When I make a job offer, I always tell candidates what percentage of the company their equity offer amounts to. So many companies only say the number of shares and not the percentage, which is meaningless and deceptive.

I'm curious to see a postmortem on the procedure failures and/or testing holes that led to this outage. You'd think a flaw like this would be caught by a basic integration test.

July 20th, 1969 is a date that stands high in my life – it was the day that we, together as one people, achieved the greatest scientific and technological feat in human history. The specific role that Neil, Mike and I performed in the event was but one part of an immeasurably

I heavily prioritize immutability and referential transparency when programming because of the extreme benefits they provide, so it's interesting when mutability is exactly what's needed for a problem.

I suspect some of the job applications I'm receiving have cover letters written by AI. I can't know for sure, as humans are also capable of that level of inauthenticity and vagueness.

Rama has ~3,700 tests and ~22,000 assertions, about one test per 53 lines of source and one assertion per 9 lines of source. We have 18% more test code than source code. This doesn't include our automated chaos testing against real clusters, a separate 20,000 LOC project.

Simplicity is more effort in the short term, a lot less effort in the long term

Static typing is better for large codebases if your development team is mediocre or weak. Dynamic typing is far superior for a strong team. If you think static typing is better, then there are huge improvements you can make in how you structure your code and how you test.

Some general responses to the comments on my last tweet, which blew up way more than I expected: - Another way to phrase what I said: all else being equal, dynamically typed languages have a higher ceiling and lower floor than statically typed languages. Most of the negative

If 10x programmers exist, then most programmers are significantly below average.

I needed a name for a new concept in Rama for persisting the value of a large object to a file, but the names "snapshot" and "checkpoint" already refer to slightly different concepts in the codebase. So I called it "snappoint", which for some reason seems better than "checkshot".

It's a good sign when a user's questions shift from "How do I do X in Rama?" to "Here's code that's working fine, but I have a feeling Rama can do it more elegantly. Can it?" Their instinct is usually correct.

If a strong programmer uses "best practices", it's a coincidence

Your future self will curse your current self for not putting readily available information into log messages to debug unexpected issues

Strong testing can cover for less-than-perfect code, but great code can't cover for poor testing.