I just use a big text file where I keep all notes and todos (do use a [] as well), separate dates in a structured way with most recent date at top. Typically edit it in emacs. It just keeps getting bigger and bigger, but it’s nice to have a single text file that I can search.
A lot of dynamic Econ models have some condition where my choice today depends on an expectation of the future. Like how much I choose to eat today depends in part on my best prediction of how much I’ll eat tomorrow. For aggregate decisions, getting better predictions should improve welfare.
The Chamley-Judd finding of a 0% optimal capital tax is a very sticky result in optimal taxation theory. One way to think about is you want your tax system to walk as softly as possible while getting from point A to point B. Don’t distort intertemporal decisions if you don’t have to. Don’t tax elastic things when you could tax inelastic things - impose taxes on things where the optimal allocations don’t change much with the new disincentive. A wealth tax takes this result and compounds it. I should really consume my wealth instead of saving it. So you’re rewarding the spendthrift and hurting the saver (the investor). Investment and growth (and overall welfare) suffer. The caveat to all this is that a wealth tax that’s a one-time unexpected confiscation of wealth shouldn’t distort incentives, which is kind of approximated by the estate tax (poorly approximated since its still expected). But the wealth tax is really not about optimal tax theory. It’s about saying the quiet part out loud - taking from the wealthy because there’s a belief that it improves democracy. This belief is unfounded but has high political ROI.
All models are wrong. Some are useful approximations for understanding and others for forecasting. In Econ for example, DSGE models are useful for asking counterfactuals but not great for forecasting. Lots of models are particularly susceptible to assumptions that can be tuned in an Upton Sinclair like way. Climate modeling appears to be one of these areas where the black box can be filled with all sorts of debatable assumptions, yet questioning what’s in the black box carries great reputational risk.
There are MANY CLTs. In general there is a memory-moment trade-off in assumptions for a given CLT. For example, you can relax the iid assumption to allow dependence between observations, such as a mixing condition, with the cost being you need a little bit stronger moment condition where the moment condition is a function of the mixing condition - more dependence (up to a point) requires a stronger moment.