domain:parrhesia.substack.com
That being said I mean... are there ways to legislate outside of tariffs to prevent this sort of major sell off of strategic business to adversarial nations?
Couple of tools. Golden shares is one way. Second is tax code. If you only allow some parts of the money spent outside the US to be tax deductible - suddenly US labor doesn't look so bad. The government has lots of tools. What is usually lacking is the will.
Large pizzas are usually a good deal for the same reason as 16" vs 12" sounds like only a third more pizza.
With most now-rich founder tech CEOs I could buy that. With CPUs, GPUs, OSs, social networks, whatever, there was a stable of plausible looking competitors and one caught a lead and rolled up into a progressively fatter cat that no other could compete with, network effects and all that, and maybe any one of them could have done just as well. ('anybody' seems a stretch, say, I'd most likely screw it up if asked to be a team lead of a team of any size. I might manage to manage a kitten but wouldn't bet on it. Very happy that tech companies have an IC track. And probably it takes a more select type to be a startup founder that doesn't fizzle out than, say, a Starbucks franchise boss or line manager wherever)
But SpaceX? Why would you expect some socioeconomic factors to turn up the same thing if there wasn't a idiosyncratic space maniac Elon driving it? There have been any numbers of attempts at space startups with comparatively incredibly lame results. Probably the most serious one has been Blue Origin where (if we believe the AI slop Google gives me for the search prompt) Bezos has likely poured in 100x as much of his own money as Elon did, and managed one (1) orbital flight so far, and some tens of 'hey we edged just over 100km so we can claim our tourists visited space' which tends to be peak space startup achievement. Is there any reason to think that swapping out Elon some random other boss wouldn't end up with at most a Virgin Galactic, instead of the wildly implausible looking outcome of first catching up with the established fat cat aerospace companies that had been doing this for decades at scale and made a giant government-funded grift of it, and then undercutting them on launch cost by 20x?
I've had luck with certain time-consuming rote tasks in medium-large codebases (1M - 10M LOC) like writing good tests for existing legacy code.
Here is some code.
Here are some examples of tests for other code which are well-structured and fit with the house style [style doc]. Note our conventions for how to invoke business logic. Note particularly that we do not mock injected dependencies in functional tests, other than the ones in [this short enumerated list].
- Identify the parts of the code I just handed you which look sketchiest.
- Write some functional tests for the code under test, mimicking as closely as possible the style and structure of the canonical examples of good tests
- Use this command to run your new test, iterating until the test passes.
- You can use this tool to identify which lines were tested - try to have passing tests that exercise as many lines of code as practical of the ones you identified as sketchy in step 1.
- Perform these linting and code quality assessment steps in order, redoing all previous steps on each change.
If at any point during this process, you identify a bug in the code you are writing a test for, describe the bug, propose a fix for the bug, and stop working.
It's not doing anything I couldn't have done, it's not even faster than me in terms of wall-clock time to get a good functional test, but I can kick it off in the background while I'm doing other things and come back to some tests that definitely pass and probably even test the stuff I want to test in something approximating the way I want to test it.
I come from a Unix background where we are taught that programs should do one thing and do it well. Seeing all this bullshit makes me seethe in a way very little else does.
More options
Context Copy link