I think this was one of the better meetups that I've run. I'm optimistic about this exercise becoming a pretty core LW/CFAR-esque cornerstone. It seemed worth a retrospective. tldr:

• I could have explained the step-by-step-instructions a bit better (see below for what I think they should have been).
• We had 20 participants. I think roughly 6 reported getting something clearly useful out of the exercise. My rough guess is another 4-10 would get something useful out of it with a bit more practice and/or clearer instruction.
• When the exercise is working, it is obviously working. (I've done the exercise about 4 times. Two of those times involved floundering slightly and making subtle progress. The other two involved clear 'aha!' moments.
• I think the subtle progress was worthwhile and pulled roughly the same weight as other cognitive exercises (which I think tend to yield something like a 3% return. See Sunset at Noon and 3% Improvement). Over time this adds up to something very powerful, although hard to check progress on.
• The aha moments felt like major discontinuities that were immediately useful.

I'd summarize the exercise and philosophy, but honestly the original blogpost did a very good job being succinct. So I'm going to focus mostly on how I operationalized it and how I currently understand it.For easy reference, I'm including what SquirrelInHell listed as the key insight:

• Your brain has the ability to update its cognitive strategies, but the usual mechanism has unnecessary levels of indirection, ie:
• Cognitive strategy -> Thought -> Action -> Reward or punishment
• You get rewarded or punished for what you do (as measured by your brain's chemical responses). Good thoughts are more likely to be followed by good actions. Good cognitive strategies are more likely to generate good thoughts. On average, your brain will slowly update its cognitive strategies in the right direction.
• Cognitive strategy -> Thought -> Reward or punishment
• You have learned to be happy or unhappy about having certain ideas, even when you don't yet know how they apply to the real world. Now your brain gets rewarded or punished for thoughts, and on average good thoughts are more likely to be generated by good cognitive strategies. Your brain can update cognitive strategies faster, according to heuristics about what makes ideas "good".
• However, by carefully looking at the "deltas" between conscious thoughts, we can get rid of the last remaining level of indirection (this is the key insight of this whole page!):
• Cognitive strategy -> Reward or punishment
• You have learned to perceive your cognitive strategies as they happen, and developed some heuristics that tell you whether they are good or bad. Now your brain can update cognitive strategies immediately, and do it regardless of the topic of your thoughts.
• Even when you generate a useless idea from another useless idea, you can still track whether the cognitive strategy behind it was sound, and learn from the experience.

Ray's Core Summary of Exercise

My attempt to distill the exercise down:

0. Previous Skill: have practiced at least one style of introspection, such as Gendlin's Focusing or Mindfulness Meditation.1. Mindfully solve a puzzle.

2. While doing so, record your stream of consciousness in as much detail as possible (the ideal goal is to include things on the sub-second level, that includes both verbal thoughts and vague impulses, flinches, desires etc). Recording the stream of consciousness is more important than actually solving the puzzle.

3. Whenever you either complete the puzzle, or come to natural stopping points in the process, review your mental stream and look at the deltas in between mental snapshots. In as high a resolution as possible, pay attention to steps in your mental process were most fruitful.

Some snags in the exercise that needed clarification/operationalization:

1. Thinking in batches

I often found I was solving a subproblem within the exercise, and interrupting that solving process to record my second-by-second experiences was disruptive. I resolved this by "thinking in batches" – let my train of thought run interrupted until it hit a natural stopping point, and then going back into my list-of-things-I-experienced and filling in gaps.

2. Look for positive deltas more than negative ones

One thing a few people ran into was something like "ending up in a loop of useless thoughts", sometimes literally thing "I'm stuck, I'm stuck, I'm stuck" over and over, and otherwise second guessing their worth as a person ("I'm not smart enough to solve this.") Noticing when you're stuck is helpful, but ultimately the thing that's more relevant is "what gets you unstuck."

Eventually, you will become unstuck (maybe by stopping to think about an entirely different thing). Notice what happened in the moment when you became unstuck.

It's hard for human brains to learn not to do things, and easier to learn to do things (i.e. "don't think about an elephant" and all that). So noticing and then reinforcing/rewarding yourself for breaking out of the loop is the most important bit here.

3. Stuck on the 2-3 second timescale

The blogpost says to be looking for the subsecond timescale of internal mental states. I'm not sure whether I actually got to the point where I could identify multiple states per second, and some people at the meetup also struggled with that. (I did get to points where, in a given 2-3 second timescale, I could identify maybe one flash of impulse/insight that took less than a second, but then it came with a 2-3 second batch of thought)

My current take on this is something like "you can get a lot of value out of the 2-3 second resolution. You can (I assume) also get value from practicing noticing what's going on at the subsecond level. I wouldn't stress out too much about which area you're focusing on until you've started exhausting the low hanging fruit.

Some Aha moments

a. "Remember to look for cheap tests"

The first time I did the exercise, I had a hypothesis, and then I figured I'd need to check a large-ish number of things, calculate something, and then check another set of things to find out if it were definitely true. I thought about that for 10 minutes or so.

Then I checked the very first thing, and it completely falsified the line of reasoning. If I had done that sooner, I'd have saved 10 minutes.

This has led to a new TAP of "look for cheap tests".

b. Claim: Spin cycles / willpower are a waste of time.

Something that I didn't quite get the first time I did the exercise, but did get during the workshop, but am not sure if I think is real, but seems probably real...

...is that the claim that spending willpower to think is just a wasteful byproduct that some people accidentally learn, as they cobble together their cognitive patterns as they grow up. Thinking shouldn't be effortful. You should just be able to watch your thoughts happen, the only resource you're spending is time. Any extra effort you put in is misguided wasted motion.

I'm a person who generally gets headaches when I think strategically. For awhile this meant I didn't think strategically. Then it meant I just dealt with headaches, and limited myself to what I could do sustainably. (i.e. I could think hard for about 2 hours before my head hurt). So, figuring out if this was true was important to me.

The guy at the meetup who ended up in a loop of "I'm stuck... I'm stuck..." described it as feeling like their brain was spinning real fast. And I had something of a crystalization of "ah, so is the claim that every time I'm doing a mental motion that feels effortful, I should just... not be doing that?" (which other thing to be doing instead being a bit context dependent).

I dunno, hard to explain exactly what I'm thinking here. I'm not sure I buy the claim that no thinking should be effortful but now I do buy that there's *enough* wasted motion that you can trim out a lot of it.

Important: Bring a laptop.

The exercises we'll be doing involve trying to pay attention to sub-second transitions in your thought process, and I find it much easier to write those down and review them with a keyboard. Pen+paper will be available but if you can bring a laptop that's preferable.