As a teenager, I tried to get addicted to cigarettes so I could stand by the cool kids and smoke with them. I started smoking 4-6 cigarettes a day but hated it so much, I couldn't continue after a week or two...
> Completely different results as LLM is non-deterministic.
You'd need to produce this like 20 times by each model and then do 2x20x20 cross comparisons by both models and ultimately distill the 2x20x20 comparison results into two reports of how they differ.
In this non deterministic computing future, everything else is voodoo, feelings and "vibes".
It's really bad when you let opus do investigations on broken java or infrastructure stuff. It starts decompiling .jar, sometimes multiple versions of the same dependency, reading every single kubernetes/terraform file and loading all the logs and info kubectl offers.
This actually makes a lot of sense, it's testing the luck of the candidate through the rng feeding the LLM. You wouldn't want to hire unlucky employees after all! Hiring managers of the past would solve this by throwing every second resume in the trash, now this is a built in feature of ATS.
My private next.js fullstack slop runs dockerized on my kubernetes cluster and for auth I use auth0, because I am too lazy to run keycloak or whatever dockerized auth slop is currently en vogue.