dllthomas

> But then, at that point, you're saying something pretty profound about computer science (roughly: that there was no meaningfully more sophisticated and complex place to go from where we are today).

Or that human taste and expertise cannot helpfully guide a project there.

> that's what "productivity", in the economic sense, means

That's too strong. It does mean we can do the same work with fewer people, but depending on the shape of the demand curve that can mean more jobs exist doing that thing than otherwise would (an instance of the Jevons Paradox).

I mean, in practice Rc is the one I run into the least, but pin only comes up occasionally, yes. Maybe once or twice a year on one of my projects.

Of those, Pin, by some margin.

Neat, but I'm pretty happy with my disintegrated development environment.

Yours is a response to someone providing clarifying information, something I often find helpful on HN.

Sometimes I sing under lights that are purple. Sometimes I'm shirtless and I tilt my head.

> [Performance cost] is (de facto) required for dynamically typed languages, but not for statically typed ones where the newtype constructor/deconstructor can be elided at compile time.

For a single value, it should reliably be free in any reasonable statically typed language that meet your other criteria.

For a collection, it may still be de facto required. Unwrapping a set of addresses into a set of strings takes unnecessary cycles, an unsafe coercion, sufficiently sophisticated affordances around coercion that it can be safe, or a smart enough optimiser. At least some static languages have at least one of these; I'm not sure all do, and certainly not all have always had.

Nitpicking, but this has bugged me for a while and I'm taking this opportunity to vent:

"Enslaved people were counted as three-fifths of a person, not as full citizens."

They were not counted as three fifths of a person in a way that matters for what we talk about today. They got zero percent of the vote they deserved. They probably got, on average, quite substantially less than three fifths of the respect and dignity they deserved. They were counted as three fifths of a person for how much they magnified the power of the votes of their captors and how much taxes their State had to pay. Slavery would have been every bit as wrong if they were counted as whole persons (or half persons, non-persons, double persons) for apportionment and taxation.

I think I'd say that the batch file is not itself autonomous but the system as a whole is autonomous (if limited) but I'm not prepared to argue that's the correct definition.

If you learned that a piece of the brain where meaningful computation takes place was stateless, would that cause you to question whether the human mind was "highly autonomous"?

You can avoid specifying agency in the active with some sort of placeholder. Hopefully, maybe, that placeholder is going to be more noticable than the omission of agency in the passive... but it seems more useful to simply ask directly whether agency is clear.

"Missiles were shot at Gaza" is passive and avoids specifying agency. "Someone shot missiles at Gaza" is active and avoids specifying agency. "Missiles were fired at Gaza by Israel" is passive and specifies agency. Sometimes you don't even need a placeholder: "Missiles hit Gaza" is active and avoids agency.

Importantly, specifying reasoning can have communicative value while falling very far short of formal verification. Personally, I also try to include a cross reference to the things that could allow "this" to happen were they to change.

Wildly, the Polish word "nagle" (pronounced differently) means "suddenly" or "all at once", which is just astonishingly apropos for what I'm almost certain is pure coincidence.

> so it's no surprise that we're not Boltzmann Brains

I think I agree you've excluded them from the definition, but I don't see why that has an impact on likelihood.