Add a draft for The Black Hole Problem

astro.config.mjs

@@ -22,6 +22,8 @@ export default defineConfig({
         mdx(),
         sitemap({
             filter: (page) =>
+                !page.includes("blog/the-black-hole-problem") &&
+                !page.includes("blog/beat-the-drum") &&
                 !page.includes("blog/the-power-of-the-link") &&
                 !page.includes("blog/software-engineering-manager-tactics") &&
                 !page.includes("blog/building-a-hyper-key-tree") &&

src/pages/blog/the-black-hole-problem.md

@@ -0,0 +1,77 @@
+---
+layout: ../../layouts/BlogPostLayout.astro
+categories:
+  - communication
+date: "2024-09-27"
+unlisted: true
+title: The Black Hole Problem
+---
+
+A [quote](https://scrapsfromtheloft.com/movies/interstellar-2014-transcript/)
+from [Interstellar](https://en.wikipedia.org/wiki/Interstellar_(film)) talking
+about a [black hole](https://en.wikipedia.org/wiki/Black_hole):
+
+> ROMILLY: Nothing escapes that horizon.
+>
+> Not even light.
+>
+> Oh, the answer’s there, just no way to see it.
+
+In companies, one problem that I've occasionally run into is what I've come to
+call the black hole problem. It means making a request of some kind and feeling
+like it has gone into a black hole. You're not sure who has seen the request.
+You're not sure when you'll get a response. You don't even know *if* you'll get
+a response. This could take many forms, like asking a question to a person or
+group or putting in a ticket for something.
+
+The black hole problem is one way to gauge how dysfunctional a company is. No
+company is perfect, but the more it feels like my requests are going into a
+black hole, the less likely it is that the company's internal processes and
+systems are designed or operating well in general.
+
+It reminds me of building user interfaces, which have varying degrees of
+responsiveness to their users. A bug might cause some requests to fail. Doing
+something else might take so long that some users give up on it and leave before
+it finishes. A third thing might silently succeed, with no feedback to the user
+that something did in fact happen. It takes effort to build user interfaces that
+feel fast, reliable, and predictable.
+
+Similarly, companies should actively think about how to make its systems,
+processes, and people more responsive. Whenever someone makes a request, it's
+going into some sort of queue. That queue could be implicit or explicit.
+
+If you ask someone a question, you are effectively giving them a piece of work.
+You expect a response, but you probably don't know all the other work that the
+other person has. That person has an implicit queue in their head of everything
+that they need to do, and you don't know what position your question has in that
+queue.
+
+Or maybe you put in a ticket to IT for help with something. Your ticket is
+explicitly going into whatever queue your IT department has set up to triage and
+deal with incoming requests.
+
+When I run into the black hole problem, I sometimes wish that we would make
+these implicit and explicit queues more like the queues that we use in software
+systems. I'd like to have
+[SLAs](https://en.wikipedia.org/wiki/Service-level_agreement). I'd like to be
+able to see metrics like how backed up a queue is and what is its typical
+throughput. I'd like to have monitoring and alerting so that when something
+isn't working well, I can easily find out about it and hopefully improve it.
+
+Of course, that all takes work and introduces overhead that wouldn't always be
+worth it. At one end of the spectrum, every queue in a company would be totally
+transparent. I'd be able to see every little thing that someone or something has
+to respond to. I would know where all my requests sit in those queues, and I
+would have confidence in when those requests would be fulfilled (or at least
+when I'd get pushback). And when a queue is failing, there would be
+accountability of some sort.
+
+On the other end of the spectrum, you have the black hole problem everywhere.
+For any given company, there's a point on that spectrum that is optimal. But by
+default, companies tend to gravitate towards the black hole end, since it takes
+no work or effort to make things opaque.
+
+In the meantime, the most effective strategy I have for dealing with the black
+hole problem is to keep track of my outbound requests. I add following up on
+things to my own todo list so that if something does go into a black hole, I can
+proactively deal with it.