Looking through these text messages, if you take a look at the final one Elon texts about the informed productivity of twitter engineers. I'm wondering if Elon Musk is misinformed or what's going on; I'm thinking he is referring to lines that actually go into production or something of that nature which I imagine would go through intensive review and so on?
Because 100 lines of code a month, not to mention that LoC is a poor metric of performance anyway, still seems like very little work.
What do you guys think? Is this an accurate measure for twitter engineer's productivity?
I'm more interested in the engineers who are removing 100 LOC a week and making the system simpler and more reliable.
Musk is definitely not one you should listen to for advice about software engineering. Or rather, his mental model of software engineering is very big on "a small number of geniuses solving impossible problems to make a breakthrough" which is only a tiny fraction of the total effort in the software engineering space.
I love how billions are being thrown around ideas like let's do a blockchain Twitter, whatever the hell that is, and Musk is like yeah that's not a real thing, and the other guy is like whatever, my client Sam Bankman-Fried is good either way to sling that 5 billion. Money is apparently not an issue when you have an offshore cryptocurrency exchange totally not used for financing crimes around the world, and also Biden it looks like. In the United States a political party accepting millions from a scumbag crypto lord is not a problem, they are totally able to serve the interest of the people instead of their own interest and the interest of this guy.
Lines don't mean anything. In Haskell a few lines accomplishes a few hundred lines in other more verbose languages. If you use a library, one method might be thousands of lines. Not to mention, you could use very long variable names or have lines wrap in less characters. This is a really bad metric.
6 comments
[ 2.9 ms ] story [ 38.9 ms ] threadBecause 100 lines of code a month, not to mention that LoC is a poor metric of performance anyway, still seems like very little work.
What do you guys think? Is this an accurate measure for twitter engineer's productivity?
Other than that, I agree on a bad day I might pump out only 100 lines, but on a good day it can be 1,000-2,000 lines.
LoC isn't really a good metric by itself, because as every great hacker knows:
Hours of planning can save days or even weeks of programming.
Understanding the problem domain is vertically essential.
Musk is definitely not one you should listen to for advice about software engineering. Or rather, his mental model of software engineering is very big on "a small number of geniuses solving impossible problems to make a breakthrough" which is only a tiny fraction of the total effort in the software engineering space.
The creator states - "They were interpreted by AI. We've aimed to be as accurate as possible, but we can't guarantee accuracy or validity."