"Applied Rationality Training Regime" Overall

    <  |  ^  | 

 January 31, the end of this subproject. (Click the "^" link above for the subproject index page, on which this page will be the last entry.) I've spent at least a little time (and one blog post) for each day of January, on the idea of "applied rationality" -- specifically on what Mark Xu describes as his take on CFAR's (the Center For Applied Rationality) take on ways to get your thinking and your actions to be a little more in line with your goals, a little less counter-productive; how to improve a little in almost whatever direction you wanted to improve in. In the end, how to improve your methods of self-improvement.

And has it helped? Yes, I think so. Some. I think doing this has 

  (a) pushed me into forming somewhat better models of what I'm doing when I try to form habits (although they're very awkward to phrase, which probably means I don't really understand them);

  (b) made it easier to form some new simple-but-useful habits;

  (c) made it easier to stop doing/saying annoying things that I wanted to stop;

  (d) made it easier to approach some problems more systematically.

  As outcomes go, that's not bad.  TAP, Noticing, Murphyjitsu+Socratic Duck; in fact I'm now starting each morning with a brief meeting inside my head, a meeting with the characters I invented as labels (at least as labels) for what I need to do for each of those. The habits I'm trying to reinforce get assigned to Checklist Charlie, my TAP-dancing Laputan flapper-spider. Warnings about what I'm about to say or do come from Marian, the Noticer. For working through problems and anticipating troubles, I talk with Spider-Duck as well as Sarah Som and Jim Pright who form my Murphyjitsu team. I'm trying to use sketches more, simply because I'm better at symbols...I might end up trying to create a spider who draws.

  Each of these is really, I think, serving as a low-bandwidth messenger (or at least message-locus, a label for a mail-drop) between my "conscious self", my central story as continually updated, and the other systems that don't communicate very well with it. You can describe it as communication between System 2 and varying parts of System 1 if you wish, but I don't frame it that way.  I'm starting from the same point as Arnold Geulincx, the 17th-century follower of Descartes who noted that his identity was his consciousness so what he wasn't conscious of, such as muscle management, wasn't, couldn't be, part of his identity. "Since, then, the movements of my body take place without my knowing how ..." it followed that "I do not cause my own bodily actions" and in the end, "I am therefore a mere spectator of this machine." I do think that's going too far: the label "I" within my story applies not only to having viewed, but to having shared some of the decisions. As Mark Xu  puts it in his followup post of TAPs for Tutoring, I may not be deciding how my central pattern generator walks, but I will remember having asked it to "do the walking thing." (Okay, he didn't really put it that way, but that's what I got out of his use of that phrase.) Of course it might be less direct than that. I know that I'm asking some part of me to "do the walking thing", and I know that a central pattern generator is involved, and I think I can claim that our understanding of CPGs is sufficient for me to say that my attempt to "walk consciously" will interfere with it, making me more clumsy rather than less. But I strongly suspect that there is at least one additional layer between my consciousness and the CPG: "I" don't talk to it directly. 

  And why am I talking this clumsy way? Well, partly because it's Geulincx' birthday, but mostly because this really is my current view of what I'm working with when I deal with the not-so-conscious parts of me. Of us. Whatever. I would like to develop a good vocabulary for this, and I'm going to try, and in the process I'll be trying to extend my messengers in what I think of as Bayesian directions. But I'm also going to try to learn some specific stuff that simply interests me; I'm going to go on with trivial form-building and try to watch myself learning IPA. (As a programmer I've worked with IPA as adapted for some endangered languages, but I didn't understand the IPA layers any more than I understood the Russian or Japanese translation layers. So I'm curious, and I didn't like being unable to pronounce "Geulincx" by Wikipedia's IPA: https://itinerarium.github.io/phoneme-synthesis/

didn't accept it, although I was able to work it out with

https://www.internationalphoneticalphabet.org/ipa-sounds/ipa-chart-with-sounds/ .) It's a small thing to learn, and worth learning. We'll see. Happy Birthday, Arnold.

"Applied Rationality Training Regime" Review#5

    <  |  ^  |  >

 

 January 30, Review#5. 

   Murphyjitsu with puzzle as form-building toy problem.... a sixth-grade puzzle of filling in a simplified and relabelled periodic table. The puzzle was sent to me, with a separate solution, by a granddaughter who lives a few thousand miles away; so, for me at least, there's the issue of solving it, and there's the much more complex issue of what to say about it. From the end of the day, I see it as a conversation involving Sarah+Jim, my Murphyjitsu team, on both issues; Spider-Duck on the puzzle; Marian on what (not) to say... and my wife, to approve the result. Now I'm wishing I could do it all over again, for "form-building", but I guess what I've spent the last little while on is at least partially just that -- doing it over again in my head, but with better form, with my sub-problems and checklists better organized than they were this morning.  I think I can claim that I'm not just working it as I would have worked it before all this (model "A"), not just working out some self-improvement (model "B"), but working on improving my self-improvement (model "C"), and thus to some extent applying Douglas Engelbart's ABC Model of recursive self-improvement.  Yes, it's his birthday. And yes, I'm aware of the Yudkowsky "Insufficiently Recursive" critique:

1. Engelbart committed the Classic Mistake of AI: underestimating how much cognitive work gets done by hidden algorithms running beneath the surface of introspection, and overestimating what you can do by fiddling with the visible control levers.
2. Engelbart anchored on the way that someone as intelligent as Engelbart would use computers, but there was only one of him - and due to point 1 above, he couldn't use computers to make other people as smart as him.

Those both strike me as legitimate, but OTOH I think that Engelbart's work arguably did start a recursive improvement process in easier/faster creation of each next generation of software leading to a new generation of hardware leading( to opportunities for a new generation of software, and so on. Epistemic status? Well, I'm not confident, but a lot of stuff happens that isn't hidden and without Engelbart, even some of those hidden algorithms would have run quite a bit more slowly. As a graduate student in the UPenn CIS Dept in the late 70s, I think everybody I knew thought Moore's Law would peter out in a while; that came up in discussions of the need for massive parallelism (and proving theorems about massive parallelism, in my case). It didn't peter out. Why did new kinds of circuitry, not just shrunken versions of the old, keep appearing when they were needed? Partly, I think, because interactive software and the mice that helped it squeak did make engineers and even physicists effectively just a bit more intelligent than they would have been without it. I'd really like to have a better sense of how strongly I should believe that, if at all, but it's not something I'm going to work on now. Tomorrow ends the month.

"Applied Rationality Training Regime" Review#4

    <  |  ^  |  >

January 29, Review Day#4 --  I focus on applied rationality as per Anton Chekhov:  "Don't tell me the moon is shining; show me the glint of light on broken glass." (Even if it's rephrased a little.) It's his 160th birthday, and I try to make my TAP-dancing Charlie Checklist as vividly concrete as I can, with TAP-shoes and cane (the cane TAPs on items I'm supposed to remember: this morning at 4, after taking the dog out when I got up, I checked the garage temperature and didn't close the garage  door quietly. My daughter's bedroom is right overhead. Okay, I have practiced closing that door quietly, and an echo of Charlie has started to remind me for other doors as well as that one. (The "echo of Charlie" or "version of Charlie" idea is basically that I can have different Charlies who are themselves evoked as TAPs by triggers like doors, coffee pots, chairs, whatever. I can't handle many, but a few seems okay.))

I am going a bit beyond Mark Xu's ten-times rule for TAP reinforcement in that I try for spaced repetition, and this does help. To make it work better, Charlie has a pocket-watch. This doesn't work well yet. And, at least today, I haven't been asking him for help Sarah and Jim with solving actual problems via Murphyjitsu, even toy ones. Okay, tomorrow I will look at a toy problem, a puzzle from today's email, and I will try to think about "form-building".

 Today is also W.C. Fields' 140th: "Start every day off with a smile, and get it over with." This, too, is applied rationality, but Marian remarks that Fields' misanthropic humor is exactly the kind of distancing that I created her to warn me against. It has its place, I'm not gonna leave it behind entirely, but I do appreciate the warning. Maybe Fields' over-the-top version of it is really all right, some of the time: "I am free of all prejudice -- I hate everyone equally" is certainly an interesting approach to overcoming bias. But I think I'll start tomorrow with a toy problem.

 

"Applied Rationality Training Regime" Review#3

 

   <  |  ^  |  >

 

January 28, Review Day #3, and this morning while walking on an icy road with my daughter I explained that the imaginary cartoon spider that was TAP-dancing on her head had just reminded me of what she'd said earlier on the walk. Through the day it happened again, a couple more times -- Checklist Charlie has actually become somewhat useful. Can I "systemize" on that basis? Maybe, at least partly. Marian also spoke up when I was about to say the wrong thing -- well, I think it would have been the wrong thing. It would be nice to believe that I can become less wrong than I was. Charlie did fail me once, and I think  it was because I was insufficiently real in my cartoon-spider-TAP-dance-visualization. I need a proper mix of fact and fancy... it's not quite enough to say with Kathleen Lonsdale (the crystallographer who flattened benzene, whose birthday it is) that "in science as in the arts, there is very little worth having that does not require the exercise of intuition as well as of intelligence, the use of imagination as well as of information." I've gotta imagine the right things, making the right use of the information.

And it's also Arthur Rubinstein's birthday: "Love life, and life will love you back." This, too, is applied rationality. 

I keep thinking that "Applied rationality" is such an awkward name. I like "selves-control" or "selves-training" better, but I can't imagine anyone else would.

I was just looking at Mental subagent implications for AI Safety:

Assume, for the sake of argument, that breaking a human agent down into distinct mental subagents is a useful and meaningful abstraction. The appearance of human agentic choice arises from the under-the-hood consensus among mental subagents. Each subagent is a product of some simple, specific past experience or human drive. Human behavior arises from the gestalt of subagent behavior. Pretend that subagents are ontologically real. ...Some agents are going to be dis-endorsed by almost every other relevant agent. I think inside you there's probably something like a 2-year-old toddler who just wants everyone to immediately do what you say.

I thought at first "A-ha! That's exactly how I think of it, it's what I've been doing throughout." But now I don't think it's quite right. It's not that your stories, your subagents, are ontologically real so much as it is that they will have been ontologically "real", as real as your overall self. They are, and you are, stories that you tell your selves. Tentatively, I see no reason to believe that more than a small fraction of them are products of simple, specific past experiences or human drives, either. And I don't think this has the same implications for AI safety, but it might. 

"Applied Rationality Training Regime" Review#2

 

   <  |  ^  |  >

  January 27, and I've been trying to go over the methods and adapt them to current issues, with some partial success (maybe). Yesterday, I thought I was going to give Marian, the Noticer of unhelpful humor,  a bunch of named sister-spiders to help in noticing some of my failure-of-social-skill bad habits, but something happened which I remember from trying to write fiction: once you create a character, they've been created and to a considerable extent they say what they want to say. 

  Marian says she'll deal with her sisters, or leave them out of it; she just wants a list of the bad habits with examples she can use for TAPs. She wants these TAPs to be managed as a checklist, and so she asks for a Systemizing spider named Checklist Charlie (I'd thought about defining him on day #10 but hadn't quite settled him in my head; I guess I was too busy being balky.) Charlie starts as an animation of an old shopping-list trick. You can associate numbers with rhyming words one-gun, two-shoe, three-tree, four-door, five-hive, six-sticks and so on, and make a little cartoon-image story for your eggs, butter, milk etc with a gun firing eggs, a shoe filled with butter, a tree whose fruit is milk cartons... whatever. Repeat the list a few times and it really does help memory. Okay, Charlie's standing, TAP-dancing, on two of his legs but he has six for memory and as a cartoon he can pull out more if needed, so number-one is his gun-claw and it holds a TAP with a trigger and the action being the first item on the checklist. That's fairly easy, even when I find that a given trigger (being in the kitchen by the coffee pot) ought to trigger several things. In that case, even though Charlie's not really recursive, he can have a baby checklist spider hanging from that claw. This actually worked this morning -- but in one case I had used the wrong trigger and didn't get triggered until too late. Working on it...

And does this really help Marian? She thinks so, because she thinks my social-skills failures are grouped together, and she's not at all clear about what I said in the  We are having a continuing discussion with Spider-Duck about how that works.

"Applied Rationality Training Regime" Review#1

 

   <  |  ^  |  >

When I go back over the last 25 days of the Mark Xu take on CFAR, I tend to want to connect the tactics to a common story-thread that helps me explain to my selves how they fit together, why they work. I'm looking at the models of the human mind as a city-state (Plato's Republic), or as a mountain-range with scattered villages and a broad central valley  (Scott Alexander), or as an elephant with rider (Haidt), or as an elaborate puppet (perhaps an elephant, why not?) managed by story-threads managed by cartoon spider-swarms grouped in working rooms, cavities within the structure (me)... or maybe as a distorted homunculus corresponding to actual brain circuitry, as mapped by neurosurgeon Wilder Penfield. (It's his birthday.) That homunculus, I believe, is incomplete in the sense that it doesn't include the central pattern generators and other circuits that may be actually outside the brain, but even inside the brain I suspect that it doesn't represent the spider-swarms themselves, but rather the individual low-bandwidth threads, the doorway spiders, that connect each swarm to a central swarm.

  What I'm rather clumsily trying to get at here is that when Haidt wants his rider to train the elephant, he's actually trying to train spider-swarms that he mostly can't get at: he can only get at message-passers. And that, to me, helps make sense of the CFAR emphasis on noticing what you feel --  it's all about taking advantage of the messages from the outer swarms, which I do accept that I've been basically ignoring almost all of for almost all of my life. The communication paths are good enough for my ancestors to have survived, but not good enough for my actions or theirs to have been well-aligned with goals. Should I really even think in terms of goals? Well, they form a convenient shorthand for... something. I'm not sure how good or bad that is, but labels are convenient and I propose to have more of them, not fewer. I'm going to give Marian, my Day#8 Noticer, a bunch of sisters who are (I hope) sensitive to individual kinds of issues; in particular I'm going to ask her to keep them sensing my tendencies to fail in attentiveness, in sensitivity to others' thinking, and in staying on task. I might eventually go to Yudkowsky's 12 virtues, or for that matter Franklin's 13, but I have to start small.

To make this work for me, I guess I have to make them fairly concrete -- as concrete as cartoon spiders need to be; they will need names. Like Sarah Som the Murphyjitsu Sense-O-Meter, they're supposed to be oriented to failures that haven't happened yet but are implicit in our model of what is happening -- that's part of Murphyjitsu, but I'm extending the idea, trying to improve my self-improvement, by asking them, and Sarah, for odds.  I'm going to think of Sergeant Sarah as being in charge of Marian and her sisters, Sarah's not just going to imagine that something has gone wrong, and see if that seems surprising. she's going to assume a hundred copies of the situation and see how many go wrong. 

And will Jim Pright, the pre-hindsight guy that works with Sarah, be similarly extended? I suppose so. I have to work that out, describing it to Spider-Duck. But in general I'm trying to apply rationality after the fashion of another guy with a birthday today: Wayne Gretzky, who says to "skate where the puck is going to be." Live according to your selves' models of what's about to have happened, in the context of an overall selves-model with rather drastic communication limits, but where reprogramming with spaced repetition can generate the TAPs that will, maybe possibly sometimes, get the puppet to move in a useful direction. 

Rationality? Well, maybe. :-)

 

"Applied Rationality Training Regime" #25: Recursive Self-Improvement

   <  |  ^  |  >

January 25, the final frontier (indeed, it's High Frontier Day,  Lagrange's birthday): we close with Training Regime Day 25: Recursive Self-Improvement wherein we plan to improve our rationality/improvement techniques, and apply our rationality/improvement techniques to the plans, on object and meta-levels (well, it's meta- to begin with, I guess.) So I think I should go back, a bit: where am I anyway? 

1. It's all about "winning", about being the pattern you want to be.

2. It's good to have a list of the things that bug you.

3. (tips): Boggle at reality, (semi)commit to working at it, build "form" with trivial problems, push your comfort zone, "seat-adjust" all the tips always, including this one sometimes; focus on your feelings, take no sides on trivia, paradoxes should be paradoctored, win!

(I think I'm thinking of 1--3 as introductory material here. The Real Stuff begins..)

4. Murphyjitsu: plan around failure, or through it.

5. TAPs: build up your reflexes by (spaced) repetition; this includes know-what, the standard memory stuff supported by flashcards, but it also includes how-to, e.g. sewing.

6 Seeking the Sense in other people's points of view

(I'm thinking of 4--6 as basic tools for internal and external plan formation...)

7: Factor goals into actions and vice-versa, Notice your reflexes, seek the Double-Crux for agreement, Systematize yourself into Huxley's clock, Talk to the Socratic Duck. Practice Focusing, cycle your resolve, dump your brain occasionally, expand your comfort zone. Ask yourself the Important Questions, ...

Is this as random as it seems? Maybe so, and that's not necessarily bad, but I'm going to start over at this point, and go back to What is This Self That "I" Am Trying to Improve? I've spent parts of the day going over this, and I think I'll have to give myself a little more time. Well, I was originally assuming that this was a 30-day plan, anyway, and now I think I'll be continuing it for quite a while after January ends. I will be doing resolve cycles, but my seat-adjustment for them is that I'm not forming my expectations of what I can do in five minutes (or whatever) by self-deception; I'll form them by repetition, trying to work as intensely as I did last time  (or sometimes a little better). Just like physical exercises. I will be naming and consulting more specific story-spiders; I'm not sure that Spider-Duck needs a name, but he has certainly been developing a personality. And I seem to be going back to the physical index cards that used to be a central part of my approach to almost everything. Learning how to learn how to learn... changing how to change how to change.... I'm definitely not done with this sequence, but there are no more numbered days of the Xu sequence, so I'll have to go on a bit without them. Robert Boyle apparently said that "God would not have made the universe as it is, unless he intended us to understand it" and it's his birthday too.

"Applied Rationality Training Regime" #24: Resolve Cycles 2

   <  |  ^  |  >

January 24th, getting  quite near the end of our not-quite-30-days, is
Training Regime Day 24: Resolve Cycles 2 and this I find to be even more odd and awkward than the original Day #13: Resolve Cycles. The idea is to push yourself really hard to solve a problem really fast, and in particular you're supposed to "to figure out a way to reliably generate expectation for yourself.... Tell your friend [a likely falsehood, or...] Pretend..." And he cautions that "A caveat with this technique is that if you can't solve the problem after you've expected yourself to be able to solve it, then it hurts..." 

   My problem is not the expectation of hurt but the loss of internal credibility; if my internal selves are lying to each other, then nothing much makes sense to me. So, I describe this issue to Spider-Duck: he points out that seat-adjustment has been needed for just about everything else, so why not this? And I ask my murphyjitsu team if they can tell how this kind of resolution problem-solving could have gone right for me, instead of how it will have gone wrong, and we get distracted with thirty-year-old memories of problems solved by resolution in theorem-proving, but it seems that Sarah Som has a probable sister. Serena wanders in possible worlds, so that if there's a chance that things go right then Serena might be able to sense it, and I might be able to do what would have worked in that world, and my selves might even be able to work as hard as the selves in that world, too. Maybe the consideration of possible selves is a way to generate expectations, as Mark (and presumably CFAR) recommend. Mostly I prefer to generate expectations based on prior experience and on taking the external view, mostly I'm more interested in guarding against expectations that are too high, not in pushing them higher than the evidence suggests, but I guess I can expect my selves to push as if they expected everything to go right. After all, some of them do.

   Hmmm....this doesn't make a whole lot of sense, but I think I can try it. But I also think that I'm deviating more and more from the "proper" (?) approach. I'm actually wondering if trying too hard isn't a serious issue here. Maybe because of Smullyan's sage, back towards the beginning, but maybe because today is Edith Wharton's 159th birthday, and "If only we'd stop trying to be happy, we'd have a pretty good time" seems like an excellent application of rationality to me. But yes, I'll try to try to do better -- specifically with study methods. I think I can solve the problem of getting myself to study and remember, using what I already knew plus what I've been doing for these 24 days past.

"Applied Rationality Training Regime" #23: TAPs 2

 

  <  |  ^  |  >

 

January 23 is Training Regime Day 23: TAPs 2 and it's an odd one; I'm a little bit confused. This is specifically a "Noticing" TAP from Day 8 where Mark said:

Essentially, you want to install the TAP 

"feel <the thing> -> <do some action>".

In this case, the "thing" is "something's wrong with the world" and the "action" is "fix it". I don't know how to treat that as a TAP, which I've been viewing as a more or less atomic action on its own. It's not specific, which was a big piece of what being a TAP was all about. I hereby do a seat adjustment: for each irritant, I will draw a mental circle to represent a story-spider for handling that irritant, and I will try to attach it to thread-lines for input and output. 

  Mark is talking about his examples being "dumb-sounding" little stuff. I'm not sure how the Pareto Principle does apply: maybe the 20% of irritants that cause 80% of the irritation will turn out to coincide with the dumbest ones? Whatever; it seems to me that paragraph would be better rewritten with a Mustachian point of view. As Mr. Money Mustache says in Eliminate Short-Termitis, the Bankruptcy Disease:

As a future young millionaire, you need to start thinking about all of your purchases as LONG-TERM events, not short-term ones. That means each decision should be carried forward in your mind for at least 10 years

   (Hmm, he said that about ten years ago.) So each of these minor irritants should likely be ignored if they only happen once but should be considered as larger if yearly, much larger if monthly, drastically larger if weekly, and so on. These story-spiders only warrant concern, they only get names, if they keep coming back.

   But -- if I'm interpreting the "fix it" mentality correctly -- this reflexive Notice/try-fix is what I've always done, to the point of being annoying to a couple of my family members. In my own notes on Day 8: Noticing, I mentioned trying to commission Marian, a story-spider, to notice when I was taking something serious as a joke and bring it to my conscious attention... but this was not the only flaw of the day. Marian is also supposed to remind me that often, I should Stop Trying to Fix Things, Just Listen! 

   And why, you ask, should I think that the same Noticer should remind me (1) to take things seriously rather than as an opportunity for wordplay, and (2) to sympathize rather than trying to identify and solve a problem? I'm glad you asked that. It's because both of these create a sense of distance, which may not be intended as defensiveness but certainly serves for defense, and I now visualize Marian as giving me a shout (a rather squeaky spider-shout) when a thread stretches. Marian is in fact coming up now, on day 23, because she thinks of herself as helping me address a problem that is not very important once or twice or three times, but that repeated over and over again for decades does get to be a big deal. Any tiny little thing that's repeated enough times will be a big problem or a big opportunity or both... As Fred Morrison (whose birthday is today) put it, "you could buy a cake pan for five cents, and if people on the beach were willing to pay a quarter for it, well—there was a business." A business which turned into the Frisbee I got as a graduate student at IJCAI-77... 

  But I don't think this day is about new tactics, it's about rethinking what I already do.

"Applied Rationality Training Regime" #22: Murphyjitsu 2

  <  |  ^  |  >

 

January 22, we go on to Training Regime Day 22: Murphyjitsu 2 with the goal of raising Murphyjitsu to a higher level...figure that the Universe is against you, the laws of probability are being warped to halt your plans, and work out a plan to win anyway. 

umm.... Huh?

I admit this is one of the points where I would be asking an instructor for examples or other hints, because when I asked Sarah Som and Jim Pright, they looked as blank as spiders can look. And then I thought I'd try something really small. I'd try eating a piece of cake without making a mess, okay? It so happens that I like the texture that other people call dried-out, so there's even more of a risk of crumbs than usual, so I took a piece and actually cut it into pieces before putting it on to the plate. Then I thought about the fork I was about to get, and Sarah and Jim thought that if I stuck a fork in a piece it would split, and if I lifted a piece on the fork it would slide off. So I picked a toothpick out of the drawer beside the stove, and decided to eat my pieces of cake with that. Piece of cake, right? Ummmm.... the toothpick, unlike any other toothpick pulled from that drawer, was cracked. It broke. And Sarah and Jim, watching me from inside my head, pulled in their legs and rolled around and around and around, laughing at me. Yes, this really happened, this morning. I am not making it up. 

  So this evening I decided to have another piece, taking no precautions at all except to hold it in a napkin; this generated a wifely protest, but nothing went wrong. Nothing. The empirical result is clear: the gods were not bothering to be against me, unless I take precautions. Since today is the birthday of Francis Bacon, father of empiricism in science, I cannot go against the empirical result. I will bow to the gods and have some more precaution-free, anti-murphyjitsu, cake tomorrow.

"Applied Rationality Training Regime" #21: Executing Intentions

   <  |  ^  |  >

 

January 21 brings us to Training Regime Day 21: Executing Intentions and that's all there is to it: just form an intention and execute it. Do that again. And again? Sure. That would be a "form-execute" cycle, sort of like a computer's fetch-execute cycle, especially since my understanding is that human memory is active and actually forms a memory rather than just fetching it. I take this as an exercise in mindfulness, trying to make a TAP more accessible to other parts of your mind, other selves and other stories. I've spent parts of today trying to do it, and it's doable.

It's not cost-free; if I try to walk along, making each step into a conscious intention-execution, I find that I can do it to a very limited extent, but it doesn't work well. If I try to exercise with dumbbells as I walk, as I do every morning, I'm abruptly unable to move properly with arms or  legs; I can think my body through the motion a little bit at a time, but it takes at least a couple of seconds per step. I suppose I'm interfering with one or more central pattern generators, circuits which work just fine as long as my conscious/verbalizing selves don't try to talk to them while they're doing their jobs. In this case, perhaps the highest application of rationality is to stay out of the way. Still, the CPGs are obviously to some extent trainable, and learning to walk (with or without rhythmic arm-swings) is precisely such training, and if all I can lift into my consciousness is the difficulty of this communication then I really ought to do that. 

It would certainly be nice to have a map of all the CPGs that thus facilitate and block my processing, an atlas of the whole brain (and a bit more than just the brain?). If I Google for "brain atlas" I find the first match is the "Allen Institute", and it just happens (you guessed it, right?) that today is Paul Allen's birthday. "You have a certain number of dreams in your life that you want to fulfill." Happy birthday...and happy dreams.

"Applied Rationality Training Regime" #20: OODA Loop

 

  <  |  ^  |  >

January 20 brings us to the Training Regime Day 20: OODA Loop. The OODA loop idea is adapted by lots of people in lots of ways; I'm going to adapt Mark Xu's adaptation of the CFAR adaptation of Colonel Boyd's original (or whatever intermediate form CFAR was working from). Let's try a textual diagram:

1.Observe   [update model of world-region]

      => O-O transition =>     [*end updating world model]

          2.Orient [update model of available operations]

              => O-D  transition =>   [*end updating ops/self model]

                  3.Decide     [update model of choices -- What I Do Next]

                       =>   D-A transition =>      [*end updating choices model]

                             4.ACT   [update world]

      <= A-O transition <=  [*end action, back to 1.Observe]

It's easier to type it this way, but actually I visualize these on a clock-face with Observe at 12, Orient at 3, and so on. Each one might take a fraction of a second, or years. I think of each of the stages as having its own heuristics for making pretty-good guesses, and each transition as having heuristics for ending the corresponding stage, saying "never mind perfection, we've got to move on." If you use bad heuristics, that's your failure in an OODA stage or transition; if you use good heuristics and fail anyway, that's.... well, that's too bad. Note that 4.ACT can make changes in the outside world or in your internal structures; the other stages and transitions are internal.

  Consider an example: I'm going through this "Training Regime" as an OODA loop in which each day I try to update some part of my world-model to understand the day's exercise, update my self-model of what actions I can take, update my model of my actual choices (i.e., make one), and then write something.  Do my heuristics make sense? A sub-example: One peculiar heuristic that I don't see other people using is that I often throw in some random alternate-point-of-view boggling by looking up whose birthday I'm at. Today is the birthday of Buzz Aldrin, the second man to be able to say (as he does, dunno if Armstrong ever did) "Shoot for the moon; you might get there." That's an interesting heuristic about the Orient phase of a loop. Come to think of it, there's a meta-heuristic about that, something connecting identity with abilities, implicit in a quote from DeForest Kelley: "Dammit, Jim, I'm an actor, not a doctor." Something like that. It's his birthday too; he'd be 101. I'm not going anywhere in particular with these, but they do make me think about the stages in a different light, forced by random input.

  I suppose that every sequence of conscious actions (or actions which might be raised to consciousness by Noticing) can be usefully viewed as an OODA loop, even if one or more of the stages/transitions may have been made automatic, i.e. made into TAP lines.

  Indeed, consider a "TAP loop" -- a phrase I may have just now invented, but surely the idea is obvious: the pattern you detect (in yourselves or in your surroundings) triggers an action which changes the pattern to something new which triggers an action which changes the pattern.... I hereby claim that a TAP loop is an OODA loop, stripped to the bone by pre-computing as much as possible with a set of simplifying assumptions which, if true, imply the appropriateness of the Action associated with any given Trigger.  Right? Observe, Orient, Decide all get folded together into a single pattern-recognition, and of course under some circumstances that'll work. Of course, this morning as I was going through the TAP-loop that gets my breakfast ready, I happened to have a side-thought about cleanup while I was still on preparation, which led to my putting the granola away before I'd added a bit on top of my chopped apple with yogurt. TAP loops are fragile, fully-worked-out OODA loops less so, but working out the stages is a lot more work than just following a TAP.  

 OODA loops can obviously be nested: any of the four parts can and likely will be a composite development, as when we start a loop of building up world-model information for a decision.  OODA loops can obviously overlap, as when a sequence of discussions happens to overlap with a sequence of meals, each with its own observations, orientations, etc. I think that I'm gonna change the way I diagram events in my head: I'm gonna try to add the clockwise-circle-arrow symbol, U+2941: ⥁Maybe that'll remind me to look for the OODA loops within and surrounding each human (or AI) action I see. 

  (And as yet another example, this particular January 20th has seen a presidential transition, and all of us have to Observe and Orient and Decide and Act once more.... I did vote for this outcome, but I can't claim to be happy about it. It's less bad than the alternative, at least in the short and medium runs, but my 2016 hope that the Blue Tribe would take up the banner of constitutional limits seems to have been bad modelling on my part. So it goes.)

 

"Applied Rationality Training Regime" #19: Hamming Questions for Potted Plants

  <  |  ^  |  >

January 19th, Training Regime Day 19: Hamming Questions for Potted Plants. Here the exercise is to imagine yourself as a potted plant; to vegetate with pot! (I haven't done that for almost fifty years.) 

  Oh-kay, it's a reasonable metaphor -- maybe really an analogy? Anyway, it's a pattern which you can try to fit your selves within, and I take it that it's really, underneath, an exercise in boggling: to look at your life, your stories, your selves, with fresh eyes. Think through each aspect of being a plant, and consider what it's a metaphor for in your own actual life. More concretely, make a diagram of circles and arrows, or spiders and legs and threads, connecting the crucial factors in the plant's existence -- and then relabel them as crucial factors in your own. Can do. In the icy little park where I sat this morning on an insulated folding pad and had my cappuccino and croissant, I found myself looking at some plants in giant pots, waiting for spring.... but before that I thought about the way Xu's notes think of fruit. If you want a potted plant to produce fruit, you need to worry about how much sunlight it gets because 6CO2 + 6H20 => C6H12O6 needs light to drive it... and if you're thinking about potted plants that produce fruit, then you're talking about strawberries, avocados, (strawberries are not berries but avocados are) and maybe a few others.... think about companion plants that help strawberries grow, like bush beans, and similar issues. Growth and productivity are all about setting up the environment. Maybe. Here I plant myself, I can do no other. But I guess I'd rather think about flowers, specifically the hybrid tea rose called Rosa 'Dolly Parton':

Sun: Full sun

Water: Medium

Maintenance: Medium

Flower: Showy, Fragrant

Attracts: Butterflies

Other: Thorns

 

It's fun to muddle metaphors, isn't it? And it's Dolly Parton's birthday.

 

"Applied Rationality Training Regime" #18: Negative Visualization

  <  |  ^  |  >

January 18, "Training Regime Day 18: Negative Visualization" and this one is all about Murphy without the jitsu; visualize that things are uncertain, so things go wrong, and then we go on from there. If we still can, says a story-spider in the back of my head as it points a leg at "Unfriendly AI Takes Over!" and then "Tiny Black Hole Falls Into Earth's Core!" and finally at "Heat-Death of the Universe!" But no, this  Training Regime item is about smaller failures; the failures we can in fact go on from.

  So, Mark Xu describes two TAP-arcs. The first is from consciousness of uncertainty, that things might go badly, to a pre-imagined consciousness of things going badly; failure. The second goes directly from that imagined failure as trigger, on to imagining whatever comes next. We go on.

And what good is this? It's an exercise in self-control, a response-override. Without it, your reaction to failure will probably be dominated by negative emotions even if the situation doesn't prompt the traditional "When in danger or in doubt, Run in circles, scream and shout". Those emotions would make it difficult to choose a good path, or even a least-bad path, going on from failure; better to override them. But even if there were no practical value in overriding negative emotions, they're still negative, not fun, and best replaced. So... I ask my Murphyjitsu team if they can deal with this for me, and Sarah Som (Sense-O-Meter) can certainly imagine things going badly in any case where they might do so, but Jim says he'd rather stay specialized on pre-hindsight. His cousin Peter, however, knows the same model situation and says that yes, it's pretty depressing, but he'll project it forwards and just accept the outcome, making lists of all kinds of stuff to manage his feelings. (It's Roget's birthday.) Okay; that does sound like a use of stoicism, sort of the way Spinoza did in saying "Do not weep. Do not wax indignant. Understand." 

  If things go well, we go on from there. If things go badly, we go on from there. Yes, I can imagine that, and have practiced. So it goes. (And among the things Billy Pilgrim could not change were the past, the present, and the future.)

"Applied Rationality Training Regime" #17: Deflinching and Lines of Retreat

  <  |  ^  |  >

January 17 brings me to Xu's Training Regime Day 17: Deflinching and Lines of Retreat and Trying Not To Be Offended, indeed more generally Trying Not To React Emotionally to What People Say To/About You. And it seemed to me that it was reminding me of something, and then it clicked... long ago I had read Ben Franklin's story of cycling through the "virtues" and trying to, umm, apply rationality to the general goal of aligning his behavior with his goals, and #9 is

Moderation. Avoid extremes; forbear resenting injuries so much as you think they deserve. 

I'm pretty sure that I thought of it because I knew today is Franklin's birthday, but I wonder if his virtue-training course was in some sense an ancestor of this one. And I found myself imagining combining the two, and then more generally cycling through lists of tactics on the one hand and lists of goals on the other. Imagine 6 cards with Haidt's moral foundations, and 13 cards with Franklin's virtues; they are relatively prime, so if you just cycle through each, as a morning cue for contemplation, you'll get all 78 pairs and then you'll repeat. That's not quite the way I'd do it, but it gives an idea.

Meanwhile, the "deflinching and lines of retreat" didn't seem to be an issue when I tried, and that may be because I already have my lines of retreat more or less as Xu describes, but as usual the seat is adjusted. Imagine yourself as a cartoon story-spider, the Anansi head of a story-spider-swarm; imagine the person you're talking to the same way. When people say things in my hearing, that updates my map of the part of my world that includes them; it gives a fact about them. I'm not very good at making this out as a fact about me, and I guess that's an advantage as well as being always a disadvantage... it's easy for people to say things that make me feel puzzled or sad or oh-so-very-tired-of-not-being-able-to-communicate, but it seems harder for me to be "offended". This, too, I remember from high school if not earlier. So my relationship to deflinching is a little off, I guess. My main reaction is a set of thoughts about diagrams for these relationships, and then maybe diagrams for earlier training-regime tactics. Maybe I'll be able to write something about those; maybe not. But happy 315th birthday to Ben, anyway.

"Applied Rationality Training Regime" #16: Hamming Questions

    < | ^ | > 

 January 16. I look at Training Regime Day 16: Hamming Questions and see that Hamming, of the eponymous error-correcting codes, had a habit of asking Hamming Questions:

"What are the important questions in your field?" (and later: "Why aren't you working on them?") 

 And suddenly the year is 1981, my PhuD on "Infinite Structures in Programming Languages" is less than a year old, I'm teaching at the UDel CIS Dept which is very much in crisis mode and the guy asking me the basic Hamming Question is Bob Caviness, newly arrived as department chair.  Umm... okay, now I know it can be called a Hamming Question. Back to 2021. Mark Xu's take on CFAR's take is that at this point we ask what are the important questions in our lives and why aren't we working on them? And of course, to some extent I think I am, right now: the basic prerequisite for answering questions correctly is (some version of) rationality, and the basic prerequisite for working successfully on a project is (some version of) applied rationality. So I'm okay with that. I was going to write something about the questions I think are important in my life, but not now; maybe if I come back to this post and update it, which I'm now thinking about doing systematically with this sequence.

 (And back to remembering 1981, at the beginning of which I'd learned that the compilers course I was about to teach had last been taught by a guy using fake ID and of course fake credentials, with different names at the different places where he taught. So, rather than bring in my actually relevant PhD diploma which hadn't physically arrived yet [I'd spent a full day in December back in Philadelphia printing out pieces of dissertation in between computer crashes, shoving most of the pages through an IBM Selectric twice with a change of typeball in between]  I showed my students my Doctor of Zoology diploma, from Epsom Chapel Nursery School & Kindergarten, 1958.)  

  And today is Bruce Schneier's birthday: Bruce Schneier knows Alice and Bob's shared secret.

"Applied Rationality Training Regime" #15: CoZE

   < | ^ | > 

January 15th is of course Training Regime Day 15: CoZE for "expanding your comfort zone" by listing the things that delimit it (in my case already mentioned, social issues and manual dexterity) and doing something that goes beyond them. Well, what with lockdown and all that I'm not gonna do much to stretch myself socially, so I was thinking I'd sew, but I wound up with furniture assembly instead. Or maybe as well; we'll see what happens tomorrow. 

  One "seat-adjustment"; I find it helpful to laugh at myself. I do hope that some future tips and tricks will talk about the rationality-value of comedy.  Today's the birthday (actually baptismal day, day of birth unknown) of the comic playwright Molière...I did read some of his stuff in college, though I admit it didn't make much of an impression. "Of all the noises known to man, opera is the most expensive."

"Applied Rationality Training Regime" #14: Traffic Jams

   < | ^ | >

January 14 brings us to Training Regime Day 14: Traffic Jams with the notion that if the parts of your mind are focusing on too many different things that you'd really like to achieve, then that in itself will tend to block productive goal-oriented activity, so what you need is a Brain Dump in which you write down everything you need to do, then "clarify" by putting little things on a ToDo list to simply get done, medium-sized things get plans and Murphyjitsu, the biggest things may get pruned, or simply cut down and legitimately grieved over. Some aspirations, some story-lines, just aren't plausible and you may need to give them up in order to go on with your life, but it doesn't help to try to deny that it makes some of your selves really sad. The Litany of Tarski applies, especially because it's Tarski's birthday.

My brain dump looks an awful lot like my bugs list; in fact I'm viewing this exercise as being mainly a way to deal with my bugs list. Okay, that's interesting, and the part about grieving as part of giving up an unreachable goal is extremely interesting even if I'm not doing that today. I have tried just a little bit of Murphyjitsu on subplans for a bigger plan, and done a couple of small things. Good enough for today.

 

"Applied Rationality Training Regime" #13: Resolve Cycles

  < | ^ | >

January 13, time for Training Regime Day 13: Resolve Cycles where the idea is kinda the goal-seeking version of high-intensity interval training: you push hard for a short time, in this case 5 minutes by timer, to decide how you're going to handle a bug; then you push hard for 5 more minutes to solve it. Really? I wasn't at all sure how to interpret this, and found it helpful to read a post by Owen Shen connecting this enhancing-your-resolve,on-a-cyclic-basis CFAR tactic with two other (CFAR? Rationality community?) labels

[1] most people are not Actually Trying in life. It seems to me that, often,
when people are doing a task, they are looking for excuses to stop,
rather than reasons to continue.... [2]Taking Time is a related idea, where you’re, um, actually taking time to do things. Where Actually Trying is about powering through, Actually Taking Time is about sitting down and doing things right....formalism...ritual...

 

 But maybe not helpful enough. Solve it? Never mind. Okay, having written this in one 22-minute Pomodoro, shut up and set the timer.... Go!

 

[Later that day...] Well, I'm wondering what size of problem is supposed to be actually solved in five minutes. I really don't get it. I have made progress, it was a useful way to make progress, but solved? And I should Resolve to approach the problem as if my life depended on solving it in five minutes? I need some examples of what actually one solves this way. Oh well, I will use a less ambitious version of the Resolve Cycle for the moment. It's quite true that I can push myself harder for a five-minute timer than I can for a Pomodoro timer, and that's a good thing; a story worth telling pretty often. Not a wasted lesson. Meanwhile, happy birthday to Clark Ashton Smith, the sort of story-teller to whom H.P. Lovecraft wrote fan mail -- which started their friendship, lasting until Lovecraft's death. But as Lovecraft put it. "And with strange aeons even death may die." I think I need a strange-aeons timer for solving Big Problems.

"Applied Rationality Training Regime" #12: Focusing

  < | ^ | >

On January 12 we naturally do Training Regime Day 12: Focusing which naturally relates rather strongly to Noticing, but also to several others;  in fact what I've been noticing, and am now focusing on, is a theme which is leading to a systematic revision of my mental model of my mental modeling, based most directly on Scott Alexander's Mental Mountains, about Unlocking the Emotional Brain:

UtEB’s brain is a mountainous landscape, with fertile valleys separated by towering peaks. Some memories (or pieces of your predictive model, or whatever) live in each valley. But they can’t talk to each other. The passes are narrow and treacherous. They go on believing their own thing, unconstrained by conclusions reached elsewhere.

Consciousness is a capital city on a wide plain. When it needs the information stored in a particular valley, it sends messengers over the passes. These messengers are good enough, but they carry letters, not weighty tomes. Their bandwidth is atrocious; often they can only convey what the valley-dwellers think, and not why. And if a valley gets something wrong, lapses into heresy, as often as not the messengers can’t bring the kind of information that might change their mind.

 I see that as key to Noticing as well as Murphyjitsu and TAP itself, as well as Focusing; in each case, we have an effort to make better use of limited communication between the "conscious self" and the processors up there in the little valleys. And I'm going to go on thinking in terms of story-spider swarms in Haidt's elephant visualized as an enormous inner-spider-managed puppet, but now I'm focused (Hmm....) on the threads that go from one spider-swarm to another. I do not want to think of consciousness as "a capital city on a wide plain"; not a city at all, but maybe a single haunted house with many rooms...There ought to be a fairy-tale in there somewhere, it's Charles Perrault's birthday. (And he said: "I myself am a labyrinth, where one easily gets lost." Me too, Charles.)  In fact I'm just going to to stay with consciousness as a story, an output of edits which try to retroactively reconcile the multiple narratives being constructed by separate spider-swarms. And there may or may not be only one Anansi-swarm to do the final edit to construct the Authorized Edition at any given moment; I don't think that's crucial to my thinking about consciousness. What is crucial is that a whole lot of applied rationality, of getting your actions to be better aligned with your goals etc, is a matter of getting messages to travel along those threads, with very limited messengers who need specific attention. Messengers like Sarah for Noticing's Sense-O-Meter, Marian for Murphyjitsu, and maybe even Spider-Duck.

  The special role of Focusing is to deal with the case where the threads of communication are not simply thin but misconnected: the case where some processing in a non-central swarm is giving you a headache or some other indicator of tension/pain/whatever, and you need to notice that. So sit comfortably and let your body talk to you...and when you get a feeling of something off, something that might be a misconnected story-thread you need to hear, try first simply to give it a label or combination of labels. 

Okay, as so often, I'll be trying this again. I get nothing nohow. Happy Birthday, Charles. (And Jeff Bezos.)

"Applied Rationality Training Regime" #11: Socratic Ducking

 < | ^ | >

 January 11, so we look at Training Regime Day 11: Socratic Ducking and now the story is all about dialog. As a programmer, I'm familiar with the idea of debugging by telling it to the duck, and there's a (toy) duck right here by my desk. I've also applied "explain it to the duck" as a rule for learning...(adapted by somebody, sometime, from the standard "Docendo discimus - by teaching, we learn".) I really feel embarrassed that I've never generalized this to trying to debug the sort of issues identified as "bugs" on Day #2: Searching for Bugs.  Okay, for this I want a different duck, a symbolic story-duck for handling story-threads, so that would just have to be Spider-Duck, wouldn't it? Can I construct dialogue for me explaining the bugs in my life, and him occasionally asking questions? (Why not? As long as he doesn't have to come up with answers...)

So, SD and I have been talking about disorder on my desk and elsewhere, and ways that I've tried to combat it, and why they fail and what I should do instead. So far we haven't moved past the obvious -- stuff like the obvious relevance of today's being NATIONAL CLEAN YOUR DESK DAY. Still, we will come back to it again and again; I think I can commit to that, because it's mildly fun (takes very little willpower) to keep coming back to it. Actually doing something is another matter. I haven't accomplished much, except that while I was thinking of variations on Spider-Duck, he asked why not try variations on me? Which actually might make sense. Or not. Am I really able to decide which self to walk into, which story of myself to tell? Dunno. Meanwhile, happy birthday to... what's your name, man? (And maybe who tells your story is the right way, one of the right ways, to look at choice-making generally.)

"Applied Rationality Training Regime" #10: Systemization

< | ^ | >

January 10, Mark Xu's "Training Regime Day 10: Systemization" and we try to become more systematic about the choices we make, in fact to keep them from being choices at all, along the lines of the Alfred North Whitehead quote I started with on day 1:

Civilization advances by extending the number of important operations which we can perform without thinking about them. Operations of thought ... are strictly limited in number....

And yet this may not be along the lines of Smullyan's Sage and Huxley's clock-self, because it's not clear that choices are being removed altogether. Instead they are being lifted into meta-choices, where we choose spools or skeins of story thread, then let them unwind. It's the choice to live by repeating calendar entries and checklists.  Yes, it's a good idea; yes, I have a negative reaction to parts of it, probably because I've had the experience of trying it but not actually saving time, or saving some time but at the cost of extra willpower because it was hard to do and kept on being hard to do. Maybe that's all there is to it. Very likely, it's the same balkiness that some of my selves feel with commitment, as I said in #3:

I think there are parts of me that tend to confuse commitment with loss of autonomy.

Okay, I'm looking over the bullet journals which I kept for a few years but had abandoned. I've been thinking what parts of that I should resurrect, or extend. We'll see. 

Meanwhile, happy birthday to Lord Action; "power corrupts". But also, "Liberty is not the power of doing what we like, but the right of being able to do what we ought." And maybe that can be true, and maybe it's the right way to see commitment to Do The Right Thing. Maybe. Or then again, maybe not. We'll see.

"Applied Rationality Training Regime" #9: Double-Crux

 

< | ^ | >

January 9 brings us to "Training Regime Day 9: Double-Crux" which startled me considerably; I'd read some other essays which talked about this technique, and now I think I know why I had a problem with them. 

  An argument (which may be embedded in something more complicated) is a sequence of statements where each is supposedly justified, whether because it's just plain obvious ( = an axiom) or because it follows from previous statements ( = an inference.) If you accept (the conclusion of) an argument, the crux of that argument for you is simply a crucial point...one for which you're aware that you'd change your mind if that point's justification failed. If Joe doesn't believe your argument, then there are one or more specific points you disagree at. The double-crux claim is that you and Joe should normally be able to find some one point which is crucial to both of you: a double-crux. (It might be a conjunction: (the sun is a fried egg AND 2+2=17)). If you can convince him it's true, he'll change his mind about the overall argument, but if he can convince you it's false, then you'll change yours.  Something like that; a point on which to focus. And with a little luck, it'll be something you can look up or otherwise settle easily. Without that luck, at least you'll have identified the disagreement, which is progress anyway. You can start looking for parts of the crux to serve as sub-cruxes.

  Mark suggests two kinds of crux-searching. The first, "proposer/listener", has you taking your argument and thinking which steps are cruxes for you...and you propose these to Joe, to see if they are crucial for him too. If nothing comes up, swap roles and try again. The second crux-search, "model unpacking", may involve exactly the same argument-steps, but the change in focus is total. Instead of thinking of a proposition as coming directly from data, assertions about the world, think of it as reflecting a model which is influenced by that data. Instead of thinking about the argument as a sequence of justifications, think of it as a traversal, an exposition, of your underlying model. The crucial statements are those which expose crucial aspects of that model -- so that if Joe accepted them he'd accept the conclusion.

  I'm not sure how much difference it should make in practice, but over the years I have moved away from thinking about arguments as step-by-step justifications. Perhaps it comes from actually teaching proof techniques in computer science classes long ago, or perhaps it comes from further back ...  My wife says some of our college seminar classmates were really annoyed that they couldn't tell what I actually supported, since I would argue pretty much any side of anything. But I was stunned by one who said "I just don't feel the need to believe the logical consequences of my beliefs." I'm not sure how double-crux would have found something to work with there. 

  Nowadays, I tend to think of almost everything through models, models which I express to myself, when I want to get explicit, as Bayesian bookshelves where each model is an imaginary book whose width is the probability I assign it. Evidence makes them thicker or thinner...it doesn't generally make them go away. There might be a crux somewhere, but it's not the focus of the enterprise, it's not the way I generally try to think. So I've been thinking, today, about how to unpack models; thinking about models that I've depended on for a lifetime, and about models that have changed a lot quite recently. My views on minimum wage have changed sharply...but not because I've rejected some set of previously convincing-to-me arguments or even changed mechanisms in the model, but rather as a function of gradual and continuing changes in my models of the prospects for AI alignment. Tentative and indirect, but it seems to have become the way I think.Changes happen, yes, but are cruxes really the right way to think about it? This is not clear to me, except in a trivial sense. Well, I'll go on thinking.

Changes happen; today is National Balloon Ascension Day, and it's Karel Čapek's birthday. We keep dreaming up new tech, and the dreams keep coming true. Happy birthday, Karel.

"Applied Rationality Training Regime" #8: Noticing

 

 < | ^ | >

  On January 8, on to Training Regime Day 8: Noticing. The idea once more seems to be to bring some of your unconscious mechanisms up towards conscious detection (so you can then adapt to them, and perhaps control them.) In particular, try to practice a TAP where the trigger is a feeling -- in Xu's case, a feeling of defensiveness -- and the corresponding action is just something that's easy to notice -- in Xu's case, rubbing the fingers of a hand together. By doing that, he turns his defensiveness into something he can work with or work against. 

  So, I'll try to ask a story-spider to help me literally bite my tongue, gently, any time I'm starting to treat something serious as a joke.  I do that so often that it may be realistic to practice as a TAP, and I don't intend to try to stop altogether but I know that sometimes I do it annoyingly. 

   It has just been pointed out to me, in connection with Mark Xu's account, that treating something serious as a joke  could even be a form of defensiveness... I'm not sure that's quite right, and I suspect that it's most annoying when I really do think that the Creator is showing Her sense of humor. "Man plans, and God laughs." Still, it is a defense mechanism of a sort: humor creates distance.

   Anyway, I want this to work, so I bring to bear any associations I can to strengthen the TAP. I have  named this particular story-spider Marian, after the "intelligent and resourceful" secret-uncoverer Marian Halcombe of Wilkie Collins' 1859 mystery novel Woman in White, After all, it's his birthday. And she's not just any old cartoon spider, the way most of them are: she's a very striking-looking Friula Wallacei, of which only one specimen is known, collected by Alfred Russel Wallace in the 1850s. It's his birthday too. And she's a very mysteriously bright lady, because it's also the birthday of Stephen Hawking.  (And the birthday -- the second birthday -- of my great-niece.) Is this kind of linkage useful? I think it may be. I will keep trying, and I'll keep asking Marian to react to possibly-inappropriate humor.  Meanwhile, happy birthday to all of them, and Soupy Sales too.

"Applied Rationality Training Regime" #7: Goal Factoring

 < | ^ | >

And on we go to Training Regime Day 7: Goal Factoring. After reading through this, and the comments, I rather wish that it had come before #6; it seems to be a methodology for seeking sense in your own actions. (The phrase "goal factoring" led me to expect that goals would be factored into subgoals...this is primarily about factoring actions into goals.) I will of course modify this as I go along...in fact I'm going to add attributes to the recommended diagram, creating an over-the-top version which I will then probably do parts of.

  1. Name an action, preferably one that you do regularly or that consumes a lot of effort/time; put it in a small circle that's gonna have lines going out. For me, that's a story-spider... "Exercise", "StudyRationality", "ReadAStory", "Eat", "Sleep"... "DrawAStorySpider".

  2. Add a leg to the spider, with a label that's a goal helped or hindered right legs positive, left leg negative) by the action. Use a dashed-line leg "- - -" for a weak connection, a plain line "-----" for a normal connection, a double-line "===" for a strong connection. Use upper-case for IMPORTANT_GOALS and punctuation for !VERY_IMPORTANT! or for ?tentative?.

  3. Repeat #2 until it seems no further goals are involved. At this point, imagine that all the right-leg goals are fulfilled, and/or all the left-leg goals are unimportant; does the action still feel relevant? If so, go back to #2 anyway; there's a goal somewhere. More legs!

  4. Copy the spider all over again, onto an index card; this time put the legs in order of goal scale, with strictly personal goals at the top and global goals at the bottom, with family and then community in between. 

   The length of each leg indicates the time-scale that you're thinking about: for "Exercise" I have a very short double-line left leg that just says weHatesItWeHatesIt because I'm really not an exercise enthusiast at all, but the ?timeLost? leg is tentative because I'm not sure that the net effect really is time lost. There is one more fairly short left leg: a plain-line Willpower leg, because on many days it really is hard to get going or even to keep going. On the other leg --- well, actually there are several right legs. There's a  short single-line "?feelBetter?". There's a medium length double-line right leg "NO_HEADACHES!" which is a major discovery I made some years ago about the headaches that had plagued me since adolescence and which come back if I don't exercise enough for more than a couple of days. There's a still longer single-line "HEALTH", and a very long single-line "LONGEVITY"; I don't obsess about longevity, but I'd like to be around for as long as I can be useful and have fun. At this point I paused, thinking that was about it.  It's interesting that Mark Xu puts "look good" as a goal there; I guess even at 68 I don't like to be bloated/flabby, so I'll add a tentative right leg for that too. 

I want to try thinking about the goals in more standardized terms.... There are several ways to do that.

1. Take my somewhat distorted view of Haidt's 6 moral foundation axes: I tend to list them as "caring, pairing, sharing; ladder, limits, flight" where caring, his first axis, is a preferably temporary parental or other "caregiver" relationship; make it symmetric and you have pairing, two people who will care for each other at need. Spread it to a group and you have sharing. Add a hierarchy, whether based on dominance or respect, and that's your ladder, and if you go up above or down below the ends of the ladder you get religion; kick it away and fly, that's the liberty axis. But I add personal/private relationship with your corner of the universe as a seventh moral foundation, and call it "appreciation" or "development" depending on which end I'm looking from. In any case, I can go over my spider-legs and try to group them according to the moral foundations.

2. Take Daniel Pink's "Drive" components as being the inner incentives that people seem to share: "AMP", being Autonomy, Mastery, Purpose. (Lots of YouTube videos and web pages on this.) Again I can group my goals, thinking about how far I'm incentivized by his patterns.

3. However, then I want to take David Rock's external incentives: "SCARF", being Status, Clarity, Autonomy, Relatedness, Fairness. Again there are videos and web pages to discuss these.

 I'm not saying there's any magic to any of these lists, just that each of them provides a structure, a way of thinking about what pushes or pulls you. I think this is a place where a theory of values/incentives can be of value, even if it's just providing a set of placeholders for what you felt already.

Anyway, what do I do with the story spider thus created? I think it has been useful already; it helped me see how things fit together. So far, it's not actually about making choices: the "choice" to Exercise was very much overdetermined. However, having a detailed image in my head (or on an index card) for this story component does make it easier for me to imagine variations, even without writing everything out explicitly. I don't run, because trying that has in the past produced enough joint pain as to stop my exercise altogether... but I can look over that picture and think about changes in my actions, and imagine the resulting effects on goals. What I want to do next is a dual graph, where the spider-bodies are current goals and the legs are current actions which relate to them, so that the "health" goal would have an "exercise" leg. Then I can try putting these action-spiders together in different ways? Maybe so. Or maybe that would be too many spiders and altogether ooky...it is Charles Addams' birthday, after all. (Happy birthday, Charles.)

 

"Applied Rationality Training Regime" #6: Seeking Sense

< | ^ | >

Day 6, so we're at Mark Xu's Training Regime Day 6: Seeking Sense.  This one is really hard. Xu presents it as having two components: the first "is partially captured by words like "empathy" or "compassion"....[having to treat] other people like human beings instead of just action/reaction APIs."

Starting from that, we're supposed to develop "the rationality version of having social skills.... to observe people's actions and infer things about their values, how they think, what they know about the world, etc."

To me, this seems enormously important. We need to be able to infer things about people's values; that is precisely the problem to be solved by the Stuart Russell approach to AI alignment, in which (in my misinterpretation thereof) the AI is not given a goal of its own but rather an ability to infer what the boss wants, and to update those inferences as time goes on. This happens to be the first approach to AI alignment (other than my own which doesn't go here) which I take seriously, but it does require a rational approach to seeking sense. So I think about it.... 

Well, I've already linked to my post about more or less totally lacking social skills, despite having tried for more than six decades to develop them. So I've spent much of today thinking about that. I do make up story/models for people I meet, and these of course incorporate guesses "about their values, how they think, what they know about the world, etc." But my confidence in these guesses is extremely low; this is a talent which I seem to lack. It's true that Heinrich Schliemann, the guy who excavated Troy (with dynamite) did say "Talent means energy and persistence and nothing more!", and it so happens that today is his birthday, but somehow I'm not convinced.  When I encounter someone I don't understand at all, I know he has a model of the world, he has human emotions, it all makes sense to him somehow. And according to the Roman playwright Terence, it all should make sense to me, too, because nothing human is alien to me. But actually everything human is at least somewhat alien to me, because as Khalil Gibran put it,

“I am a stranger to myself, and when I hear my tongue speak, my ears wonder over my voice; I see my inner self smiling, crying, braving, and fearing; and my existence wonders over my substance while my soul interrogates my heart; but I remain unknown, engulfed by tremendous silence.” 

It's his birthday too, and his attitude, if I dare say it, makes more sense to me. I don't believe that I have any special insight into my own mental processing, except that I've spent a lot of time developing models for it. So.... I dunno. This really bugs me, and I've made very little progress today. Oh well... on to tomorrow, and happy birthday to Schliemann and Gibran.