Thread by @ztellman

For the last few months I've been interviewing for a variety of full-time jobs. Over the last two years, I've had a lot of time to think about what I want to work on, and the best answer I could come up with was "interface design, in every sense."

In any design, the only parts that last are lines we draw between the different parts. The people and software on either side of an interface come and go, but the interface remains. The more foundational the interface, the longer it will stick around (see POSIX and TCP).

I want to work on load-bearing interfaces, where the difference between good design and merely serviceable design really matters. This is why I'm excited to announce that next month I'll be joining the Semantic Machines team at Microsoft.

Semantic Machines is trying to build conversational UIs that are truly compositional, rather than just a 1001 siloed "skills" for playing music and changing the colors of your lights. This involves some deep questions about representation that I'm eager to delve into.

They're a Scala shop, which makes this my second job in a row that doesn't involve Clojure. Given my selection criteria for jobs, it's unlikely this trend will reverse. I realized this late last year, which coincides with my vocal discontent about the state of Clojure.

I hope I'm wrong about this. I hope that Clojure bucks the Lisp curse, and the ecosystem becomes more legible to newcomers. I hope that the people at the helm will come to see community-oriented improvements as a profit center rather than a cost center. But I don't expect it.

In the weeks leading up to my new job, I'll be sketching out my next book, tentatively titled "Principles of Software Design", which will refine and build upon the ideas in @elementsofclj such that they're accessible to a broader audience. Stay tuned for more details.