A year of milliblogging

A dreaded semiannual process blogpost.

On July 1, 2023, Elon Musk decided to rate limit Twitter, rendering the website unusable. No one knows why he did it, but Occam’s Razor suggests it was because he didn’t want to pay his cloud computing bills. Whatever the case, that website was a nightmare. I had been looking for an excuse to get back to blogging and away from Tweeting, and Musk turning off his multi-billion dollar website for a few weeks was as good a forcing function as any. 

I decided to start blogging here on substack. Initially, my thought was to use this space as a Twitter replacement. To do some sort of Twitteresque microblogging. I had been a sporadic blogger, but I used to slave over posts, often for weeks. This led to some pretty good posts, but I wanted something a bit more like Twitter without Twitter. A place where I could write spontaneous thoughts and post “questions, rants, and shitposts.” Thinking about it now, I was pining for a 2003 blog.

Being a process-driven guy, I made up some arbitrary rules for myself. My idea was to post Twitter threads with better grammar on Substack. If a tweet was 140 characters and a Twitter thread 10 tweets long, writing a 300-word note every day seemed pretty doable. I set aside my morning coffee time to write and forced myself to post after an hour. I wrote,

“I’ll still give this milliblogging a go for a week or so. Let’s see how long this lasts!”

Here we are a year later. I promised myself I’d write a reflection if I made it a year, so that’s what we’re getting today. I promise you that I’ll limit the number of these naval gazing writing posts to one or two a year.

I have been pretty good about using every morning to try to write. I spend at least an hour with Google Docs open, and—most days—avoid other doomscrolling distractions. I have found the morning blogging similar to practicing an instrument. The consistency itself makes the process easier. This is so cliche, but it’s true. I bet you can even find convincing science-y studies proving that it’s true. Blogging has made writing other projects flow better, even the ultra-tedious ones professors are regularly saddled with.

The blogging has certainly been time-consuming, but I’ve tried to rate-limit it like a Twitter engineer. If I don’t feel like a post will be ready at the end of the hour, I push it to the next day. On all days, I keep my blogging to under two hours. Though I had been shooting for about 300 words, I’ve found my natural cadence, where I can say what I want to say, ends up being about 1000 words. Today’s post is 908 words. 

My arbitrary rules of practice explain the volume you get from me. I’m able to get about a thousand words I like about three or four times a week. I could, I suppose, wait until Monday and send whatever I had written the week before. But that would be against the spirit of my initial microblogging project.

In any event, this has been an exceptionally beneficial experience for me. I’ve been able to think more deeply about my teaching, explore weird research topics I might have neglected, and rant about topics I didn’t want to cast in stuffy academic arxiv papers.

So thanks to everyone who puts up with the newsletter volume and follows along.  I imagine getting 3-5 emails from me a week must be annoying. Maybe you like email more than I do. I like everything about Substack except for its “newsletter-centrism.” Maybe Substack can recreate Google Reader? I suppose in many ways it has. Regardless, know that I appreciate every one of you who put up with the email volume.

And extra thanks for your feedback in emails, comments, and Twitter replies. I’ve learned so much from readers in so many surprising ways. The Substack discourse has been far more productive for me than on any microblogging site, whether it be Twitter or Bluesky or Mastodon. So far I only have two people who have sworn to never speak to me again because of my blogging. That seems better than the rate of enemies you gain Tweeting.

Let’s see if I make it to July 1, 2025. I’ve got a plan to get me to January. First, I need to finish these Meehl lectures! I had thought this would only be a few weeks of blogging, but there’s so much depth and complexity and so much valuable perspective on the state of science, engineering, and decision-making. It’s definitely worth seeing through to the end! I think I’ll need until the end of July to work through those lectures. After that, I may take a short break as I have a manuscript deadline on September 1.

In the fall, I’m teaching a course I love for the first time in over a decade: Convex Optimization. I want to try live blogging it, following my process from last fall. I’m looking forward to the challenge of writing about Mathematical Optimization with almost no equations. The material is decidedly more technical than machine learning but has so many interesting, practical applications. What is the role of Convex Optimization in the era of LLMs? Tune in to find out!

And after that, who knows? All I know is I hope you’ll stick around to find out.