To Make Unbounded Progress, Do Similar Activities to Past Successes

I see people overreach and do overly difficult activities even after reading my essays about overreaching and claiming that they agree with me. Then they may fail a bunch, get stuck, and give up without ever trying easy enough activities. Why do they do this? Perhaps they don't understand my essays and don't know how to have productive conversations about those topics. Also, many people think they won't get stuck as others did, so they think it's OK for them to do hard, ambitious, complicated stuff without working up to it. Then when they do get stuck, they don't change their attitude effectively, so they stay stuck.

I think that writing more about effective learning processes can help everyone, stuck or not.

Judging the difficulty or failure probability for activities can be hard. People often misjudge it – especially beginners who have not yet had much success with anything philosophical.

What is a simple way that people can judge what activities would work well for incremental progress for them?

Only do learning activities that are similar to past activities that you succeeded at. If you can't point out major similarity between your next activity and a past success, don't do it. (Do more planning instead – but keep in mind that planning should be the minority of your time spent. Find something reasonably small and easy that you can do, that doesn't require much planning, and just do it! Repeat that a lot!)

How similar should a project be to past successful projects to count as "similar"? As a simple guideline, it should be similar enough that most people in your city would consider it similar if asked. Assume they would give a quick, intuitive answer without over-analyzing and without listening to any lengthy logical arguments from you.

(What about exploratory projects like reading a book on a new topic to see if you'd even be interested in learning about that topic? If you've done an exploratory project successfully before, and read some books successfully before, then you have clear similarities to prior projects. There are many dimensions of potential similarity.)

Any adult (or teen, but not infant) can (theoretically) become a great philosopher, economist, scientist, businessman, sports player, chess coach or whatever else by a series of steps where each step is similar to a step they've already succeeded at. There's no logical reason preventing such a series of steps from existing. Nothing about the structure of knowledge requires big, hard jumps. There isn't a bunch of irreducible complexity that can't be broken up into smaller parts. (You may already be familiar with irreducible complexity because it's one of the failed attempts to argue against biological evolution.)

This helps explain what I mean when I say that no hard steps are necessary for unbounded progress. Every step can be small and easy. Just doing things that are mostly similar to previous successes works for unbounded progress.

Will progress be slower due to small steps? No. It's much faster than people are used to because they'll actually make progress instead of getting stuck. Also, small steps take less time per step, so they aren't inherently slower than using bigger steps. You can imagine 500 small steps taking approximately as long as 50 steps that are 10 times bigger. But in real world scenarios, smaller steps tend to be faster.

Can steps be so small that there's an efficiency loss? Yes but that's not the problem I see anyone actually have.

Does it waste time if steps overlap with previous successes? Do you lose time repeating what you already learned? Basically, no. You don't have to relearn the part you already know. Already knowing it saves time and lets you finish faster. If you already know part of a step, then it makes the step smaller – you only have to do the new stuff.

I don't know of an abstract principle that makes progress rates don't depend on the size of the steps you use. However, if the steps are big and increase your error and failure rates, then they do lower your learning rate. And if the steps are too disconnected from your existing knowledge, even if small, that also raises your error and failure rates. So use small steps that overlap with past successes. They have basically no downside and avoid common ways people fail.

In general, if you want to learn and progress, you should figure out what knowledge you already have, then figure out ways to build on it. This mental model of learning is much more effective than trying to figure out how to directly learn something hard and complex like philosophy, without relating it to your current knowledge. You need to grow your knowledge from where it is now to include philosophy, not create a new, separate body of knowledge for philosophy. Your current knowledge – stuff that you already know and can succeed at – can be called your "baseline". It's important to test your beliefs about your baseline by actually successfully doing some stuff you think you can do. People often think some stuff is within their baseline, don't try doing it, and would fail if they did try.

A good learning plan involves figuring out what knowledge you currently have, then figuring out how to get from here (your current knowledge) to there (your goal knowledge), step by step. You start where you are now (current knowledge) and take steps, from there, in the direction of your goals. If your goals are far away, divide them into sub-goals (and sub-sub-goals, and sub-sub-sub-goals, as necessary) so that there are some goals that are within view and reachable without a ton of steps. How many steps away is close enough for a sub-goal? You can answer that by looking at your past successes. What's the most steps that worked before? If you've done lots of projects with a goal that was 10 steps away, and succeeded at almost all of them, then up to 11 steps would be fine, and a little more might work.

Don't try to plan out every sub-sub-sub-goal at the start. There's flexibility, but the most typical or standard approach is to figure out a series of reasonably high level goals from your current knowledge to your goal. Include just enough sub-goals and details for each step to have some confidence it will work. Then fill in more details as you go along. Start on the first sub-goal, and figure out sub-sub-goals for it, and if they are too big then take the first one, that you're about to do, and divide it up more. Whenever you actually start doing a (sub) goal, divide it up into smaller parts if necessary (and when you start doing each of those parts, divide them up if necessary). Do just-in-time step division instead of planning everything from the start. Also, when it comes to written planning, a small enough step means you can handle further divisions without written notes; it doesn't mean there are no further divisions. Written planning is meant to aid your ability to keep track of things and deal with complexity. Planning should break things into small enough chunks that you can handle them with a high success rate, not into the smallest chunks you can think of.

What counts as a small increase over past success? It depends on context, but as a very loose rule of thumb you can use 5%. (To get a doubling with 5% increases, with compounding, takes 15 increases. And it's 20 increases to double without compounding.) Aiming for increases above 20% is suspicious and often unreasonable (and I see people do that a lot); that's not aiming for incremental progress via many small steps.

How can you decide where milestones (the ends of sub-projects/sub-goals) can go? Milestones can go anywhere that you're able to judge success or failure. You want to have frequent steps where you are able to evaluate whether things are working or not. That way, if things aren't working, you don't spend a long time and do a lot of steps without realizing it. Virtually everyone uses too few milestones; I don't think I've seen anyone use too many; so loose guideline is you should use every milestone you know of that will work. But if you do a great job of brainstorming dozens of milestones for one project, then be more selective.

Another way to view breaking down sub-projects is you should reach a sub-project size involving one practice activity. If you aren't practicing anything, you should be suspicious. And if you're practicing multiple things, then it'd be better to make multiple separate judgments about success and failure – one for each thing you're practicing. So break it into smaller chunks to separate the judgments, so each chunk can focus on reaching success at one type of judgment.

One of the most important drivers of success is that your evaluations of milestones are accurate. You don't want to think you've succeeded when you failed. What helps with this? Do projects similar to past successes. If you did a project, succeeded, and knew you succeeded, then you're in a better position to also judge a similar project accurately. In general, you'll have a better chance to judge a similar project than a dissimilar project.

The phrase "incremental progress" means making progress in small increments (increases, steps) from where you are now (what you already know, the progress you already made in the past). Karl Popper's epistemology advocates incremental progress (also called piecemeal progress or gradual progress). It's one of Popper's many ideas that Critical Fallibilism agrees with, builds on and expands on. Popper never wrote an essay similar to this one, but I think this essay is compatible with what he was saying; I think it's adding new information and analyzing more details rather than contradicting him. Incremental progress is also a fairly well known idea that has been advocated by many other thinkers besides Popper. I've seen people (who've never heard of Popper) have success with incremental progress when learning things other than philosophy (e.g. art, math, sports or speedrunning video games), but in philosophy people tend to do it poorly (ambitious steps that try to tackle complex issues; being content with unsolved problems instead of doing a succession of problems they can actually solve to make progress; not doing practice drills; having few milestones; trying to develop ideas where they have little ability to objectively judge if their ideas are correct).