Matt Ambrogi

Recurse Center Weekly Recap #7

I’ve decided to change the format of these weekly reflections. I’ve also decided that there will be no format. I’m going to use this space to share whatever is on my mind at the end of the week. I'll still focus on how I’ve been spending time at RC and what I’d like to do next.

This past week was unique. I spent the first few days reflecting, writing, and reassessing my priorities. I dedicated the rest of my time to meeting Recursers from the new batch and beginning a thorough reading of Skiena’s The Algorithm Design Manual.

Although less technical, it was still a productive week. Here’s some things that stuck out:

  • Reflected on my time so far at RC. Shared that reflection as a blog post.
  • Put finishing touches (for now) on my Strava Analysis app. Notably, adjusted plots to that they look ok on mobile.
  • Had some fun discussions with Recursers, including one about building tools to help scientists explore data.
  • Got a solid start on The Algorithm Design Manual.
  • Had a very inspiring talk with an old friend about programming, tools for thought, and how to a live a good life with an old friend.
  • Put together and shared a little presentation on my Strava Analysis app.

It felt strange going a few days without programming or working on a project. But I’m hopeful that the clarity gained from reflection will make the next weeks more productive than they would have been without it. I’m looking forward to getting back into a groove this week.

Back in that groove, I’ll be focused on two things:

  1. Reading The Algorithm Design Manual
  2. Beginning to explore machine learning resources and projects

Regarding (1), I had planned to start practicing algorithm problems this week, but I’m going to push that back and dedicate one more week to reading ADM. Having a strong understanding of the fundamentals is important to me. I'm also optimistic that the time spent reading will pay dividends in the future by giving me a more powerful set of principals to reason from. This sounds a bit tedious, but I'm actually enjoying reading the book.

Still, diving into ML is probably more exciting. I’m going to allow myself to poke around and explore a variety of resources before settling into any substantial project. Hopefully I’ll get there by next week. This week I’ll be going through blog posts, resources from Open AI, and this course from fast.ai.

This week I’m also going to run structure of my days experiment. The experiment is simple. I'm just going to commit to one simple thing everyday: wake up and do focused work (reading ADM, ML resources, or programming) for 3 hours before I do anything else.

This isn't so different than how I have been working. But there is a one notable difference. For the past seven weeks I’ve allowed exceptions, maybe I go to the gym before work a few days, check my calendar, look at email. This week, no exceptions. To paraphrase Tim Urban (2 hrs in this conversation), by doing 3 hours every morning, I can ensure everyday will be at least a B for productivity. Which means the week will also be a B. That’s a pretty good lower bound.

Lastly here are some things on my mind:

  • How do you know when to move on for a project?
  • Should I be deploying more of the toy apps I’m building? Probably... This is a perfect example of leaning into stuff I feel uncomfortable with.
  • How do you best create systems that encourage you to work at the edge of your ability?
  • Is reading (docs, tutorials, theory) up front necessary? Or is it almost always better to just try and jump in?
  • What would anxiety free productivity look like for me? How do I get stuff done but have minimal residual anxiety or the feeling that I should be doing more? Can I get there just by following what I’m most interested in?

- 1 toast