“My formative memory of Python was when the Quake Live team used it for the back end work, and we wound up having serious performance problems with a few million users. My bias is that a lot (not all!) of complex “scalable” systems can be done with a simple, single C++ server.”
They aren't because we're not committing prompts. The analogous would be compiling to and then committing unmaintainable assembly when the higher level language, which is a deterministic compiler, already exists
I mean, iteration and interaction builds your understanding which verifies and validates what you built. Relying on agents without formal validation is like saying a tree fell in the forest
Sometimes you just hit impossible life scenarios like a pretax clawback in a hostile workplace which drains your liquid savings and leaves you broke, or breakup with a partner which forces you out the home, or learning that the landlord you split your duplex with is a charged criminal of some pretty awful crimes.
You can’t fail sometimes, even if you thought it were failure
This is what Anthropic did with agents and $20k to write a C compiler that survived gcc’s torture suite. But the LLM knew:
1. What a C compiler was
2. What a C compiler looked like
3. What the C compiler had to do at runtime to pass gcc’s torture suite through some sort of collaborative iteration (compile, run, did it get stuck at some torture suite test or fail?)
Remove 1 and 2, or replace it with imperfect business logic, and you’re left with a system that is built to _only_ pass the tests you supply it, or in the most extreme case, print(“unit and functional tests pass!”)
Are you okay? They were teenagers. We were all there and we all tried. The logistics of having 7 billion people quarantine while international flights and asymptomatic carriers carried on made it _impossible_ to not play out like it did
Hard numbers, no. Even high level concepts and theory you need to triangulate and prompt in different angles, across different models, and figure out what overlaps to build a mental mode that’s - even then - roughly 80% correct. It’s better than google, but the information isn’t free