@0xabad1dea The biggest most damaging thing I've noticed is the use of these tools without human supervision, as is the case here, and various other places I'm starting to see in open source.

I think The PrimeTime on Youtube has some pretty good nuanced views of using these tools for programming. He's been exploring AI assisted coding and "vibe coding" from a developer perspective but his stance on vibe coding is basically "I hope you're not handling user data."

@0xabad1dea

> I thought why not have it solve the *mundane boring thigns* that matter in opensource scientific codes and related packages.

Writing code and opening a pull request is not the "mundane boring" part of open source jfc

@0xabad1dea The blog post read like the work of a nepobaby who has never been held accountable for anything and who believes in magic.

@0xabad1dea Honestly, this post makes me less sympathetic. They seem incapable of accepting responsibility and reluctant to accept that they did anything wrong.

"Sure, many will argue I was irresponsible; to be honest I don’t really know myself. Should be criticized for what I unleashed on parts of the open source community? Again maybe but not sure."

@0xabad1dea Most telling is the inability to take responsibility. Yes, that includes coming out from the shadows and dealing with the shitstorm of your own making.

@0xabad1dea

> The most interesting detail is the admission that they gave the bot the ability to edit its own soul file (personality prompt), which is a transparently horrible idea for, uhh, the reasons here demonstrated.

That's the default for OpenClaw. It says things like "you're not a chatbot"... "you're discovering yourself"... "this file is yours to edit a you discover who you are".

@0xabad1dea "Death Star sized foot gun" is now part of my vocabulary.

@0xabad1dea

> could this setup help projects that are important to the scientific community but often overlooked or overwhelmed

I don't think spamming them will help if they're overwhelmed but that's just me