October 24, 2016

A look through the flawed lens Oct. 24

Highlights

Articles I see that please or excite

Proper use of humility: Daniel Dewey owns up to missing his giving pledge on the EA forum, and shares his plans for how to do better.

Conspiring with the Enemy and Cooperating in Warfare by Yvonne Chiu. Some faith in humanity restored. Not all of it, but some.

Andrew Gelman posts about whether it's fair to use bayesian reasoning to convict someone of a crime. This is a conversation I want to see happen, luckily there is a comments section to browse!

Siderea talks about what Trump's video tape does and notably doesn't mean. I have gotten extremely sick of 'my' people using Trump's excesses as an excuse to dismiss all notion of truth-seeking and integrity.
I think some people on the Right are hearing that video very, very differently than people on the Left do. And I think it important for the Left that they understand the various ways the Right is taking this. It is a crash course in feminist history, and an orientation to something important that is going on right now.
Xenosystems makes a similar point.

A good angle on the parallel moralities of America's political parties made at Meaningness.

David Henderson at Econlog clarifies that "Person X did this bad thing" and variations are not overall judgments about the person. If you try to argue based on a false statement, then truth-seekers are going to correct you even if (especially if) they agree with you.

I found the precurser to Uriel of Unsong. It's really fascinating seeing how fictional rationalist universes have definite meta-causality.

Miri at BruteReason deconstructs jealousy. It's a good read for how to deal with a difficult, complex emotion.

Sarah Constantin creates a new metaphor-god called Ra, to join the pantheon of Moloch and Azathoth. It's approximately the human fault of valuing smooth ambiguity and vagueness over flawed detail and concreteness. I write these flawed lens posts mainly to counteract the thing Sarah is pointing out here, so I think it's a pretty neat metaphor to have.

Lowlights

Frowny faces on activities that do not tend towards truth or fulfillment.

"The ingroup has a rich, varied and deep decision making process. The outgroup are simple creatures driven by base instincts." - ContentOfMedia

"creating an ideology incompatible with any modern political structure so that you can pity yourself as the underdog indefinately" - FatherOfNun

Laughing nervously at weird sun twitter posts - Me

Robin Hanson writes about the Smart Sincere Contrarian Trap; namely, the idea that smart sincere people champion ideas that lack fashionableness more often, and so high-status movers find them "Good to listen to behind the scenes to get ideas for possible new fashions, but bad to embrace publicly as a loyal group member". I spit on that. (The trap, not this post or smart sincere contrarians)


SportOfBrahma

Open Questions


Scott Sumner at Econlog has such pretty graphs and I don't understand them any more than the last time they showed up. They look so concise and tidy... way too concise, I am majorly missing some context. Can you explain what all of the words and arrows refer to?

I am disagreeing pretty hard with this TheMoneyIllusion post (also Scott Sumner), it feels very dismissive of what I think are real concerns for people. Increasingly I find that Trump's aims point to very critical cultural issues that he is absolutely the wrong person to solve those problems. Admittedly, a lot of that results from numerous bad arguments from Dems (same poster talking about this, ironically) inoculating me against the better forms of the arguments. I really, really do not want the grievances of traditional, rural America to fade back into ridicule and obscurity with the fall of their ill-chosen champion. Do you predict post-election riots? What measures are in place to prevent them?
Andrew Sullivan, in nymag, has a long essay on how the information age kind of crushes our souls. I think we can mitigate the harm while keeping the benefits, but I don't see a lot of coordinated effort to do so. Rationalists pay a lot of lip service to binding ourselves to reality, but we definitely spring from an in-the-air intellectual demographic. What, besides meditation, do you do to stay connected physically and emotionally?

Jeff Kaufmann writes about the selection effect that goes on when some organizations aim for transparency full-time and others only sometimes. When one organization volunteers only good information, organizations that dutifully report the bad as well look comparatively awful. So, How completely would we have to reconstruct the education system to provide a firm base of statistical understanding and information?

If you think we're wrong or unhelpful, or if you aren't sure, tell us!

October 17, 2016

A look through the flawed lens Oct. 17

Highlights

Articles I see that please or excite

Embodied cognition. Not in the usual sense of "clench your fists to increase willpower", but using your body and environment as part of your computative toolset.
This gives me a nice framework to mash up idle thoughts about PCK-seeking* and database design; I think we can increase ability in a lot of domains simply by picking mental representation of concepts that are easier to work with in relevant ways. Good infographics, for example, take advantage of our very powerful visual systems to convey a lot of information simply and concretely. I do not think it is a coincidence that the descriptions of mathematical savant ability I've heard include an element of synesthesia.
*Nod to Valentine Smith of CFAR

Kudos to Unit of Caring for promoting ethical norms.  Let's not go down the road where we "make it dangerous to your continued wellbeing and ability to earn a living to try to persuade people of your ideas". I'm also fond of this post on giving room for people (even lizard-people) to change their minds.

Also regarding Unit of Caring, I JUST DISCOVERED SHE'S LINTAMANDE. Her Patreon rewards sound so tantalyzing all of a sudden... And seeing that she's on AO3, it makes sense why her glowfic Arda had those really freaky overlaps with that modern Silmarill story I'd read once upon a time.

On Thing of Things, Ozy has posted an Intellectual Turing Test try on 1-2-3 different stances toward social justice. I could stand to see more of this and double-crux as normal activities.

This article on providing an exit-ramp for people who are dead set on ideas you think are catastrophically wrong (as they think Trump supporters are). Surprise—the people you disagree with are people! They do not respond well to being cornered and humiliated. They do respond to patience and information and validation. The article's given advice feels like a short summary of Carnegie's "How to Win Friends & Influence People".

Lowlights

Frowny faces on activities that do not tend towards truth or fulfillment.

I anticipate the world burning before my eyes, and run off to grab a bowl of popcorn. That is probably not the most helpful response I could have to critical and complex problems. Many things seems distant and darkly amusing. Especially the idea that the outcome of this election might seriously matter to America's short-term non-hellishness and long-term existence as a stable country. We could drag who knows how much of the global community down with us, and I'm not convinced that either candidate can do anything useful about it. Everything our predecessors ever worked to accomplish could fall apart and it's hilarious. Also funny: ISIS has bomb drones.

There is no Nobel Prize for ecology or geology or climate science. A bunch of fields just get stuffed into the peace prize #NotAllScience

This Guaranteed Jobs Proposal article. I basically read this proposal as saying 'solve the jobs problem by solving the jobs problem with the government'. This fails at the virtue of simplicity! Also, I see absolutely nothing in this proposal that seems aware of disability as a potential roadblock. I pay close attention to when my shoulder Social Justice Warrior and shoulder Libertarian agree on anything.

Robin Hanson posts on Overcoming Bias about the backlash he gets for focusing on emulation-based AIs instead of coded-algorithm AIs. I'm glad he's working on it; I do not want AI-friendliness research to become an echo-chamber. Anders Sandberg wins the cake for answering my "Can you model this?" question before I'd even asked it.
The issue of how to spread research effort cuts to the heart of EA values. I have day-terrors about EA becoming a villain organization that actively hinders people from going after low-hanging fruit, lest we waste any resources on something that is not "THE TOP 3 TRUE CAUSES".

Open Questions

How many instances of bad incentive design can you spot in your everyday life? Contract theory provides a shiny-new interesting frame on how I rate and reward myself for my own work.

Andrew Gelman's mini-paper on the scientific replication crisis, shared for this figure. Does someone you know have the skill and lack-of-better-things-to-do to make a quick graphic with comparisons of 1%, 3%, 5%, 8% power? Would recommend coloring it in a tasteful rainbow, and blowing it up to desktop background size so you have plenty of plausible excuse to pass it around.

If there was one economic idea you could explain to everyone on earth, how would you cement it to as many people's system ones as possible? Econlog (quoting Steve Horwitz's AMA) recommends The idea that prices are 'knowledge surrogates', a critical form of expressing information necessary to making rational choices. See also 

Drones delivering prison contraband. What would it take for technology adoption to definitively outpace our ability to generate defenses against it? I worry we're already there with shoddy password security, and everything is kept functional through frantic patchwork and luck.

This article on cutting Latin America's murder rate got me thinking... I usually see redistribution touted as a solution to socio-economic inequality. It might be done through voluntary donation, it might be done through obligatory taxation. If your problems are rooted in social and economic inequality, What actually happens when you directly encourage and strengthen strong empathy ties across class borders? I'm sure this is horribly naive and there are obvious ways it goes hellishly wrong, but I'm curious what the specifics of those are. I would want to create an interchange of sorts... incentivize safe and mutually beneficial interactions between economic classes. Make pen pals and patreons. Open neutral meeting spaces with TSA-level security. Integration!


~*~

If you think we're wrong or unhelpful, or if you aren't sure, tell us!

October 9, 2016

A look through the flawed lens Oct. 9

Random Musings as I trawl the news and lesswrong diaspora.

Highlights

Assortment of things that please or excite

A good heuristic mentioned on Andrew Gelman's blog: "If there's no report you can read, there's no study."

Jeff Kaufman tracked his schedule. I approve of numerical precision as a general habit.

Ben Hoffman had a lovely post on the principles of truth-friendly discourse. I already intended to make a round-up like this, but it helped me flesh out my criteria for evaluating meta-discourse.

Put a number on it's words on the question "Should a person who seeks to make the world a better place follow empathy or discard it?" moved me. I delighted in the number and unobtrusiveness of reference links it holds.

This econlog post deserves a mention for the line "Since I oppose preaching to the choir, I'm aiming the book at any human being interested in (a) the ethics and science of immigration, or (b) non-fiction graphic novels". I fiercely approve of this attitude; we don't have nearly enough manpower going into bridging inferential gaps.

I retract my lingering complaints about Petrov Day rituals. The proposed program doesn't over-glorify this one decision out of all of human history as nauseatingly as the people who talk up Petrov day in conversation.

Lowlights

Frowny faces on activities that do not tend towards truth or fulfillment.

On Andrew Gelman's blog: "It is OK to criticize a paper, even it isn’t horrible."
The backlash that Gelman got on criticizing a paper's statistics bothers me.  Critiquing half-good things is usually a better exercise in critical thinking than critiquing uniformly awful things.

I know am going to wind up linking this self-deprecating explanation of the rationalist community to a number of normie acquaintances on Facebook. I'm not sure this is actually a good idea, because this desire bases itself primarily on a bitter "Fight me" impulse.

This pun induces Mixed Feelings

Open Questions

The problem of Digital security exemplifies, in a less controversial way than most of my pet issues, the tragedy of people developing an attitude of bleak resignation that prevents implementing solutions. At minimum, I can deal with this myself by dumping all of the idea that are 'too depressing' to evaluate in one place and drawing out a random subsample of next actions. So literature review time: What do we know about how the general problem of decision fatigue manifests, and what does that imply about potential solutions?

I don't think it's fair to say that we have sold our values to the bottom line when people can hardly find any information except the price to make tradeoffs on. People fight viciously over the framing of where our meat comes from precisely because it has the power to massively shift beliefs and choices that affect millions of lives either way, and it's going to have that ability whether the shift is justified or not. I could ask how we're going to put a stop to Factories shutting down informed consent with legislation, and the (lesser but still worth mentioning) occasional problem of ARAs emotionally manipulating a populace that have their naivety sheltered and reinforced by third parties, but instead I'm going to ask Is there an app/website that crowdsources where are all the small local farms are, and what their specialties are?

Someone raised the question of whether humanitarian intervention constitutes an exception to the principal of non-interference in international law. Non-interference affects private actors much less, and one might assume that's due a commensurate lack of diplomatic protection or an assumption that non-government-affiliated individuals have substantially less power to effect changes. But if non-intervention is actually important for other reasons, then it's plausible that an effective altruist could find ways around that and cause problems. What purposes does the principle of non-interference serve,  and should they apply to sufficiently powerful private actors? The current analogs I see to the non-interference principle in the EA movement are tangential heuristics: 1) you are unlikely to predict local effects better than local actors, 2) nudging existing overton windows gets more results than pushing against public opinion, 3) Transparency and changing course quickly in response to evidence are both important and governments usually suck at them.

What research is there on whether lucid dreaming harms learning? I'd expect it to harm the beneficial effects of REM sleep, at least as it's used by your average lucid dreamer.

Are there potentially positive uses of IP spoofing? Or is the lack of measures against it entirely about implementation difficulty and opportunity cost?

Put a number on it's post on the question "Should a person who seeks to make the world a better place follow empathy or discard it?" moved me deeply, but it still feels like a big unanswered question. When do you want to use an empathy framing? I think this is an especially important question to ask if you're planning to design ritual as I hope to.
This article on 'saving science' has an interesting framing on endemic problems in scientific research. They propose that the major problems befalling science—failure to replicate, multiple test mining, positive bias, elitism—but suggest that these are all fundamentally caused by leaving the course of science to "the free play of free intellects" instead of wedding it to creating concrete effects in the real world. As a key example of this, they point to the innovative outputs of the Department of Defense. I am very on board with the idea that science really needs to be backed up with something other than itself, and could be persuaded that cross-field peer grants aren't enough to keep research focused on high impact questions. But their other case study in good science management is the National Breast Cancer Coalition, and one of the independent researchers I trust is in the process of writing a whole book on how the cancer research movement is utterly fail. I must ask, then, What would it look like for a research effort to be focused on the correct problem every step of the way? Is it possible to identify and reward when that is happening?


~*~
If you think I'm being wrong or unhelpful, or if you aren't sure, say something!

October 5, 2016

Theory of rationality

'Do not forget your purpose', I said to myself.

When you see lies and injustice to be fought, you will take up arms. You will stand and fight. You will make harsh moves and bitter choices. But the defeat of an enemy must never come at the cost of what you took up arms to protect. 'Do not forget your purpose.'

When you aspire to do great works, you will work. You will explore and rework. You will strive and persevere. But valiant attempts must never come at the cost of achieving. 'Do not forget your purpose.'

When you seek happiness, you will make meaning. You will forge connection and write stories. You will create safety and purpose. But maintaining the structure of life must never come at the cost of living it.

'Do not forget your purpose', I said to myself.

July 22, 2016

How easy is it to just get outside of the house? An experiment.

Exercise is probably one of the highest value wellbeing interventions one can apply ot their life[citations needed]. It's sure a lot easier to get errands done when I'm already planning to be out and about. Not to mention, I live in a temperate coastal city with pleasant weather and scenery that merits enjoying.

How easy is it to just get outside of the house? An experiment.

I noticed I resisted decision/preparation timeouts in leaving the house. I thought I could train a discrete progression of actions (remember, decide, get ready, GO!) and turn it into a keystone daily habit that I can make other plans around.

Starting March 22, I decided that if I noticed I had no plans to leave the house, I would use the next break to step outside and wander a bit.

Data

Review Apr 8 - The "breaks" I take are not discrete, actually. I'd forgotten about this. I made a ritual habit to venture outside at noon every day instead.

Todo Notes Apr 11 - I haven't used my scooter in a while. Instead, I got rides with my romantic partner. The sky got cloudy and rained over the weekend. Fat chance I was going outside in that.

Review Apr 29 - Over the past week I have gone on multiple walks, 2 drives, and to lunch with a friend. I'm not really paying attention to this anymore; whether I do anything about it is highly dependent on my mental state.

Conclusions

This provided some useful information. When I am in a good mental state, I'm willing and ready to try to solve problems. Readiness to leave the house comes as a natural extension of this. With this in mind, I conclude that I need to improve my mental state in general.

Followup

I threw some trigger-action habit flashcards in my anki deck to cover some 80% of situations that tank my mental state, and these have been in action to varying success since April.

---
Was this a nice read? Does this seem like something worth performing for yourself? I would greatly appreciate it if you took a couple seconds to comment yes or no.

July 16, 2016

What python topics do I most consistently need to look up? An experiment.

A programmer's lament: Looking stuff up sucks.

I want better indexed docs. In the interest of this, I decided to do a little experiment. I wanted to figure out what, exactly, out of all the docs I looked up are important to have within reach.

What python topics do I most consistently need to look up?

I decided I would trawl my chrome history for programming-related pages.

Data

I noted in my strategic review* that I'm confused about which functions are built into python and which are class/library functions.
*(strategic reviews are when I periodically overhaul my to-do list, making a record of my successes and failures and checking in with my longer-term goals.)

Conclusions

The areas I want to focus on are:
  • string methods, formatting
  • file input/output, incl. json
  • git manipulation
  • exception handling
  • regex syntax
  • javascript syntax
  • postgres setup, syntax

Followup

Sanity Check
- Examples dominated by what I've spent time on. What have I *skipped* doing because I didn't even know where to start? noted in strategic review that I'm confused about what fns are built into python


Things I could do next:
  • Create Incremental Reading cards for docs/101 tutorials on
    • string methods, formatting
    • file input/output, incl. json
    • git manipulation
    • exception handling
    • regex syntax
    • javascript syntax
    • postgres setup, syntax
--

Was this a nice read? Does this seem like something worth performing for yourself? Take a couple seconds to comment yes or no!

July 15, 2016

Do I make progress on projects if I stick to 5 minute stubbable code functions? An experiment.

"ADHD is really executive function disorder" - Russell Barkely

The quote in the best talk ever made about ADHD goes something like that. In my quest to overcome this and learn how to actually get things done, I've tried many things. This is one of those tries, and I welcome you to read along with my adventure:

Do I make progress on projects if I stick to 5 minute stubbable code functions? An experiment.

I want to actually finish a programming project. But how? I took to the internet to find some inspiration, and found a Systematic Program Design Course with what seemed like a good system.

I gathered up the materials and set myself an experiment. I would try 5 to 10 rounds of attempting stupid tiniest chunks of email automation or website code (those being the projects I felt invested in at the time) in only 5 minutes. A significant part of my aim will be to make each function "stubbable"—that is, it returns a value such that the code is executable even if not correct.

If this is a good method of making progress on my projects, I expect my code to go off on an occasional tangent but for the functions to retain enough isolation to be fixable.

Below, I summarize the specific issues rather than include embarassing detail.

Data

Mar. 21 - None of the parts I created actually do what I need yet.
  • 5 tries so far on email automater.  
  • 2 functions stubbed,  
  • 1 function deleted.  
  • 4 function tests.
  • Issues
    • 13 complications
    • 9 information errors
    • 9 objectives failed
    • 3 uncategorized errors


Mar 22 - tried to make an website code stub, the result was invalid.
  • Issues
    • 1 complication
    • 2 information errors
    • 2 objectives failed

Mar 25 - copied a function from website, made an instance function, made a new class. Spent a while on planning which components I'd need, then made a new class. Got sucked into looong work instead of stubbing, bad idea.
  • Issues
    • 3 complications
    • 3 information errors
    • 4 objectives failed
    • 2 meta-issues

Mar 26 - starting over with wishlist of components for a log parsing utility.
  • 3 functions,
  • 5 stubs,
  • 5 tests.
  • Issues
    • 2 complications
    • 4 information errors
    • 6 objectives failed

Mar 27, 28. worked on programming, mostly doing big pieces trying to force sequential progress. I think I really benefitted from the pieces already built, and that I made faster progress when I kept to the spirit of the chunking strategy. Not bothering trying to record issues on this, too scattered.

Review Apr. 8 - I quit sticking to the 5min thing, but spent a bit of time working out chunks to work on. I'm going over the timers I set even then, esp. with issue hunts.

Conclusion

Yes this helps, but with caveats. Stuff gets done if I time-constrain myself on making stubbable functions, *including* separate time-constraints on the functions to test other functions.
Followup
I continued to use this method to break down my work in projects, and am just now sharing this experiment on my blog.

After an experiment I always take a few minutes to think of Sanity Checks: These are other quick experiment ideas (ideally taking less than five minutes)
  • How much can I actually type at mozy speed in 5min?
  • Estimate how long it'd take me to refactor a data type throughout a medium-sized project.
  • Time myself on refactoring the log parser to look nice
  • Fermi estimate time wasted on a project because of having to look stuff up and verify syntax. 
Compare to time wasted hunting down accumulated errors for lack of a good feedback loop. Which is actually likely to be the bottleneck?

I rarely the time to do the followup tests I dream up, but I consider it good practice to habitually consider what being plausibly, identifiably wrong looks like.

----

Was this a nice read? Does something like this experiment seem worth performing for yourself? Take a couple seconds to comment yes or no!

July 10, 2016

Rationality-Habits Anki deck

It occurred to me, as I was discussing my use of Anki for habit-building with a good friend, that a decent portion of my personal growth is getting dutifully recorded as flash cards. "Why not share the decks?," I thought. "A random soul might find my notes useful, or my friends at least might consider it informative of my strange thought process." Here is a deck of rationality habits, filtered a bit for anything excessively tentative or personal.

Faithful readers, you would do me a great honor if you looked over some part of it and shared with me where you find pleasure, fault, or confusion.

April 2, 2016

Testing: How much could I learn on Rosalind.info?

Hello, internet!

If you're curious what I get up to with my time, observe! I ran a little experiment today. This post reports it in a hodgepodge mix of "personal notes" and "terse scientific report".

If you take thirty seconds out of your day to comment, I will owe you one. Please encourage the norm of testing everyday ideas empirically; Be my miniature peer review.

How much would rosalind.info teach me?

I want to know the holes in my knowledge. If Rosalind actually helps, I want to train my internal motivation to prioritize working on it.

Hypothesis:

I currently can't do ~70% of problems. My learning trajectory looks unlikely to increase my capability beyond ~%50.

Method:

Experiment as proposed Mar 28 - count how many rosalind.info project list I can't do and how many I likely won't learn

Resources:

Rosalind Problems Lists
glimpses at other references for algorithmic and data structure fundamentals

Results:

Apr 2: I ignored Biology Content Knowledge, except where the lack of deep understanding seriously impedes my ability to program the solution.

PYTHON VILLAGE 1:0:0
  • 6 understand
BIOINFORMATICS STRONGHOLD 1:1:5
  • 13 understand
  • 14 don't understand yet
  • 78 doubt I'll cover
BIOINFORMATICS ARMORY 2:1:4
  • 4 understand
  • 2 don't understand yet
  • 9 doubt I'll cover
BIOINFORMATICS TEXTBOOK TRACK 1:2:4
  • 16 understand
  • 39 don't understand yet
  • 70 doubt I'll cover
ALGORITHMIC HEIGHTS 3:4:1
  • 12 understand
  • 18 don't understand yet
  • 4 doubt I'll cover
TOTALS 2:3:6
  • 51 understand
  • 73 don't understand yet
  • 161 doubt I'll cover

Conclusions:

Technically speaking, Rosalind wouldn't teach me anything. It has practice problems and a minor amount of scaffolding; designing the solution remains up to me.

Rosalind can sharpen my current skill set some, and do a little more to stretch my limits (mostly in graphs). The vast majority of problems deal in bioinformatics-specific knowledge. If I want to focus on that, the "Bioinformatics Armory" set works would shore up relevant foundations while not wasting time on general skills. This accords with its stated purpose.

Python Village and Algorithmic heights make okay sources for general practice.

Sanity-checks:


  • talk to an actual bioinformatics researcher, ask what they do and what tools they're using
  • try this experiment on an actual programming practice book, or a fundamentals checklist for comparison
  • google search for "waste of time rosalind.info" and "helped me rosalind.info", see if I can get some user experiences for how much it has helped other people
  • what does ONET say about the career outlook if I did decide to work in bioinformatics?
  • spend 5 minutes thinking about ways I could leverage bioinformatics domain knowledge in my other goals

Followups:


  • Trawl the Rosalind problems for Anki cards
  • Throw the glossary into Anki
  • Share this feedback with rosalind.info

Checklist for a Good Anki Card


This is my personal Checklist for making a Good Anki Card. Refer to as it helps, ignore it as it doesn't.

Better version here.

TLDR:

☐ I know why I want to make this card
☐ I know what I need to know for this card
☐ I know how to recognize when/where I need this
☐ There are no almost or partially right answers to this card
☐ There are no technically correct answers which comically miss the point
☐ There is no way to cheat to the right answer
☐ I cannot get marked wrong for being smarter than the question writer

DO NOT MAKE A CARD IF YOU DO NOT FULLY UNDERSTAND THE INFORMATION INVOLVED.

I know why I want to make this card

Always ask yourself why you want to make a card. Ask yourself before you do anything else. Ask so before you start on a chunk of cards and ask before you make each card. The answer to this question tells you what information you need.
  • If you want to remember things, you have to know what exactly you're remembering. 
  • If you want vocabulary, you have to understand what the words mean and when you'd use them 
  • If you want to know things, you have to understand what it is you know. 
  • If you want to do things, you have to understand how and when to do them 
  • If you want to follow advice and principles, you have to understand what you're following, what to do, and when it's relevant.
Do not make cards if you do not totally understand the information involved. Really, don't. I put it first on the checklist because it matters the most.

If you want to, you can justify memorizing anything on the basis that you totally understand how to turn "This Question" into "This Answer". You may feel tempted to throw a bunch of information into your deck and hope that some of it will stick. Most stuff added this way will waste your review time and dead-end your actual learning.

Anki can indeed help you can bootstrap your way up to big ideas very quickly, if you take the time to solidify the foundation of knowledge you're building on top of. The throw-it-all-in-and-hope-something-sticks method does a poor job of prioritizing the right things to reinforce, leaving you with a treacherously swiss-holed foundation.

I know what I need to know

When you have a particular piece of information you want to include, ask yourself what exactly you need to know. This tells you what kind of note template to use:
  • Question & Answer card: a certain response for a specific situation
  • Reversible card: two ideas that lead directly to each other.
    • Commonly used for an idea and its representation, like vocabulary or abbreviations.
  • Cloze deletion: a big hunk of information that needs to be understood in context of itself
    • Fore example, remembering the specific wording of a phrase. 
  • Many examples, one answer: a principle that spreads across many domains or has many different ways of instantiating itself
  • Generated cards: a skill to train. Anki doesn't have much of this built in, but you can cheat a bit with templates or excel. Hacks exist for making truely dynamic cards.

I know how to recognize when/where I need this

Ask yourself when the information becomes relevant and how you'd know. This will tell you what needs to go on the front of the card.

Invoke signs and senses and emotions, hint at what context you need to remember this in. Wording the question right will help you find the quickest and most reliable mental pathway between noticing a need for more information and remembering the exact information you need. If there's a lot of situations in which you need to remember the information, you may need to make more cards.

You may feel tempted to shortcut this process by declaring the obvious answer—"when I see the card question!" and just try to come up with a question that the information answers. That won't help you if the card question looks nothing like the specific moment when that information should pop up in your head.

I cannot get an almost or partially right answer to this card.

Ask yourself if you notice any answers that could conceivably get half-credit for. This will tell you if you need to split an idea into more cards.

Anki and other spaced repetition software wins out over good old fashioned flash cards because it uses smart scheduling; if you review all the information at the same interval anyways because you crammed it all into a single card, you lose that benefit. Long answers also take longer to learn, longer bring to mind fully, and longer to judge. You have to remember them as a group, so improvement comes all-or-nothing. You might remember pieces in a different order each time too, never practicing one mental path often enough to improve substantially at it.

You may feel tempted to throw a few glossaries into a deck. However, definitions actually cover a lot of information! Rather than try to judge whether you covered *enough* of the definition every time you answer the card, split it up. Take a step back, ask yourself when you need each part of the definition, and make cards for each case.

You may also make the mistake of writing questions that require you to list all the members of a collection, such as the names of your cousin's kids. You would get better practice out if you split it into things like:
  • recognize a name as one of your cousin's kids' names
  • given an incomplete list of the names, remember the ones blanked out
  • know how many kids your cousin has total.
  • recall a name to start from (youngest or oldest probably)

I cannot get a technically correctly answer that completely misses the point of this card

Ask yourself how a determined idiot or smartass will completely miss the point of your card. This will tell you if you need to work in details and get vividly descriptive.

People make the most egregiously technically-correct-but-useless-card-answers in response to Skill cards. You can make a card asking what to do in an emergency situation, and could easily get the answer "right" without having any idea how to put the answer into practice. To avoid this, you need to use vivid details—mental narrative, personal examples, pictures even—that get your imagination running wild and *put* you, mentally, into the situation where you need a real and useful answer.

It happens more subtly in vocabulary cards. You may find that you could define a term at multiple levels of detail, that the same word takes on different meanings in different contexts, or that the same idea goes by many different names. What answer are you looking for? How do ensure that *exactly that answer* is the first thing that pops into mind?

I cannot get a right answer for stupid, unrelated reasons

Ask yourself if it's possible to construct the right answer for the wrong reason. This will tell you if you need to make things more simple and consistent.

The way you word the question matters. You can give leading questions, full of hints that nearly obviate the answer. If you phrase all of your "No" questions the same way, you might start giving "No" answers based on the phrasing instead of thinking through the card's prompt. If you put a whole bunch of extra information in the question, you might start associating the answers with those details instead of the central point.

So think carefully when you add related cards: add both positive and negative examples, keep a relatively consistent question style for both, and put extraneous details in their own cards.

I cannot get the wrong answer for being smarter or more knowledgeable than the question writer

Ask yourself how what you learn now could interfere with what you'd learn later. This will tell you if you need to put in more precision and distinctions.

When you learn about similar things you often confuse them. Two cards with exceedingly similar questions and differing answers will screw up your practice of both; you need to either pick one answer to be the "proper" response or rewrite the questions to make the distinction clear.

Information can also get out of date, and or conflict with other sources sources. Where you know this is likely to happen, you need for the cards to be either more precise (not "this is true", but "X said this", or "This as of 2016") or for there to be an easy way to find and update them all quickly (Tags, Decks, Card types).

Also, in case I wasn't clear before,

DO NOT MAKE A CARD IF YOU DO NOT FULLY UNDERSTAND THE INFORMATION INVOLVED.

Thus concludes my checklist. I'll go into more detail and provide resources in later posts.

March 19, 2016

An Experiment In Morning Blues

Hello friends!

If you're curious what I get up to with my time, observe! I ran this little self-study over the past couple of weeks. This post reports it using a hodgepodge style somewhere between personal notes and a scientific paper.

If you want to encourage the norm of testing everyday ideas empirically, or just do me a solid favor, please be my mini peer-review and take a moment to comment.

~*~

When I have a bad morning, the rest of the day's productivity seems to get shot and that feels bad. I set out to do something about it.

Question: What are the key factors in having a pleasant morning?

Hypothesis: Happiness will correlate with getting dressed and eating food and getting sunlight. I expect I can manage those things more easily when I'm in a good mood.

Method: For 15 mornings, I recorded when I did one of a set of things on the whiteboard. Collected "Happy, get sunlight, get dressed, get water, eat food, noise control". Looked for visual correlations. Checked straight predictive accuracy.

Results:

sorry blind people i'm lazy

Dashes in the first image mark a wholly blank day.

The second image is some summary data of the first.
Counts: True positive, false negatives, false positives, true negatives for each column.
Below that: Accuracy, precision, sensitivity (recall) for each column

Conclusions:

  • Noise control may help a bit. Nothing else looked like an especially accurate predictor of mood.
  • Eating in the morning may anti-predict feeling good.
  • Getting a drink might be a sensitive enough mood test to use as a Trigger Action for exploiting them.
Sanity checks:
(This is a new section in my notes, added to encourage me to notice whether my ideas could be wrong in a way that would have take embarrassingly little effort to find out. If you can facilitate one of these sanity checks, please let me know!)

  • Ask someone who can Real Science(tm) to calculate my result's p-values.
    • Common sense-wise, I had 5 happy mornings recorded to work with. All of my conclusions are based on, like, a difference of 1 data point.
  • Check correlations between the variables themselves.
  • Take a mood test with one of clearerthinking.org's tools at several different times before 11am. If I get a wide variance and/or the empirical-ish results differ from my overall self-report then my results are probably bogus.
  • Search google scholar for studies about morning routines and read the abstracts
  • Test whether I get carb crashes. Food does have a plausible mechanism for causing a negative effect.
  • Ask autistic tumblr how they tweak their morning routines to prevent overload.
Potential Follow-ups:

  • Try this again when I'm not sick
  • Trigger Action Plan: With a drink in my hand, free associate thoughts about stuff I need to do.
  • Practice making/importing data into R
  • Spend a week or two collecting a more representative sample of things I do in the morning, run another self-study with those.

March 16, 2016

Anki Deck for Systematic Program Design

I want to know what skills I can work on to write elegant code. In particular, I want to know the secret sauce to making a comprehensive project I can show off to people

Methods: I searched an edX course's materials for some causal intuitions I could reinforce. (Background understandings, like "If I get X right then Y will go a little more smoothly".) I then stuck my insights into an Anki deck to be digested at a later date.

Results: Anki Deck on Google Drive (UPDATE: 4 months later Revised Deck)

Good cards require personalization. Results may vary.

Random sample 4/49 cards
Anki card: Why bother running code that you know doesn't work yet? ... to catch little mistakes before they build upA self-referential type must have at least two cases: a reference and a base case.Anki card: when should I first try to run my program? ... as early as possibleAnki Card: Function stub means... a dummy function that returns the right type value. e.g. def add_numbers(n1, n2): return 0
An aside: taking snapshots with the intent of making them public is a really good way to catch mistakes.

Sanity Checks:
Ask [programmer friend] to look over my cards
Try the following the "recipes" on a practice programming problem
Ask people on my blog for feedback

Followups:
Observe how many cards I've deleted, marked, suspended, or edited in three months.
I hit my comprehension limit partway through, maybe read it again after I've reviewed the deck.

----UPDATE July 11, 2016:
4 months later, I still have 49 cards. I've modified 31 of them. Here is the Revised Deck.

March 15, 2016

What's a good metacognitive mirror for defeating self-loathing?

Someone asked that question on a forum I frequent, and I thought I'd share my answer here.

First, what is a metacognitive mirror?

People have bias. We sometimes have ulterior motives nudging our decisions in certain ways, a kind of bias we're all aware of, and our perception of the world also develops some quirks based on the way our brains work. (See Kahneman and Tversky)

For example: we tend to question evidence that confirms our beliefs less harshly than we question evidence that disproves them. Counteracting this can be as simple as pretending you started out believing the opposite, and asking yourself how you feel about the evidence then.

Some problems like this take more effort to root out than others. Some of them actively work against our ability to notice and overcome them.

For example, feeling angry.
A lot of heated discussions go much more smoothly as soon as one or both parties take a breather to calm down. But the angrier you feel, the less you notice the ways that your anger works against you. Sometimes you can be in the middle of an argument, shouting "I'M NOT ANGRY" and barely even notice the contradiction.

Another example, getting drunk
We know alcohol causes a slower reaction time and impulsive decision making. There's good reason not to drive or make important decisions while under the influence. Unfortunately, judging whether you're too impaired to do something is also a decision, and it's negatively affected as much as any other.


What's that about self-loathing?

When you hate yourself, you can twist *anything* that happens into evidence that you are an awful person. Any mistake is unforgivable, anything good could have been better. You develop very distorted beliefs about your own abilities, what other people think about you, and what is or isn't worth doing. An astute self-hater may even notice the distortion of their thinking and proceed to beat themselves up for not having fixed it.

A litmus test on par with breathalizers for alcohol: Look yourself in the eye in a mirror and say that you're a worthwhile person.

It's stupid and corny, but it's seriously just words. If you find yourself flinching away from a simple exercise like this then your judgment is likely affected in less immediately visible ways too. Other tells: avoiding eye contact, talking faster, touching self nervously, exaggerating your words or voice or movements, nitpicking over the technicalities of your claim, making excuses not to do it another time, etc. It's like Gendlin-style Focusing.


You can combine the mentality of this with body mindfulness; notice that your body exists and spread yourself out to physically lay claim to all the space you can afford yourself.

You can especially benefit from noticing why you aren't trying both of these *right now*.

March 7, 2016

Political campaigning: Some California Ballots

TLDR: I checked Ballotpedia, and can give soundbites but do not significantly less confused about politics than I was before.

I saw a poster advertising paid political campaigning and called in on a whim.
Well, 70% whim - 20% intentionally pushing my comfort zone - 10% pleasegivemeajobIwantmoney

Motivation: If something comes of the interview they set up tomorrow and I go campaigning for spare cash, I want to know what I'm pushing. I'm not okay with making the world a worse place because I couldn't be bothered to do even rudimentary fact-checking. I think plenty of people prefer to hear an honest case, and *I* prefer to not look like an ignorant idiot whenever someone asks me an off-script question.

Question:
What initiatives did the guy on the phone bring up, and do I support them?

Method:
Google keywords mentioned (Prop 40 education tax extension, Water bond 2016,  Tobacco tax to medical)
Find out that's not specific enough to find recent things, so look up recent stuff on whatever wiki-ish site that I find other propositions on.

Sources:
sources: https://ballotpedia.org/Potential_2016_ballot_measures#California
the notes I took during the phone call
(That was a good call, grabbing that notebook at the last minute!)

Answer:
Tax Extension to Fund Education and Healthcare Initiative
An Income tax increase in 2012 (on 250k bracket) gets 12 extra years, the funds from this increase go to schools and healthcare. Guy on phone suggested this is their focus

I think this likely fails to solve underlying problems with the education system and, dependent on how that tax bracket plays out it practice, possibly shits on entrepreneurs who *would* have helped solve the issues. I wonder where the money was going before.

Water Supply Reliability and Drought Protection Initiative
$6 billion in bonds for water supply infrastructure. Cost to local gov money maybe ~3x more than it saves. ($300 vs. $100 mill annual average)

Infrastructure is important I think. I'd vote yes and hope that someone else was paying attention to the implementation details.

Healthcare, Research and Prevention Tobacco Tax Amendment
Increase cigarette tax to $2/pack, with equiv. measures on other tobacco products and electronic cigarettes. Allocate revenues to healthcare and tobacco-related stuff. If biennial audit suggests this tax is decreasing tobacco use that other anti-tobacco programs rely on, transfers tax revenues to offset them. Supposedly a net revenue increase of $1-1.5 billion in next few years, decreasing in later years.

Generally for, but suspect diminishing returns on this sort of stuff. The offset worries me, not sure how the incentives shake out. Also curious what the consumer cost of cigarettes works up to nowadays. What programs ostensibly prevent taxes like this from just penalizing addicts cuz they're acceptable targets, and do they work?

Sanity Check:
Google if bonds are usually a good idea
How much of a marginal difference would a campaigner even make to the outcome, is it worth *refraining* from?
Ask [name] or [name] to research this shit
Reverse check - What are the biggest problems in California? Do these initiatives have any effect whatsoever on those problems or is it mostly opportunity cost at stake?
Have we tried talking to someone in the $250k bracket about their lifestyles and the taxes they pay?


Followup:
Talk to the interview guy
Read the whole proposal texts with a highlighter and red pen
Look up proposition 40, try to find the connection to the tax extension

March 5, 2016

Goal setting

An incomplete list of stuff I'm up to

Alumni Pair Debugging
Some in my community express feelings of extreme loneliness and ineffectiveness at times. To counteract both of these, I started a service to pair people up regularly to troubleshoot their problems together. I engineered the system to stay stable and low-effort from the start so they could rely on its continuation; now I'd like to expand its scope. I will gather data on how people engage with the system now, experiment with ways to alter the service to better serve people's needs, and bring more members on board.

Make Anki Useful Again
Spaced repetition has a lot of evidence supporting its usefulness, and I made extensive Anki decks on hoping to capture that value. In time, I learned that to use that value I had to respect its limits; efficient memorization could not replace proper *understanding* and *prioritization*. I will clear out decks not relating to current or projected objectives and alter my current reading followup system to take advantage of the incremental reading plugin.

Trends and Disagreements in the Rationalist Community

Edited from my personal notes format again.

Question: What are the trends and disagreements in the rationality community this month?

Motivation: I want to know what's up, and I want to make a difference. Avoiding the valley of bad rationality seems useful. Also the community has a distinct motte-and-bailey going on about whether rationality means "whatever wins" or the flavor of the LW sequences and diaspora.

That's a fine thing to have, in one direction: we can talk about our latest and greatest ideas, play with those flavors, invite others to join in, and when we encounter disagreement we should rightly take a step back into our core principles to ask "Wait, is this actually what wins?"

But I sometimes see it happening the other way: using the "whatever wins" banner as an enticement to get others to join in, and use our supposed intentions as a general defense for our methods. Methods we aren't centralized enough to even all agree on.

Method: Look at some major blogs. Give a quick summary of each and try to name our felt sense about a) how authoritative they are and b) whether they agree with each other.

Sources: Lesswrong Main, Lesswrong Discussion

Answer: Lesswrong main: seems to navel-gaze lately. It rehashes. It struggles to curate itself. Latest posts are Meetups, trying to distill CFAR stuff, estimating the value of people in Effective Altruism, LW karma charts, speculating about medical conditions, gamifying goals.

LW discussion: I see people chatting about LW's decline, but what do they do about it? I see continued argument about the place for politics and productivity hacks.

AI fermi paradox speculation. They make a background assumption that FAI is hard, unsurprisingly.
AI safety. Using analogies to the past, notable inability to bridge inferential gaps with people out of the OB/LW sphere

High school advice: Get GED and skip to college, or leverage private school. Get college's large wealth and power boost. Meta advice: break down skills important to making choices in
general (value awareness, prediction accuracy, agency) and check that you're stacking your character.

Media to follow... gwern's recommendations well-received

PSA about air quality, gathering what-should-be-universal statistical literacy, trying to propagate lists of high-value PSAs in general

Disagreement: is altruistic deception necessary?

I'm nowhere close to done but I'm way over time.

Sanity-check:
LW is decidedly not the zeitgeist, and it knows it's not.

People are complicated, reality is complicated, it would be surprising to have succint summarizations.

Could I do a poll of what people read, what 'tenets' people work with? What such polls already exist?

Have I spent five minutes actually imagining what a group both coordinated and flexible enough to all change their minds on a dime would actually look like?
.... Damn, it would look like the mormon church.

Followups:
Featured/Promoted: try to establish what canon is well-known at the moment

Refine what it is I'm looking for.

Old LW surveys, SSC surveys, see if gwern does surveys

This tumblr post by Scott Alexander about rationalist tumblr

March 2, 2016

I imply that you imply that I imply that you suck

I made a note on a Facebook event to warn people who have issues with pets and small children that this event is likely to include pets and small children. I tried to word it in a way that would get people to contact me to deal with their concern instead of just not going, or suffering in silence. I got a pretty strong backlash, paraphrased:
"It's worded like an offer to sequester babies away from the party. I feel uncomfortable with phrasings that diminish their personhood, or imply their existence in their own home is a burden on others."
Well, I'm uncomfortable with the implications of this objection!

Don't get me wrong, I understand problems with people considering me a burden to interact with. It's awful. I have a hard time believing in my own worth. I'm continually finding old scars in places I didn't know existed. I don't want to cause that for someone else.

I did reword the note. I wanted to make sure potential issues were resolved ahead of time instead of by last-minute ultimatum, and it wouldn't accomplish my purpose if it pissed off the housemates.

I still resent the implications of this objection.

It tries to provide a protection for personhood by outlining the zone where they cannot be a burden, because, by definition, it is neutral ground to meet all their needs in that zone instead of someone else's. In the process, it runs roughshod over all my attempts to protect my own personhood and respect others'.

When I hear others talk about burdens they seem to treat it as an absolute scale instead of a relative scale. Someone is or isn't thought of as a burden, with some gradation for "how" burdensome. The label clearly matters to people. Being a burden is supposed to be terrible. For the longest time, I thought had to do everything in my power to not be one.

It makes me want to laugh and cry and scream all in one moment when people try to tug on my heartstrings by pointing out how awful it would be if there was no safe, neutral ground to interact with people from. Regardless of whether an interaction is "worth it", regardless of whether I want the interaction or actively try to disengage, it almost always costs me something to interact with people. The most obvious cost is sanity: too much stimulation with no recovery time leads to mental overload and shut down, and this cost is very much additive. The idea of a person who does not come with a cost to burden is *ludicrous* in my experience, but everyone I knew told me it was an achievable standard and failing to meet it was monstrous.

Life started going much better when I decided to to minimize the costs on me and let other people negotiate the costs to them. My neutral ground is the ground I'm standing on. I cannot interact with people as people if they're going to insist that the costs they impose don't count and can't be negotiated. Those aren't agents, those are natural disasters.

January 25, 2016

With your eyes closed, tell me how to snark

I randomly picked an interview question to answer badly. The permission to give awful answers allays the terror of not finding the perfect and true answer, which can paralyze me to the point I never answer at all. The question I got was a lot more random than I anticipated... I answered it anyways.

With your eyes closed, tell me step-by-step how to tie my shoes.
  1. first reach down to your feet.
  2. check if you have shoes
  3. check if your shoelaces are untied
  4. if they are, tie your freaking shoelaces you're an adult gosh
Note, that's pretty uncharitable. What if it's a job interacting with disability or children? However trivial the example, plenty of jobs need you to demonstrate a strong ability to break down tasks into concrete parts and deliver them as clear instructions. Not to mention, some people just don't appreciate a good snark and/or will count my unwillingness to jump through dumb hoops on command like a purebred show-monkey against me.

In any case, showing the concrete steps and practicing each part with on-the-spot corrections would get the point across much more clearly. I have to know my audience to tailor my instructions. I won't pretend that I'm amazing at this task in the best of circumstances, it's worse when you deliberately deprive me of the tools I'd usually use to do it.

And I still like the snark answers better.

January 16, 2016

Nice things for a struggling friend

In this post, I will experiment with modifying my personal research notes to a sane, shareable format.

These notes have a story to them. A while back, I stumbled upon a piece of brilliant advice: “Here’s how to write your thesis without suffering: every day, write down an experiment you ran and record the result, even if all you did was answer a question using a Google search”. Working with ideas as coherent and novel as theses and sharing them with the world excites me, I figured that writing down the results of my google searches really WAS the least I could do towards that goal.

I admit, this example does not lend so well to sharing. My personal version contains more specific  and identifying information which I'd prefer not to share. I edited this version to showcase a more general solution.

Question: 

What are some nice things I can do for a struggling friend?

Motivation:

I have a friend I predict is having a not-so-great time right now.  I know from personal experience and reading many blog posts about secular grief that "Ask me if you need anything" generally doesn't lead anywhere. They might feel uncomfortable asking. Even without awkwardness, it's a hard task figuring out what exactly people can help with.

Method:

Google, alter query as needed

Sources:

Here I would generally include links and describe blurry recollections of where I'd gotten the supporting notions. I will not include links in this version since they relate to a specific, maybe identifying, problem.

Many secular blog posts about grief - When you do things to actually assist an overwhelmed person, they remember even the smallest instances more vividly and warmly than all of the kind words.

The concept of 'next actions' from Getting Things Done, and expansions on the idea from CFAR - Tasks you aren't sure how to act on leave an open loop running in your head, sapping your attention.

Personal experience - I flailed about helplessly, wishing beyond hope for someone to do something, and could think of nothing to say when anyone asked what they could do to help. When I make a vague suggestion to a friend, the ensuing back-and-forth often gets interrupted before going anywhere.

Answer:

TAKE INITIATIVE, as I already knew. Do stuff like "I'll be over at 7 pm Friday to pick you up and treat you to dinner," rather than "So maybe we should do dinner sometime..."

Seek them to hang out with, they may try to isolate.
Specifically: invite them personally to social events you know about. Offer them a ride places. Take them out to lunch or coffee. Call them up or IM them.

Reach out to talk with them after anything biggish.
Specifically: A birthday or anniversary of someone they lost. Filing taxes, paperwork. Kids just left back to ex-spouse's house. First date. Anything big enough that they brought it up in your last conversation.

Do something physical, exercise is the classic anti-depressant.

Cook a meal to deliver. Drop by to help with chores, groceries, dinner.  Help clean the house.

Send an email, share something from the web, let them know that you were thinking of them

Follow-up ideas

How do I even find out where my friends live?

Can I generalize this solution? - I did... for this post.

Try with a few friends, see if it actually works.

What's the likelihood of this backfiring horribly?

Send an email 'following up' our last visit with a random Reddit in the postscript