I discussed epistemology and cryonics with
Aubrey de Grey via email.
Click here to find the rest of the discussion. Yellow quotes are from Aubrey de Grey, with permission. Bluegreen is me, red is other.
Thanks again Elliot. I have several issues below, but they have a single common theme. This approach involves no open-ended creative thinking and not actually answering many specific criticisms and arguments. Nor does it come up with an explanation of the best way to proceed. It does not create knowledge.
I was probably unclear on that: that’s part (most, in fact, for interesting cases) of step 1, i.e. "Gather, as best one can in the time one has decided to spend, all the arguments recommending either of the alternative courses of action.” I didn’t mean to imply that this would be restricted to pre-existing arguments. So in other words, yes actually, I did use exactly this method in my evaluation of Estep’s criticism of SENS, and in my reply I articulated some of the results of that evaluation, namely some refutations of elements of the criticism. Consider your position as a reader: why did you accept my rebuttal as the last word? Why didn’t you write to Estep to ask him for a more thorough re-rebuttal than TR gave him the option of? Answer (I claim): because you subjectively decided that my rebuttal was impressive ENOUGH that Estep PROBABLY wouldn’t have a persuasive re-rebuttal, so you chose not to allocate time to contacting him. Note the quantitative, as well as subjective, elements of what I claim was your process (and I claim it confidently, because I can’t think of any other process you could have used for deciding not to write to Estep).
It's interesting you specifically express confidence, and can't think of any other process. This description isn't close to how I approached the Estep debate.
First, your rebuttal wasn't important here. I had already decided Estep was wrong before reading your rebuttal. That was easy. His position was largely philosophy, rather than being about detailed scientific points that I might have difficulty evaluating. While reading his text, I thought of criticisms of his arguments.
Actually, rather than being particularly impressed, I disliked three aspects of your rebuttal. But these criticisms were tangents, and are standard parts of academic culture. If I'm right about them, they don't make SENS wrong or Estep right. 1) Complaining about Estep's invective and saying you'd take the high road, but then returning some invective. 2) What I consider an overly prestigious writing style, partly intended to impress. 3) Arguing some over who has how much scientific authority and what they think (rather than only discussing substantive issues directly).
My interest in your rebuttal wasn't to learn why Estep was wrong – which I already knew. Note I say why he was wrong (explanation) rather than considering who is more impressive (ugh). Instead, I read to see how closely your thinking and approach matched my own (if I found important differences, I'd be interested in why, at least one of us would have to be wrong in an important way), to see what passes for debate in these kinds of papers in your field, and to see if you'd say an important point I'd missed or a mistake.
The main reason I didn't write to Estep is because I don't think he wants to have a discussion with me. My usual policy is not to write to paper authors who don't include contact information in their papers.
Now that you brought it up, I tried google and didn't find contact info there either. I think discussion is unwelcome. I did find his email in the GRG archives, but that's no invitation.
I actually would be happy to talk to him, if he wanted to have a discussion. Like if Estep volunteered to answer questions and criticisms from me, I'd participate. I like to talk to a variety of people, even ones I consider very bad. I want to understand irrationality and psychology better. And it helps keep my ideas exposed to all kinds of criticism. And I don't get myself stuck in unwanted polite or boring conversation.
You're right that I wouldn't expect Estep to change my mind if we talked. This is because I guessed an understanding of what he's like, which I have no criticisms of and no non-refuted alternatives to. Not probability. But this is minor. I'd talk to him anyway, the issue is he doesn't want to.
And I didn't just leave this to my judgment. I exposed my view on this matter to criticism. I wrote about it in public and invited criticism from the best thinkers I've been able to gather (or anyone else). (BTW you'd be welcome to join my Fallible Ideas discussion group and my private group.)
I don't do more than this because I have explanations of why other activities are better to spend my time on, and I don't know a problem/criticism with my approach or an explanation of a better approach. And all of this is open to public criticism. And I've made a large ongoing effort to have ready access to high quality criticism.
There is no such thing as how epistemologically good an explanation is.
I don’t get this. You’ve been referring to good and bad explanations throughout this exchange. What have you been meaning by that, if not epistemologically good and bad? I know you are saying that there are only refuted or non-refuted explanations, but you must have been meaning something else by good and bad, since you’ve definitely been using those adjectives - and other ones, like “clear”, “explicit” etc - in an unambiguously quantitative rather than binary/boolean sense, e.g.:
I can see how that'd be confusing. It's an imprecise but convenient way to speak. Depending what you're doing, you only need limited precision, so it can be OK. And it'd take forever to elaborate on every point, it's better only to go into detail on points where someone thinks it's worthwhile to, for some reason.
My position is that all correct arguments can be converted or translated into more precise statements that strictly adhere to the boolean epistemology approach.
Speaking of amount of clarity is a high level concept that's sometimes precise enough. You can, when you want to, get into more precise lower level details like pointing out specific ambiguous phrases or unanswered questions about the writer's position.
Saying an explanation is good or bad (in some amount) can quickly communicate an approximate evaluation without covering the details. It's loose speaking rather than epistemology.
They actually do have basic explanations, e.g. I've read one of them saying that vitrified brains look pretty OK, not badly damaged, to the unaided human eye. The implication is damage that's hard to see is small, so cryopreservation works well. This is a bad argument, but it's the right type of thing. They need this type of thing, but better, before anyone should sign up.
If it’s the right type of thing, what’s “bad" about it?
It is the right type of thing, meaning: it involves explanation and argument.
"Bad" here was an imprecise way to refer to some arguments I didn't write out upfront.
Damage that's hard to see to the naked human eye is not "small" in the relevant sense. The argument is a trick where it gets people to accept the damage is small (physical size in irrelevant regular daily life context), and implies the damage is small (brain still works well).
Why use unaided human eye instead of microscope? It's a parochial approach going after the emotional appeal of what people can see at scale they are used to. Rather than note appearances can be deceiving and try to help the reader understand the underlying reality, it tries to exploit the deceptiveness of appearances.
And it doesn't attempt to explore issues like how much damage would have what consequences. But with no concept of what damage has what consequences, even a correct statement of the damage wouldn't get you anywhere in terms of understanding the consequences. (And it's the consequences like having one's mind still revivable, or being dead, that people care about.)
- and more to the point, how bad?
Refuted.
What is your argument for saying "They need this type of thing, but BETTER (quantitative…), before anyone should sign up”? How much better, and why?
It needs to be better to the point it isn't refuted. Because it's a bad idea to act on ideas with known flaws.
(There are some complications here like they don't actually know my criticism, the flaws aren't known to them. What is "refuted" in each person's judgment depends on their individual knowledge. That's a tangent I won't write about now.)
You can’t just say “non-refuted”, because you know as well as I do that any argument about anything interesting can be met with a counter-argument, which itself can be met, etc., unless one has decided in advance how to terminate the exchange.
No, I disagree!
It's hard to keep up meaningful criticism for long.
Yes someone can repeat "That's dumb, I disagree" forever. But a criticism, as I mean it, is an explanation of a flaw/mistake with something, and this kind of bad repetitive objection doesn't explain any mistakes.
I don't think you had this kind of repetition in mind, or you wouldn't have specified "about anything interesting". "That's dumb, I disagree" can be used on trivial topics just as well as interesting topics.
I think you're saying that substantive critical discussion doesn't terminate and keeps having good points indefinitely. Until you terminate it arbitrarily.
I think good points are hard to come by. What are "good" points here, specifically? Ones which aren't already refuted by pre-existing criticism.
As you go along in productive discussions, you build up criticisms of many things. Not just of specific points, but of whole categories of points. Some of the criticisms have "reach" as DD calls it. They have some level of generality, they apply to many things. As criticism builds up, it gets progressively harder to come up with new ideas which aren't already refuted by existing criticism.
The reason many discussions don't look like this in practice is because of irrationality and bad methods, rather than discussions having to be that way.
My fundamental problem remains: you haven’t given me a decision-making algorithm that terminates, or even usually terminates, in an amount of time that I can specify in advance.
It's a mistake to 100% rigidly specify time limits in advance. Reasoning for time limits should be open to criticism.
The closest to a flowchart I can give you is something like:
- think creatively etc, as discussed previously
- when nearing a resource limit (like time), start referring to this limit in arguments, to bring arbitration to a close. e.g. instead of "I disagree with that, and here's why in detail", a side might say, "I disagree with that, but we don't have time to get into it. Instead, here is what I propose that we may both find acceptable."
- as resources get tighter, it gets easier to please all sides. like, they may agree it's better to flip a coin than not to reach a decision by a certain deadline.
- reasonable sides understand their fallibility and don't want anyone to go along with something without persuasion. and they understand persuasion on some point can exceed a resource limit. so they actively PREFER to find mutually agreeable temporary measures for now, when appropriate, while working on persuasion more in the longer term as more resources are available
- sometimes things go smoothly. no problem. sometimes they don't. when they don't, there are specific techniques which can be used.
- specifically, one considers questions of the form, "Given the context - and specifically not reaching agreement on points X, Y and Z, but having agreement on A, B and C - what can be done that's mutually agreeable? What can be done on this issue with the limited agreement?"
- while working on this new question, if there are any sticking points, then a similar question can be asked adding those sticking points to the exclusion list.
- these questions reduce the complexity and difficulty of the arbitration as low as needed.
- the more you use questions like this and temporarily exclude things due to resource limits, the easier it is to reach agreement. if it's different people, it goes to "since we disagree so much, let's go our separate ways". the harder case is either when a person has conflicting ideas or two people are entangled (e.g. parent and child). but that still reaches outcomes like, "given we disagree so much, and we need a decision now, let's flip a coin". both sides can prefer that to any known alternatives, in which case it's a win/win outcome.
- but what if they don't agree to flip a coin over it? well, why not? this is fundamentally why a flowchart doesn't work. because people disagree about things for reasons, and you can't flowchart answers to those reasons.
- but basically sides will either agree to a coin flip (or some better default they know of), or else they will propose something they consider a better idea. a better idea while being reasonable – so like, something they think the other side could agree with, not something that'd take a great deal of persuasion involving currently-unavailable resources.
- if sides are unreasonable – e.g. try to sabotage things, or just want their initial preference no matter what – then any conflict resolution procedure can stall or fail. that's unavoidable.
- this doesn't terminate in predictable-in-advance time because sometimes everyone agrees that the deadline is less important than further arbitration, and prefers to allocate more resources. i don't think this is a problem. it can terminate quickly when that's a good idea. the only reason it won't terminate quickly is specifically because a side disagrees that terminating quickly is a good idea in this case. (and if that happens, there will be a reason in context, which may be right or wrong, and there is no one-size-fits-all flowchart answer to it, it matters what the reason is)
I have one. It’s not perfect - I accept all your criticisms of it, I think - but the single feature that it terminates in a reasonably predictable time (just how predictable is determined, of course, by how close together one chooses the two cutoff probabilities to be) is so important that I think the method is better than any alternative that doesn’t reliably terminate.
The thing is, I think you DO have an algorithm that reliably terminates, and that despite your protestations it is pretty much identical to mine. Look at this example for illustration: Also we do have an explanation of why different experiments measuring the speed of light in a vacuum get the same answer. Because they measure the same thing. Just like different experiments measuring the size of my hand get the same answer. No big deal. The very concepts of different photons all being light, and of them all having the same speed, are explanatory ideas which make better sense out of the underlying reality.
Nonsense, because each measurement measures different photons, and we have no better explanation for all photons having the same speed than for all pigeons having the same mass. This is not trivial: indeed, I recall that Wheeler made quite a big deal out of the awfully similar question of the mass of the electron and proposed that there is in fact only one electron in the Universe. We have explicitly made the choice not to enquire further on the question.
If you go deeper, then yes I don't know everything about physics. There's some initial explanations about this stuff, but it's limited.
I'm unclear on why this is important. I don't study physics more because I prefer to do other things and I don't know of any criticisms/problems with my approach. Even if I did study physics all day, I still wouldn't know everything about it and would make choices about which things to enquire further about, because I couldn't do everything at once. I would think of an explanation for how I should approach the matter, adjust or rethink until no criticism, and do that.
Or this one: Person wants to buy something but hesitates to part with their money. Thinks about how awesome it would be, changes mind, happily buys. Solved.
That only works with an additional step that comes just before “happily buys”, namely “switches brain off before remembering that one might soon change one’s mind back”. And, actually, another step that says “remembers that one is really good at not crying over spilt milk, i.e. once the money is spent one is happy to live with whatever regret one might later have”. And so on. I know you know this.
But I don't know it. I deny it.
I think switching off the brain and trying not to think of some issues, because one couldn't deal with the issues if he paid attention to them, is a really bad approach. It's choosing winners in an irrational way – instead of resolving the conflict of ideas, you're playing the role of an arbiter who only lets one side speak, then declares them the winner.
About spilt milk: Sometimes people think of that and it helps them happily buy something. But sometimes people don't. It's not required. There are many optional steps that people find useful, or not, depending on their specific circumstances.
But, yet, you were fine with just writing “Solved”! I conclude that you DO have a termination procedure in your algorithm, and moreover that it’s an indisputably vague and subjective and probabilistic and epistemologically hole-riddled one just like mine, and I don’t know why you’re having such trouble admitting it.
I don't concede because I disagree.
I think a rational non-hole-riddled epistemology is possible, and that I understand it.
Let’s get back to cryonics - largely because I am now somewhat invested in the goal of changing your mind about signing up, coupled of course with the equally legitimate converse goal of giving you a fair shot at changing mine.
Let’s start with the specific question I already referred to above: They actually do have basic explanations, e.g. I've read one of them saying that vitrified brains look pretty OK, not badly damaged, to the unaided human eye. The implication is damage that's hard to see is small, so cryopreservation works well. This is a bad argument, but it's the right type of thing. They need this type of thing, but better, before anyone should sign up.
As this stands, as I just said, it is too vague to be amenable to refutation even in principle, i.e. it doesn’t meet your own epistemological standards, because it doesn’t incorporate any statement of (let alone any argument for) your criterion for how good that explanation needs to become.
my standard is: is there a criticism of it? not some criterion for how good.
As above, “non-refuted” doesn’t work, because that relies on consideration of (for example) how much time I choose to allocate to giving you refutations and how much you choose to allocate to giving me refutations, and I sense that that that’s a decidedly non-level playing field.
You mean, it's not a level playing field because I allocate more time to trying to get this issue right? Or at least to writing down my thinking, so that if I'm mistaken someone could tell me?
BTW, what is your explanation of why no one has written good explanations of why to sign up for cryonics anywhere? Why have they left it to you to write it, instead of merely link things?
(Good explanations to what standard? Your own. If stuff met your standards you'd link it instead of writing your own.)
My (unashamedly justificationist) starting-point is that the absence of gross damage feels like enough evidence for revivability to satisfy me that people should sign up.
The evidence you refer to is consistent with infinitely many positions, including ones that conclude not to sign up for cryo. Considering it evidence for a specific conclusion, instead of others it's equally consistent with, is some mix of 1) arbitrary 2) using unstated reasons
Why should a fact fully compatible with non-revivability be counted as "evidence for revivability"?
So let’s start with you amplifying your above statement, with a sense of what you WOULD view as a good enough (yes I said it) argument, to give me some goalposts to aim for.
The goalposts fundamentally are: I don't have further criticism.
This is hard because I have many criticisms. But there really have to be ways for me to get answers to all of them (though not all from you personally). Or else you'd be asking me to do something I have a reason not to do; you'd be asking me to just ignore my own judgment arbitrarily for no reason.
I also think you overestimate how problematic this is because you're used to debates that don't go anywhere, don't resolve anything, because of how terribly irrational most people are.
Another big factor is people who don't want to be persuaded. Rational persuasion is impossible with unwilling subjects. People always have to persuade themselves and fill in lots of details, you can't tell them everything and perfectly customize it all to their context and integrate it with all their other ideas. They have to play an active role, or any persuasion will be superficial.
Something that I'd see as a good starting place is explanations connecting different amounts of damage to consequences like being fine or dead, and quantifying the amount of damage Alcor and CI cause today.
Continue reading the next part of the discussion.