Let me give an example of an approach I use, it may help trigger some ideas of what will work for you (a 🧵):

I took mostly science courses as an undergraduate. As a result, I had weak writing skills. Law school exposed those weaknesses in part, but the real world ... https://twitter.com/IvyBGrey/status/1391089786648211460
revealed them in all their glory.

I worked hard on improving my writing skills and had some success (though I'm by no means an above average writer). Over my 40+ years as a lawyer, I realized that many lawyers had weak writing skills. I worked with those on my team ...
to improve as I continued to work on my own skills. But I started asking some questions. My area of focus as an undergraduate today would be called "cognitive science". So I asked: How do we learn to read and write? Why do I make the mistakes I do? Why do I write ...
with a style and voice different from what other writers use?

Today, as you know, computer scientists have delved into the area of reading and writing. But they approach it from a (mostly) statistical perspective. Put simply, the improve pattern recognition, they don't ...
"teach" computers to understand what they "read" or "write". I know something about statistics and computer programming. I understand (at a basic level) what they are doing and why. But CS has started to stall on machine learning and AI. Statistics will take them only ...
so far. To go further, we need to understand at a rudimentary level (cue Elon Musk: first principles) how humans lean to read and write. How do we learn to understand?

At this point, I've connected (meaning sewn together my interests) in CS, psychology, law, ...
and statistics (at a minimum). My overall focus: What would a program look like that did more than offer corrections to your writing based on rules and probabilities? Could we design a program that would learn, e.g., how Ken writes? Not just predictions, but actually ...
understand how Ken reads and writes and offer suggestions based on Ken's style and voice. Or Ivy's, or ...

Will I answer the question? Probably not. But along the way, I'll have fun and learn things that might help me design a program that could replace Word. ...
Also along the way, I've had to prioritize some areas of interest by saying: "I'll come back to that interest later." I do some pruning to keep myself moving forward. I try to lean into my strength of seeing the picture rather than all the dots of paint.

I've used this ...
(very rough) analogy from time to time. Beagles have a great sense of smell and insatiable curiosity. Let an untrained beagle loose and it will follow its nose from scent to scent until it can't go on anymore. The beagle will have enjoyed it's journey, but not ...
accomplished much. Train the beagle to follow only those scents that will take it somewhere (e.g., to tree a bobcat) and the beagle will have fun as well as accomplish something.

I want to be that beagle. I want to have a fun journey and accomplish something at the end.
You can follow @LeanLawStrategy.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: