodies of highly reliable knowledge in the world, and it takes intelligent people many years to understand them well enough to ask the questions that might actually drive progress. And when people who are less intelligent, less knowledgeable, and/or more prone to motivated reasoning are asking the questions, you can get flat earthers, Qanon, etc.
*Based on the guess that we've taught the periodic table to at least a billion kids and it's never happened yet.
I think a better way to look at it is that frequentist reasoning is appropriate in certain situations and Bayesian reasoning is appropriate in other situations. Very roughly, frequentist reasoning works well for descriptive statistics and Bayesian reasoning works well for inferential statistics. I believe that Bayesian reasoning is appropriate to use in certain kinds of cases with a probability of (1-delta), where 1 represents the probability of something that has been rationally proven to my satisfaction and delta represents the (hopefully small) probability that I am deluded.
Wars are an especially nasty type of crisis because there's an enemy. That enemy will probably attempt to use your software for its own ends. In the case of your refugee heatmap idea, given that the Russians are already massacring civilians, that might look like a Russian artillery commander using it to deliberately target refugees. Alternately, they might target incoming buses to prevent the refugees from getting out of the Ukrainian military's way and make the Ukrainians spend essential resources on feeding and protecting them.
Does the Russian military even have the tech dependencies that would make them vulnerable to cyber attacks? I think they're pretty analog.
I spent about 20 years in academic and industrial research, and my firm belief is that almost nobody spends nearly enough time in the library. There have been hundreds of thousands of scientists before you; it is overwhelmingly likely that your hot new idea has been tried before. The hard part is finding it; science is made up of thousands of tiny communities that rarely talk to each other and use divergent terminology. But if you do the digging, you may find a paper from Egypt in 1983 that describes exactly why your project isn't working (real example). Finding that paper two weeks into the project is much better than finding it five years later.
The US has at least 16 intelligence agencies, but we still went into Iraq.
Oddly, it's probably easier for Putin to get credible information about Ukraine's military than about his own. Fewer people have an interest in lying to him about Ukraine.
Richard Hanania's Public Choice Theory and the Illusion of Grand Strategy is worth reading on this general topic, although it's mostly about the Iraq war.
He doesn't use these words (that I recall, and I'm only partway through), but the behavior of states is less like rationality and more like autism. What I mean is that states act on internal motivations that are only very weakly coupled to the objective strategic reality.
Putin probably didn't have access to reliable information about the capabilities of his military vs. Ukraine's military. How could he? What General would dare tell the Tsar that the Russian army sucks? Who would dare tell the General? I'm sure the soldiers and junior officers knew they had problems, but probably had little idea how their capabilities compared to the Ukrainians' capabilities.
On the other hand, Putin probably knows as much as anyone about the highest ranks of Russian politics. He sees those people all day, and it's almost inevitable that they constitute the community that matters to him.
Re aliens - Fair enough. Some very simple alien, perhaps the Vulcan equivalent of a flatworm, may be well within our capability to understand. Is that really what we're interested in?
Re machine learning - The data for machine learning is generally some huge corpus. The question is whether we're even capable of understanding the data in something like the manner the algorithm does. My intuition says no, but it's an open question.
I'd like to add two pieces of evidence in favor of the weak unlearnability hypothesis:
(1) Humpback whales have songs that can go on for days. Despite decades of study, we don't really understand what they're saying.
(2) The output of machine learning algorithms (e.g. Google's Deep Dream) can be exceedingly counterintuitive to humans.
Whales are our distant cousins and humans created machine learning. We might reasonably suppose that actual aliens, with several billion years of completely independent evolution, might be much harder to understand.
We actually do pretty much the opposite of that in the U.S. Student loans have a Federal guarantee, so the incentive is to sign people up for as much education as possible. If they succeed, great. If they fail, they'll be paying off the loans until they die at which time Uncle Sam will pay the balance. With compounding interest, the ones who fail are the most profitable.
we don’t have a step-by-step checklist to follow in order to use informal mathematical arguments
If we did, the checklist would define a form and the mathematical arguments would become formal.
Terrence Tao uses the term post-rigorous to describe the sort of argument you're talking about. It's one of three stages. In the pre-rigorous stage, concepts are fuzzy and expressed inexactly. In the rigorous stage, concepts are precisely defined in a formal manner. In the post-rigorous stage, concepts are expressed in a fuzzy and inexact way for the sake of efficiency by people who understand them on a rigorous level; key details can be expressed as rigorously as necessary but the irrelevant details of a full proof are omitted.