Do Not Outsource Judgement

Reminds me of a blog post from a few days ago by Melanie Mitchell on a talk she gave about the recently-passed Brian Cantwell Smith:

Judgment, he wrote, is “a form of dispassionate deliberative thought, grounded in ethical commitment and responsible action, appropriate to the situation in which it is deployed.”

On the other hand, there is reckoning: the “calculative prowess” at which AI systems already excel. AI reckoning has lead to gold medals in mathematics competitions, to generating complex code, to predicting protein structure, and even to carrying out fluent conversations. But reckoning without judgment is a dangerous thing...

https://aiguide.substack.com/p/on-brian-cantwell-smith-and-t...

15k vibe coded PR? I'd just ignore and not engage. Hopefully that person gets the hint.

Also lead developer isn't the only "adult in the room", the room should be full of adults. That needs fixing (inc. up to firing people) or maybe get a new job if that is not possible.

> Risk is being pushed upward instead of owned locally

I don't view this as a bad thing or at least not in many contexts, in many orgs, risk is actively pushed downward. Vibe coded PR slop is poor hiring/culture blowback.

It is an amusing inversion of the situation in which higher level decisions by seniors/exec are made that then have to be implemented by lower levels in the org; with those lower levels having to smooth and mitigate the risks and negative consequences of aforementioned higher level decisions.

I'm seeing a worrying pattern in this space. Some engineers I work with treat LLMs like unquestionable machine gods. I've rejected PRs which are quite frankly trash piles and have received responses which are clearly LLM generated defending the PR using reasoning at the level of an elderly rat.

The problem is that a not insignificant portion of engineers out there in the real world don't care if something is good or not as long as someone signs it off. And another not insignificant portion of engineers out there really can't tell if something is good or not. They take no responsibility for it - it's just seen as a way of abstracting themselves from the responsibility. I can't fathom how their reasoning works there either.

As always, human problems. LLMs make the worst of us faster and more demanding at being terrible. The are definitely a net loss in a large org or a Dunning Kruger inflicted team.