It’s not a false dichotomy at all. There is a finite set of resources that can be deployed at any given time. $1 spent on a laptop is $1 that isn’t going to medicine. This usually a curve rather than a straight line (so usually you’re better off with some combination of both), but this doesn’t really apply to a situation where your limiting factor is dollars rather than the factors of production.
You're building your arguments on a false premise, that ALL resources are distributed from some magical mutually exclusive abstract bucket. While it's true when applied to some slices of reality (be it on micro or macro levels) it isn't true in general, where it's parallelized, distributed, discrete and largely independent instead.
Doing it your way would inherently lead to socio-cultural diversity collapse.
Which is what's actually happening currently, not due to failure of distribution, but due to resources being siphoned and accumulated by hostile entities.
There actually isn't a finite set of resources when we're talking about human capital. Motivation and inspiration and incentive come into play, and as it happens different people are inspired to work on different things.
Linux is not monolithic, if you have a buggy component you can replace it. With the exception of the kernel of course.
I think we're entering the era of personal desktops.
Not all code written by humans is deterministic and reliable. And properly guard-railed LLM can check its output, you can even employ several, for higher consensus certainty. And we're just fuckin starting.
Unreliable code is incorrect thus undesirable. We limit the risk through review and understanding what we're doing which is not possible when delegating the code generation and review.
Checking output can be done by testing but test code in itself can be unreliable and testing in itself is no correctness guarantee.
The only way reliable code could be produced without human touching it would be using formal specifications, having the LLM write the formal proof at the same time as the code and using some software to validate the proof. The formal specification would have to be written using some kind of programming language, and then we're somewhat back to square one (but with maybe a new higher level language where you only define the specs formally rather than how you implement them).
reply