Jun 04, 2008
"Why are you doing these posts on quantum physics?" the one asked me.
"Quite a number of reasons," I said.
"For one thing," I said, "the many-worlds issue is just about the only case I know of where you can bring the principles of Science and Bayesianism into direct conflict." It's important to have different mental buckets for "science" and "rationality", as they are different concepts. Bringing the two principles into direct conflict is helpful for illustrating what science is and is not, and what rationality is and is not. Otherwise you end up trusting in what you call "science", which won't be strict enough.
"For another thing," I continued, "part of what goes into becoming a rationalist, is learning to live into a counterintuitive world—learning to find things underneath the surface that are unlike the world of surface forms." Quantum mechanics makes a good introduction to that, when done correctly without the horrible confusion and despair. It breaks you of your belief in an intuitive universe, counters naive realism, destroys your trust in the way that your cognitive algorithms look from inside—and then you're ready to start seeing your mind as a mind, not as a window onto reality.
"But you're writing about physics, without being a physicist," the one said, "isn't that... a little..."
"Yes," I said, "it is, and I felt guilty about it. But there were physicists talking complete nonsense about Occam's Razor without knowing the probability theory of it, so my hand was forced. Also the situation in teaching quantum mechanics is really awful—I saw the introductions to Bayesianism and they seemed unnecessarily difficult, but the situation in quantum mechanics is so much worse." It really is. I remember sitting there staring at the "linear operators", trying to figure out what the hell they physically did to the eigenvectors—trying to visualize the actual events that were going on in the physical evolution—before it dawned on me that it was just a math trick to extract the average of the eigenvalues. Okay, but... can't you just tell me that up front? Write it down somewhere? Oh, I forgot, the math doesn't mean anything, it just works.
"Furthermore," I added, "knowing about many worlds, helps you visualize probabilities as frequencies, which is helpful to many points I want to make."
"And furthermore," I said, "reducing time to non-time is a powerful example of the principle, in reductionism, that you should reduce something to something other than itself."
"And even furthermore," I said, "I had to break my readers of trust in Science, even trust in physicists, because it doesn't seem possible to think and trust at the same time."
"Many-worlds is really a very clear and simple problem," I said, "by comparison with the challenges you encounter in AI, which are around a hundred times less clear-cut. And many scientists can't even get many-worlds, in the absence of authority." So you are left with no choice but to aspire to do better than the average scientist; a hell of a lot better, in fact. This notion is one that you cannot just blurt out to people without showing them why it is necessary.
Another helpful advantage—I often do things with quite a few different purposes in mind, as you may have realized by this point—was that you can see various commenters who still haven't gotten it, who are still saying, "Oh, look, Eliezer is overconfident because he believes in many-worlds."
Well, if you can viscerally see the arguments I have laid forth, then you can see that I am not being careless in having an opinion about physics. The balance of arguments is overwhelmingly tipped; and physicists who deny it, are making specific errors of probability theory (which I have specifically laid out, and shown to you) that they might not be expected to know about. It is not just a matter of my forming strong opinions at random.
But would you believe that I had such strong support, if I had not shown it to you in full detail? Ponder this well. For I may have other strong opinions. And it may seem to you that you don't see any good reason to form such strong beliefs. Except this is not what you will see; you will see simply that there is no good reason for strong belief, that there is no strong support one way or the other. For our first-order beliefs are how the world seems to be. And you may think, "Oh, Eliezer is just opinionated—forming strong beliefs in the absence of lopsided support." And I will not have the time to do another couple of months worth of blog posts.
I am very far from infallible, but I do not hold strong opinions at random.
"And yet still furthermore," I said, "transhumanist mailing lists have been arguing about issues of personal identity for years, and a tremendous amount of time has been wasted on it." Probably most who argue, will not bother to read what I have set forth; but if it stops any intelligent folk from wasting further time, that too is a benefit.
I am sometimes accused of being overconfident and opinionated, for telling people that being composed of "the same atoms" has nothing to do with their personal continuity. Or for saying that an uploading scan performed to the same resolution as thermal noise, actually has less effect on your identity than a sneeze (because your eyes squeeze shut when you sneeze, and that actually alters the computational state of billions of neurons in your visual cortex). Yet if you can see your nows braided into causality of the river that never flows; and the synaptic connections computing your internal narrative, that remain the same from one time to another, though not a drop of water is shared; then you can see that I have reasons for this strong belief as well.
Perhaps the one says to me that the exact duplicate constructed on Mars, is just a copy. And I post a short comment saying, "Wrong. There is no copy, there are two originals. This is knowable and I know it." Would you have thought that I might have very strong support, that you might not be seeing?
I won't always have the time to write a month of blog posts. While I am enough of a Traditional Rationalist that I dislike trust, and will not lightly ask it, I may ask it of you if your life is at stake.
Another one once asked me: "What does quantum physics have to do with overcoming bias?"
Robin Hanson chose the name "Overcoming Bias"; but names are not steel chains. If I'd started my own personal blog for the material I'm now posting, I would have called it "Reinventing Rationality" or something along those lines—and even that wouldn't have been the real purpose, which would have been harder to explain.
What are these series of posts, really? Raw material for a popular book on rationality—but maybe a tenth of this material, or less, will make it into the book. One of the reasons I write long posts, is so that I can shorten them later with a good conscience, knowing that I did lay out the full argument somewhere. But the whole quantum physics sequence is probably not going to make it into the popular book at all—and neither will many other posts. So what's the rest of it for?
Sometimes I think wistfully of how much more I could have accomplished in my teenage years, if I had known a fraction of what I know now at age 15. (This is the age at which I was a Traditional Rationalist, and dedicated and bright as such ones go, but knew not the Way of Bayes.) You can think of these blog posts, perhaps, as a series of letters to my past self. Only not exactly, because some of what I now write, I did already know then.
It seems to me, looking back, that the road which took me to this Way, had a great deal of luck in it. I would like to eliminate that element of luck for those who come after. So some of what I post, is more formal explanations of matters which Eliezer-15 knew in his bones. And the rest, I only wish I had known.
Perhaps if you prime someone with enough material as a starting point, they can figure out the other 95% on their own, if they go on to study the relevant sciences at a higher technical level. That's what I hope.
Eliezer-15 was led far astray by the seeming mysteriousness of quantum mechanics. An antiproject in which he was aided and abetted by certain popular physicists—notably Sir Roger Penrose; but also all those physicists who told him that quantum physics was "mysterious" and that it was okay not to understand it.
This is something I wish I had known, so I explained it to me.
Why not just tell me to ignore quantum physics? Because I am not going to "just ignore" a question that large. It is not how we work.
If you are confronting real scientific chaos—not just some light matter of an experimental anomaly or the search for a better theory, but genuine fear and despair, as now exists in Artificial Intelligence—then it is necessary to be a polymath. Healthy fields have healthy ways of thinking; you cannot trust the traditions of the confused field you must reform, though you must learn them. You never know which way you'll need to draw upon, on venturing out into the unknown. You learn new sciences for the same reasons that programmers learn new programming languages: to change the way you think. If you want to never learn anything without knowing in advance how it will apply, you had best stay away from chaos.
If you want to tackle challenges on the order of AI, you can't just learn a bunch of AI stuff.
There finally comes a point where you get tired of trying to communicate across vast inferential distances. There comes a point where you get tired of not being able to say things to people without a month of preliminary explanation. There comes a point where you want to say something about branching Earths or identical particles or braids in the river that never flows, and you can't.
It is such a tremendous relief, to finally be able to say all these things. And all the other things, that I have said here; that people have asked me about for so long, and all I could do was wave my hands. I didn't have to explain the concept of "inferential distance" from scratch, I could just link to it. It is such a relief.
I have written hundreds of blog posts here. Think about what it would be like, to carry all that around inside your head.
If I can do all the long sequences on Overcoming Bias, then maybe after that, it will be possible to say most things that I want to say, in just one piece.
Part of The Quantum Physics Sequence
(end of sequence)
Previous post: "Class Project"