Study Hacks Blog

Less Than Ultra Learning

As part of my craftsman in the cubicle project, I spent this past week monitoring how I learn new information.

I wasn’t impressed.

At one point, for example, I needed to dive into a topic I didn’t know much about: how information disseminates in random power law graphs. I went to Google Scholar and begin downloading papers with promising abstracts. I printed three and skimmed another half-dozen or so online. In retrospect, I think I was hoping to find a theorem somewhere that described exactly what I was looking for in notation I already understood.

Not surprisingly, I didn’t find this magic theorem. The two hours I spent felt wasted. (Well, not completely wasted, I did learn about the Riemann Zeta function, which turns up way more often than you might expect.)

This experience recommitted me to cracking the code of ultra-learning. Mastering hard knowledge fast, I now accept, requires more than blocking aside time on a schedule; it also demands technique.

The Chair

With this in mind, here was my first stab at cracking ultra-learning:

I bought a traditional leather chair (a longtime dream of mine). My wife and I still need to add some bookcases, a rich rug, and an old brass lamp — but my general  theory here is that this library nook will be make it impossible to avoid mastering new bodies of knowledge, and perhaps also pipe smoking.

Under the assumption that I might need more than the power of The Chair to become an accomplished ultra-learner, I do have one more strategy to deploy — which is what I want to talk about in this post. It’s actually a strategy I’ve known for years (my PhD adviser taught me soon after my arrival at MIT), but have seemed to forgotten recently.

Read more

A Novel Dissertation

I began working on my PhD thesis in the summer of 2008. I defended a year later, in early August, 2009.

There’s nothing unusual about this timing. What was unusual, however, was my approach.

By June 2008, I had a fair-sized collection of peer-reviewed publications. The standard practice in computer science would be for me to take the best of these results, combine them, fill in the missing details, add a thorough introduction, and then call the resulting mathematical chimera my dissertation.

To me, naive as I was, this sounded like a waste of a year. So I decided I would prove all new results.

This strategy worked fine for a while, keeping me engaged and happy, but then, in April, 2009, things took a turn toward the difficult. It was during this month that I accepted a postdoc position that would start in September.  This meant that I had to defend my thesis over the summer. Suddenly the allure of tackling all new results began to wane.

Here’s a scenario that became common:

• I would be working during the day on an important proof.
• At some point in the late afternoon I would find a flaw.
• A helpful voice in my head would point out that my whole future depended on finding a fix — without a fix, it argued, the thesis would crumble, I would be kicked out of graduate school and end up homeless, likely dying in a soup kitchen knife fight.
• After heading home, I would continue, obsessively, seeking a fix — ruining any chance at relaxation that night.

After two weeks of this exercise, I decided something needed to change.

It was then that I innovated my shutdown philosophy…

Read more