@runevision I think it’s very difficult to define a clean boundary between “non slop” and “slop” code when genAI is used...

@runevision I think it’s very difficult to define a clean boundary between “non slop” and “slop” code when genAI is used at all, and so changing our ethical standard from “No AI” to “No *agentic* AI” is a dangerous choice, because it opens up potentially gaping loopholes for devs to claim non-AI use that wouldn’t faithfully reflect reality. This is probably why you’ll see hard-line pushback.FWIW, I think House House’s specific example is one of the most innocuous uses of AI coding, and it seems to be a rare exception rather than the rule in Big Walk’s development. But how do we generalize that specific case to a rule that would cede that *some* LLM use is acceptable, without introducing moral hazard?A thought experiment: the widespread use of Claude Code/Codex/Gemini wasn’t widespread until maybe early last year; does that mean no meaningful amount of slop code existed before that? Would we be tempted move the goalposts again when a new system of extruding software is invented next yea...