On 13/08/26 at 10:44 -0600, Gunnar Wolf wrote: > If I ask an LLM to package something for me, I should take a good, > critical dive into whatever it is proposing; lets say at the very > least with as much detail as a package I'm being asked to sponsor (or > more than that, as I cannot say the LLM has the “intent” of behaving > well and doing things right and up to the latest standards).
This notion of "intent" echoes the notion of "essential complexity" from Fred Brooks' No Silver Bullet essay (1984), which resurfaced recently on numerous occasions because it's extremely relevant to AI-assisted software engineering. A very short summary (but I really recommend reading more on this) is: Brooks differenciates between (1) accidental complexity (syntax, tooling, scaffolding code, ...) (2) essential complexity (what to build, domain logic, edge cases, ...) AI assistance massively reduces accidental complexity, but does not really help with essential complexity. With AI, you can be blazingly fast at writing syntactically correct software that solves the wrong problem. Lucas
