sulami

I've been arguing that "AI" has very little impact on meaningful technical interviews, that is ones that don't test for memorization of programming trivia: https://blog.sulami.xyz/posts/llm-interviews/

Oh, good catch. Corrected.

Thank you.

The clicking to restart is there mainly for phone users, I realized after sharing with a friend that hotkeys aren't very mobile-friendly. I don't have a good way to detect desktop vs. mobile clients, and technically one can also have a keyboard connected to a phone.

I've also been thinking about zooming, I'll see what I can do over my next lunch break :D

  Location: Tokyo, Japan
  Remote: Preferred, but not required, been remote for ~5 years
  Willing to relocate: No
  Technologies: Rust, Clojure, Go, Python, SQL, k8s
  Résumé/CV: https://blog.sulami.xyz/pages/robin-schroer-cv.pdf
  Email: hire@sulami.xyz

Hi,

I'm Robin, (Senior) Staff Engineer with 10 years of experience. On the Will Larson staff archetype chart I register mostly as Architect and/or Solver, and I enjoy figuring out what is needed for organisational success. I've most recently worked at CircleCI on the core pipeline orchestration platform as well as config processing. Relevant experiences here include high throughput distributed systems and compilers.

I'm somewhat conversational in Japanese, but by no means business level (yet). Native in German, and can get by in French or Dutch, if need be.

I had a similar kind of idea for a long time, which I put into action a few weeks ago via a standalone transpiler of Emacs' rx macro to common regexp syntaxes.[0] I ended up getting interrupted and didn't completely finish it, but it generally works, though is probably riddled with edge cases.

The basic idea of rx is to use S-expressions to describe regular expressions, and my elevator pitch would've been to embed rx invocations in shell scripts using $(syntax), the main use case being something like sed invocations.

I still think it's a neat idea, and complex regular expressions tend to be hard to parse for humans.

[0]: https://github.com/sulami/rx

Sure thing, I'll have a look at it tomorrow. Would you be so kind and share your browser/OS so I can try to reproduce?

> a good terminal mode so I don't need to switch to another terminal, the builtin terminals either lag badly after 10K lines of input or render/behave strangely when used with ncurses programs

I found vterm[1] to be pretty reliable for any kind of "advanced" terminal use. Also much faster when dealing with large amounts of output.

[1]: https://github.com/akermu/emacs-libvterm

I'm on the backend, but all I've seen of the new frontend is TypeScript. There has been a decision to switch languages for reasons that someone else can probably explain better than I could, a lot of us lurk here.

CircleCI is mostly Clojure, and some Go.

The only problems we are seeing is the significant startup time cost (hence the Go bits, where it matters).

Memory usage also isn't always great, but that's more a JVM problem, as it's not releasing reclaimed memory as quickly as you'd like it to. Fine on servers, annoying on a 16GB laptop.

The Common Lisp problems you outlined don't really apply to Clojure, it's a modern lanaguage and has been stable and versatile for us.

I wouldn't call it good, but I'm using Pandoc to generate LaTeX. The whole thing is wrapped in Hakyll, which already uses Pandoc to generate the HTML from the source (org/markdown).

Feel free to dig through the source if you want to see how it works: https://github.com/sulami/sulami.github.io

I’m successfully building sulami.github.io from org-mode using Hakyll, though my setup is quite complex due to my requirements (e.g. PDF output).

I'm planning to read PAIP over the upcoming weekend as a follow-up, I've already got it sitting here.

I know next to nothing about AI, beyond the basics of RNNs, so I'm hoping to learn a lot.

This is the tufte-css, with some modifications.

Raw css is here: https://github.com/edwardtufte/tufte-css

My additions on top (and all other source) here: https://github.com/sulami/sulami.github.io/tree/develop/css

I enable orgtbl-mode automatically in markdown-mode for that reason. It's the table bit as a minor-mode that you can use in any major mode.

I’ll be that one person:

org-mode already does tables with formulas and all[1], if you’re interested in text-based spreadsheets

[1]: https://orgmode.org/manual/Tables.html

I’ve recently built a thing at work to generate an OpenAPI spec from Clojure specs and some annotations, which worked out surprisingly well. We already have a collection of endpoints through compojure, and we just spec the handler functions and add some annotations for docstrings, and that’s basically it. As an added bonus we have the test suite validate all api calls against the spec, so in theory the OpenAPI spec should always be accurate.

For the curious, we took and hacked https://github.com/metosin/spec-tools to output OpenAPI 3, and added a very thin shim to hook that up to our routes.

Same here, also Clojure.

I don't know if I could ever going back to an imperative language full-time, I think I'd feel very hampered.

We’ve had this (or something similar) at Amsterdam Schiphol for a while now. With the scanners here, everything stays in the bag, electronics and liquids. The time savings are quite noticeable, because people are so slow at taking apart their baggage. Now they just throw on their bag, take off their coat, and maybe their boots, and they’re done, easily saving 1/3 of the time spent at other airport security lines.

Yes, but it's not required at work. Though I do a fair bit of OSS, too.

Not allowing people to call their custom docker tooling "docker-custom-tooling" on eg. Github seems like they're shooting themselves in the foot. Of course, it might be bad and damage their brand, but it could also enable a great ecosystem of surrounding tooling, some of which they could bless if they wanted to. Imho a poor PR move.

Emacs

That makes sense. I wonder how viable this is financially, given that it's a very niche browser, at least right now.

I've been wondering about this for a while. Especially now that they're actively supplying servers to support sync (which also at least at the time of writing doesn't seem to allow running your own server).

Even after having seen _a lot_ of (S)NES hackery, and being quite a bit into reverse engineering the consoles and games, this is mind blowing to me. The result is just so perfect. It makes me happy to see that decades after the release of these, there are still people playing with them, in one way or another.

My team jumped onto IRC, which is working pretty well. No custom emoji, but that's about it.

Location: Germany

Remote: Yes

Willing to relocate: Yes

Technologies: Python (Django/Flask), Haskell (Yesod/Scotty), Go, D, C, SQL, HTML/CSS (, Git, AWS, Linux)

Resume: https://sulami.github.io/raw/resume.pdf

Email: sulami@peerwire.org

Github: https://github.com/sulami

Blog: https://sulami.github.io

I'm a German dev (fluid in English) with about five years professional Experience developing Backends, REST APIs and a whole bunch of mixed tools and software to go along, also some quite low-level. I like complex systems and the challenges that come with scaling them, and I'm also sort of religious when it comes to code quality and productivity enhancements.

Location: Germany

Remote: Yes

Willing to relocate: Yes

Technologies: Python (Django/Flask), Haskell (Yesod/Scotty), Go, D, C, HTML/CSS, SQL, Racket/Scheme

Resume: https://sulami.github.io/raw/resume.pdf

Email: sulami@peerwire.org

Github: https://github.com/sulami

Blog: https://sulami.github.io

I'm a German dev (fluid in English) with about five years professional Experience developing Backends, REST APIs and a whole bunch of mixed tools and software to go along. I like complex systems and the challenges that come with scaling them, and I'm also sort of religious when it comes to code quality and productivity enhancements.

If you want less travel on your mechanical keyboard, you can use o-rings on the switches. Also dampens the sound when bottoming out. I personally use one per switch, but you can also go for two rings per switch which gives you laptop-like travel.

> Unit tests are momentum against change.

I get the point, but feel different. When a project becomes large enough, changes can result in breakage in the weirdest places. Proper testing can help discovering this breakage before shipping. For me, tests are a safety net for changes in the tested code.

Large projects without any tests are effectively unmaintainable, often times even for the original author, and most certainly for others.