Bug Hunt 2

post by alkjash · 2018-02-20T05:00:00.491Z · LW · GW · 15 comments

Contents

  Hammertime: The Second Cycle
  Day 11: Bug Hunt 2
    Setup
    1. Identity
    2. Pica
    3. Ambition
  Daily Challenge
None
15 comments

This is part 11 of 30 of Hammertime. Click here for the intro.

CFAR has an underlying mantra “adjust your seat”: systematically modify every technique and class to fit your personal situation. It’s common sense nowadays that different things work for different people, but the extent to which is true still constantly surprises me. (Kierkegaard had a fun take on adjusting your seat which he called the Method of Rotation.)

If you wish to partake in Hammertime, feel free to adjust your seat as much as necessary. Draw out the practice of instrumental rationality over a longer period of time, pick and choose the methods that appeal to you, and scale them to your time constraints.

Hammertime: The Second Cycle

Hammertime is about cultivating a tiny number of powerful techniques for solving a huge variety of problems. In the second cycle, we will revisit and upgrade the tools we introduced in the first, and apply them to tougher problems:

  1. Bug Hunt
  2. Yoda Timers
  3. TAPs and Reinforcement Learning
  4. Design
  5. CoZE
  6. Focusing
  7. Cruxes
  8. Goal Factoring
  9. Internal Double Crux
  10. Self-Trust

The new ideas we will be introducing in the second half are devoted to developing higher levels of introspection and self-honesty, to figure out your true motivations and aversions, and what to do about them.

Before each post in the second cycle, take a moment to review its predecessor.

Day 11: Bug Hunt 2

Previously: Day 1.

Noticing your bugs continues to be our single most powerful technique. Training noticing involves lateral thinking, attention to detail, and self-honesty. Today, we focus on three high-level ways in which human beings systematically err.

Setup

First, review your Bug List from Day 1 and update it.

For each of the next three mini-essays: read it over, then set a Yoda Timer for five minutes and brainstorm as many bugs as you can during that time.

1. Identity

Paul Graham wrote Keep Your Identity Small. Being attached to your identity can often constrain your growth.

Rather than making an impartial decision on what kind of person to be, people often extrapolate their identity (and morality) from their previous actions. A friend of mine calls this coprolite: fossilized and over-fitted beliefs that originate from early childhood. Are you neat or messy, stingy or generous, introvert or extrovert, conscientious or agreeable, idealistic or cynical, engineer or artist, vim or emacs? Do you look down on people for being the other way? Take moment to notice all the traits you’re attached to, think about why you’re attached to them, and consider the benefits of their opposites.

Personalities are many-faceted, and you may not even understand your true motives, fears, or skills. Do your stated preferences agree with your revealed preferences? Do your aliefs differ from your beliefs? Do people systematically judge you to be different from your self-image? Do you often surprise yourself in terms of what you enjoy, excel at, or are anxious about?

It’s useful to think of personality growth as expansion rather than change. An introvert grows by learning how to navigate social scenes. An extrovert grows by reclaiming her capacity to be alone. Instead of asking what you would change about yourself, ask what you would add to the toolbox.

2. Pica

Pica is an eating disorder in which people crave food that don’t fulfill the need behind that craving; the folklore example is gnawing on ice to satisfy a mineral deficiency. Experiential pica is any craving which doesn’t fulfill the need behind it.

My top three addictions in high school were all experiential pica.

The first addiction was romantic novels and shows of a tragic nature, which served as vulnerability and sacrifice porn. I had intricate daydreams in multiple languages of love and loss.

The second addiction was RPG games, which served as improvement porn. In Diablo III, the Gem of Ease that boosts your leveling speed on all future characters to go from 1 to 70 in about an hour; I’d start a new character every couple months to get watch the level up messages roll in. MOBAs are perhaps the worst offender in this regard, taking your character from level 1 to fully equipped level 18 every single game.

The third addiction was just …

I know these are pica because the first and third cravings largely subsided when I entered a committed relationship, and the second when I started seriously working on self-improvement.

Lent is a good time to look for your pica. Are there any habits, cravings, or addictions you don’t understand and/or try hard to cut? If they’re pica, you’re applying effort at the wrong angle. Figure out the unmet need, meet it, and the pica will automatically subside.

3. Ambition

I’ve been jogging casually for about fifteen years. Until last year, it’s been uniformly awful. You’d think you’d get used to running four miles after doing it twice a week for a decade. You’d be wrong.

Then, I decided to aim at something.

I thought: I’m going to train for a seven minute mile.

My heart replied: Oh, ok, that’s kind of invigorating.

Then I thought: I’ll train for a six minute mile.

My heart: Woo baby, let’s do this!

Then I thought: A five minute mile.

My heart: HAHAHAHAHAHAHAHAHA…

I ran for over a decade with next to no improvement. Last month I ran a seven minute mile after two months of training for an impossible goal. These days I look forward to running.


I’ve been blogging casually for about five years. Until last year, it’s been a drag. You’d think you’d get better at writing by putting up two posts a month for a year or two. You’d be wrong.

Then, I decided to aim at something.

Me: I’ll try blogging once a week.

My heart: Oh, ok, that’s nice.

Me: I’m going to blog every other day.

My heart: Now we’re getting somewhere.

Me: I’m going to blog every day for a year, and write better than Eliezer Yudkowsky by the end of it.

My heart: HAHAHAHAHAHAHAHAHA…


There’s a level of ambition that pushes you to operate at maximal efficiency, that twists your heart with adrenaline just to think about. In every pursuit, aim at a target so high it feels immodest to whisper in an empty room.

List your goals now. Keep doubling them in difficulty until your heart bends over in hysterical laughter at the very thought.

Daily Challenge

State your greatest ambition: the one that feels most subjectively immodest.

15 comments

Comments sorted by top scores.

comment by tcheasdfjkl · 2018-08-08T00:11:06.627Z · LW(p) · GW(p)

Third experiential pica thing is missing actual content?

comment by Swerve · 2018-02-21T03:47:34.501Z · LW(p) · GW(p)

Before anything else, thank you for the small piece about "adjusting your seat". It's taken a considerable amount of pressure/dread off of doing Hammertime.

Anyway,

It seems like the heart of the bug hunting skill, is having enough gears in your understanding of a certain domain. Such that you can find specific faults, inefficiencies, and leverage points to address. The inclination to go meta with the domain is also very helpful.

For example, in my own practice of bug hunt, I initially found myself selecting things I was explicitly doing wrong. Over time I started realizing that while this is helpful, it is not the entire picture. Why? Because the purpose of bug hunt is to find things that, if changed, will produce a better result in some area. With this in mind I realized that by shifting my actionable model of bug hunting, I can produce not only way more bugs, but much more salient ones. This came about by way of modeling what bug hunt currently does, and what it would optimally do. The difference between these two concepts are what I now think of as bugs.(eg using bug-hunt on itself proved to be a powerful investment of my time).

Right now I think my most immodest ambition is to Try and Hammer one rationality skill every four days. I'm going to apply it with as much consistency and versatility as I possible can. I'll then write about what I learn, and connect each skill I've learned before, to my current skill, to try and embed each one thoroughly in my toolbox.

Replies from: tcheasdfjkl, alkjash
comment by tcheasdfjkl · 2018-08-08T00:52:01.524Z · LW(p) · GW(p)

I second the thank you for the "adjust your seat" thing!

Sometimes when a technique doesn't seem to quite fit I'm not sure whether I'm missing a skill that would actually be useful to me, or whether I would actually do better to change the technique for my benefit; probably the answer is sometimes the first thing and sometimes the second.

comment by alkjash · 2018-02-21T05:44:35.057Z · LW(p) · GW(p)

Thanks for the feedback. After your Intermission comment I've updated towards continuing to add "meta-motivation" blurbs like adjust your seat at the beginning of each post. CFAR had a bunch of these which were if anything more helpful than the actual class content.

comment by MikkW (mikkel-wilson) · 2021-01-06T22:28:12.174Z · LW(p) · GW(p)

Finally coming around to this one, I found that I was a bit disappointed in myself that I wasn't able to generate as many leads as I did last Bug Hunt.

While I'm maybe not perfect at keeping my identity small, when I was visiting the identity, I found that I've already been pushing on the problematic areas, to the extent that I didn't find anything interesting while examining where my identity may be too big. Following the heuristic of Inverting Advice, maybe I could invert this and ask where I might have too small of an identity? Or perhaps it can be worthwhile for me to just try to write out everywhere where I know I have identified too big of an identity in the past.

Regarding Pica, I suspect that there may be something to it, but if there is, most of it is flying under my radar. I was able to identify soda, video games, and walking as activities that may have pica-like motivations; I find that I most crave soda when I haven't been eating enough (I always enjoy soda, but perhaps I've been drinking it more than I normally would lately due to suboptimal food intake), video games I generally have a positive relationship with, but I find myself wondering if some of my motivation lately to play videogames has something to do with me not doing practical things that my mind is built to want to do. I also suspect that walking may have provided some benefits in the ancestral environment that I'm not getting from walks in the modern environment; walking used to be a mode of transport, a way of exploring one's locale, and something that could lead to discovering new resources; of course, none of these happen on my walks (aside from meeting people, but lately even that hasn't been very helpful). Maybe there is some deeper purpose my mind hopes to achieve when it tells me to go for a walk? But then, I'm not sure this is true. Maybe I just go for walks because it's good for the mind and body, when I'd otherwise be inside all day. It's a thread to pull on.

Regarding Ambition, I think that I'm already quite good at pulling things in the direction of setting crazily ambitious goals - to the extent that I've lately grown somewhat jaded at overly ambitious goals, and now I just groan when I consider a goal that doesn't seem realistic; instead of the wild excitement that Xiaoyu's heart felt when he kept doubling the intensity of his ambition, the groaning in my head just grew louder and louder as I raised the stakes. I instead decided to invert this advice, and try to make realistic goals out of the wildly ambitious goals my brain gives me, but halving the stakes until I felt confident in my ability to achieve them. I turned the goal of increasing by 1 kyu in Go every week to the goal of increasing by at least 1 kyu every month. I turned earning $1,000 every month to simply earning at least $10 every month. Another low-ball goal I generated was to just photograph a person; I'll let your imagination figure out what the original goal was.

For the daily challenge, I'll lean right into it. My most subjectively immodest ambition is to govern a city-state on the Moon with at least 250,000 inhabitants, all of whom are people who are selected according to criteria I set. Actually, I can do better than that. If I want to be as subjectively immodest as possible, I want to govern the entire Moon, with a population of at least a billion people. That's plenty immodest, and makes my heart throb with joy.

comment by TurnTrout · 2018-03-18T04:30:08.070Z · LW(p) · GW(p)

Upon updating my list, I found that dozens of bugs - many of which had been problems for months - "fixed themselves". Of course, I tackled a good number consciously using the techniques outlined thus far, but I’m quite surprised that I was able to resolve so many others without breaking stride.

My first thought is that the act of enumeration made me more prone to noticing when these problems cropped up, prompting me to take time each day to bugfix without even realizing what I was doing.

Replies from: Qiaochu_Yuan, alkjash
comment by Qiaochu_Yuan · 2018-03-18T05:14:00.084Z · LW(p) · GW(p)

An extreme variant of this is an exercise where you spend 5 minutes explaining to yourself why all of your bugs are secretly the same bug, then solving that one.

comment by alkjash · 2018-03-18T05:07:18.183Z · LW(p) · GW(p)

Yes, totally. At some point you realize that solving every problem consciously is an exercise in futility, and all this intentional rationality practice is really just practice for training your subconscious to automatically think in the right patterns and solve your actual problems with no apparent effort.

comment by theme_arrow · 2021-01-13T08:10:23.786Z · LW(p) · GW(p)

I would like to be the lead author on a patent (and one that actually represents a new development). I work on a team that has produced dozens of patents, but none in the last few years. I want to get my skills to the point where I'm genuinely pushing boundaries, and more selfishly I want to have something public to show for my work. I don't know if that'll be possible on my current team, or if this team/company's time has passed, but it's something I would find really special. 

Reflecting back on the bug list, I actually came up with relatively little more to add based on today's post. Thinking about things in terms of the themes discussed today certainly helped me reframe some of the items that were already on there, but it didn't make me come up with significantly more bugs. The list currently stands at 124 bugs, of which I've solved 26. Many of the solved bugs were simple ones related to my space and physical objects, since these ones could be quickly solved with a purchase. But I've made progress on more complicated bugs too. The highest bug that I consider to be solved is a 5/10 difficulty, failing to shower and brush my teeth on-time in the morning, which I seem to have been able to fix with a TAP. But I've got a 10/10 difficulty bug, not being sure if my current job is a good fit for my long-term happiness, that I'm making serious progress on. That one was actually remarkably easy to get started on, I told my manager that I had some available time and was interested in trying to broaden my skills, and he arranged for me to spend 1/4th of my time for the next three months working with another team in a different area. 


All-in-all, I'm pretty happy with how these first 10 days have gone, and I'm hoping that the next 10 will be similarly positive!

comment by itavero · 2020-12-03T03:26:11.550Z · LW(p) · GW(p)

I'd like to win a Grammy one day. It's hard to not qualify that, or hem and haw, or excuse it. But honestly, I'd like to.

Not too many identity-related bugs came up for me, but maybe that's because I can't even see the things past my current identity. I'll be thinking about this more. 

Brainstorming around pica brought out a lot of bugs. I noticed sometimes it's easier to catch the pica than to fill the void that's left. Example: I play (real) guitar at a high level, but Guitar Hero is much more addicting in the short term, probably due to clear iterative improvement and overall lack of ambiguity - there is a clear thing to do; hit all the notes on the highest difficulty level. It's hard for me to figure out how to fulfill this need. Tracking pomodoros is the closest I've gotten to channeling this positively. 

comment by Long try · 2020-02-17T06:20:04.045Z · LW(p) · GW(p)

I will profoundly and immensely change the world for the better from the ground level, by creating and/or modifying the constituent element, the building block of civilization - in other words, the very foundation of society itself.

Why does my heart not laugh hysterically at the thought? Is it because I really do believe it firmly?

Replies from: DonyChristie
comment by Pee Doom (DonyChristie) · 2020-08-17T17:21:00.089Z · LW(p) · GW(p)

Any updates on this in the past six months?

Replies from: Long try
comment by Long try · 2021-01-11T10:38:34.207Z · LW(p) · GW(p)

A few, but overall it doesn't change very much.

comment by tcheasdfjkl · 2018-08-08T00:50:19.030Z · LW(p) · GW(p)

Both the initial bug hunt and this one were by far the Hammertime tasks that took the longest - I have a lot of bugs I can list in lots of areas, so nearly every time the instructions call for a 5-minute timer I take at least twice as long adding bugs in that category. Plus this time I also added some bugs I just forgot to think of on the first try, as well as things from my Workflowy that can also be conceptualized as bugs.

I think this is useful in that it lets me organize and systematize these things and keep them all in once place. Though nearly nothing I think of is actually new. So I'm not sure if this is so much more useful than the other techniques that it justifies spending so much time on it. I guess we'll see to what extent this helps me actually solve bugs going forward!