Some things that jumped out at me from what I've been reading over the past week or so:
- "Randomized Controlled Trials with Minimal Data Retention" (Winston Chou): The author discusses methods for recursively estimating key statistics from A/B tests (e.g. mean, variance). Why is this important? Because it reduces the amount of time that we need to store anyone's data to be able to do standard online data science. I've been trying to check my understanding/stress test the content of articles like this through little simulations (here). Doing the simulations was really useful. One reason is that I found a typo (or at least a misunderstanding on my part) in one of the formulas that led me learn about the history of these types of algorithms (see Chan et al. 1983). They weren't originally developed to preserve privacy. They were created to do computations that weren't possible on single machines back in the day, but are now trivial (e.g. finding the variance of a big vector of values).
- The Advice Trap (Michael Stanier): This book has actionable advice--irony acknowledged--for coaching. The key message is to stay curious longer, avoid jumping in with advice. This has a number of benefits including making sure that you actually understand the problem and helps engage the person being coached more and develop a similar understanding in them. One way to do this is to ask more "what" questions. Here is an example of how this has already been super useful for me in my personal life. I used to ask my son at dinner: how was your day? Usually I would get the non-informative answer "fine". Then I tried "what" questions, e.g. "what did you build today". The answers were much richer, like "first I built a medium sized tower, but my friend accidentally broke it. . . "
- The Golden Thread (Kassia St Clair): a wide ranging book on the history of fabric. I'm about halfway through. Highlights for me so far include (a) the choice of wool for Viking sails is pretty counterintuitive (wool lets through a lot of air and soaks up a ton of water), but through very laborious processing it becomes a great sail material and (b) "zoom" was a Persian measure of silk equal to two bales or about four pounds.
- Home Computers (Alex Wiltshire): a coffee table book with pictures of computers (mostly) from the 1980s. Just a fun book with lots of great 1980s product names like "Exidy Sorcerer" and "Robotron KC 87".