I think having a bag of tricks is a rarely discussed (maybe less impressive?) lens to see skilled craftsmanship at nearly anything, from cooking to jazz improv to mathematics.

Computer Graphics #

• I don’t even know all these! off the top of my head:

  • surfaces (e.g., metal / reflections)
  • water movement
  • water reflections
  • anti-aliasing
  • tonemapping (color grading)
  • sky box / shader
  • textures + LOD
  • shadows
  • I’ll just say “all of lighting” (+ lots of tricks to get w/o “true” (global?) illumination)
  • bloom
  • motion blur
  • depth of field
  • procedural movement (grass, trees, birds, etc.)
  • artistic direction
    • idle animations
      • (sidestepping “all of animation” is here)

• some small examples

  • shade web engine
  • tunic visuals
  • thumper post-fx

Cooking #

• pasta

  • don’t break pasta
    • texture preference
  • cold water doesn’t matter
    • probably historical artifact of hot water picking more stuff up in pipes
  • don’t need much water
    • like all classic recipes say use tons of water
    • you can use a freaking pan
    • this in fact can help because you get higher starch concentration in water, which is then great for sauce-making
  • salting water matters
    • based on (a) cooking time, (b) pasta water usage
    • can actually taste this
    • when tasting, use cold spoon and blow on it else burn tongue so easily
  • adding other stuff to water (herbs, oil) does not matter
  • use pasta water in sauce to help cling
    • this felt like the biggest secret advantage several years ago and now everybody knows about it
  • cheese + stir + pasta water + heat can help thicken
  • pulling before pasta is fully done
    • base this on the amount of time it will be hot while making sauce with it after
  • combining sauce with pasta
    • rather than serving separate
    • so you can make a “sauce + pasta” dish in a way that won’t happen on plate
  • using a separate container to reduce residual heat for late-stage ingredients you don’t want to cook
    • pesto, eggs, lemon juice
  • undercook even slightly more for leftovers
    • seems to soften in fridge
  • whole additional world here for different sauces

• (so many other examples here)

• I think a class of tricks is understanding the leeway in different aspects of your dish. Knowing, for example, that you can keep cooking the chicken thighs (but less so chicken breasts) to get a good crisp on them even after they’re at temp because they’ll be robust to it. Or that you can salt a tomato slice with a tiny to huge pinch of salt and it’ll be OK.

Jazz Improvisation #

This is such an enormous one, that simply understanding it and doing the basics will elevate your playing an entire tier. I can’t believe nobody ever taught me this. Maybe it’s embarrassing to say out loud.

Just memorize many licks. Transcribe them. Eventually make some up.

• Learn them in the relevant keys
  • eventually all. but don’t sweat this at first!
• String them together
• Some will end up “meta-licks” you can riff off
  • Specifically, I’m thinking about purely rhythmic patterns, licks with intervals that don’t depend on their position in the cord, or “lick templates” where there’s some easy variations
• While you’re performing, once you know a handful of these, plus quote the song or other songs, maybe eventually some genuine in-the-moment creativity, then you’re doing it.

I had a few different jazz teachers and they got close to telling me to do this by having me transcribe entire songs (including solos) played by famous musicians. Then, re-playing these entire songs. But this is so different from building a vocabulary of licks!

If you listen to famous jazz solos you can even hear the musician’s favorite licks.

Math #

See “Every mathematician has only a few tricks” in Ten Lessons I Wish I Had Been Taught by Gian-Carlo Rota (1996).

Here’s a mathoverflow post collecting examples of these tricks.

This is one of the crazier ones to me because I think of mathematics as the most difficult intellectual pursuit. And even they have tricks! (I’m sure it’s still hard.)

Woodworking #

TODO: Ask woodworking friends

From my limited experience, you run up against similar problems often. One meta-trick is to make your own tools to help you achieve common tricks you need to pull off.

Meta-analysis #

What’s going on with this idea of many little tricks?

• Apprenticeship — helpful because you learn someone’s tricks

• Bottom-up — can you just learn the tricks? I suspect you need to be learning the tricks situated in a good learning environment:

  • observed measurable outcomes
    • measurable can be extremely coarse: did it taste good? did the solo sound good? does my video game look good? or very precise: did i cut the length of wood correctly?
  • repeated practice
  • some experimentation allowed

• Top-down — maybe the whole idea here is top-down modes of working (developing a “grand theory” then trying to have the actual practice fall out of it) fail because you actually need these tricks

  • actually, the tricks are maybe just lower level of abstraction details about working
  • in other words, in becoming skilled you will develop ideas about different abstraction layers of the task. to be very simple, let’s just describe two:
    1. vision / grand unifying theory
    2. tricks
  • … but in a sense, your tricks may apply to layer (1) as well! e.g., “what if we turned it upside down / made it sepia tone / added a timer / etc.”
  • in other words: tricks at the lower layer of working are easier and more concrete to understand. but once you’re proficient at them and making higher level decisions, you can come up with tricks at higher levels. however, higher level tricks are probably less universally applicable and more stylistic
  • this all leads me to (see next point)…

Grand Unifying Theory of Tricks #

Grand unifying theory of tricks:

1. Tricks are just repeatable decisions in layers of creation.⁰¹
2. The more experienced you are, the more abstract of a layer you’ll have tricks for.

To make this concrete, say you’re composing a song. There are many levels at which you’ll make decisions, and you can probably learn tricks for all of them:

• song context (e.g., scene placement in film score, album position / requirement, single’s purpose)
  • song idea (mood, point, vibe)
    • core thematic components (licks, sounds, rhythms)
      • structural pieces (key, tempo, sections)
        • tracks (instruments, vocals, samples)
          • notes
            • production (all FX / mixing / mastering)⁰²

I’ve never tried to break down composing a song hierarchically before, so probably nobody thinks about it like this, and I don’t think anybody uses this ontology to actually write music.

What I’m claiming is: experienced composers probably have tricks that fall into many different levels there. And the more experienced you are, the more abstract of a layer you’ll have tricks for.