Archive of discussions on Twitter. My current answer to the question, “How best to allocate time?” is “I have a feeling that it is right to do as much as you want to do because the difference in time allocation is overwhelmingly outweighed by the speed of learning due to high motivation. Related: Better off investing in new curves….
- Context: I saw a statement, I think it was, “I have to spend every waking hour programming,” and this was my response to it.
- You spend all of your waking hours understanding “how programs work in a computer,” which is good, but since all actions are choices, doing so means that you don’t spend time understanding “how money works in the market,” “how emotions work in people,” “how information works in society,” and so on. This also means.
- kmizu >Basically, I agree, but I thought it would be more interesting to discuss the issue of “optimizing time allocation” due to the nature that this is sort of a given and that knowledge does not increase linearly with time allocated.
- What kind of curve are you envisioning when you say “knowledge does not increase linearly with time allocated?”
- kmizu> This is a crude diagram, but it looks like this. The growth is gradual until the first time t1 is applied, then the slope becomes larger than linear until a certain time (t2), and then the slope becomes gradual again after time (t3).
- Oh, you share with me the model where the relationship between time and knowledge is an S-shape curve. And similarly, I wonder if there is a property that utility does not increase linearly with knowledge gained, what does this curve look like?
- kmizu> Yes, I agree. As for utility, I think law of diminishing marginal utility “something like” might work, although it is different from the original meaning, but it is pretty textured because it does not properly define the utility of knowledge.
- Surprising. That is very different from my model. In a market where a handful of top executives take the lion’s share of the profits, there’s not much to be gained from halfway knowledge, and it increases rapidly the closer you get to the top.
- kmizu> By the way, as for my initial motivation for replying, it’s pretty obvious that doing one action leads to not doing another, so I thought it might be more interesting to go further than that and hypothesize about “how much” to spend on a certain action (or what it is considered good). Just a thought.
- As for how much to spend, if time vs. knowledge is an S-curve and the utility function of knowledge is linear, then the tangent line drawn from the origin to that curve will have the highest utility per cumulative hours invested. So it is good to stop there. And what about when the utility function is nonlinear?
This page is auto-translated from /nishio/学びの時間配分の最適化 using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.