It's trust, because if the previous owners can't even bother cleaning it up, what else might their neglect have done to the house? Maybe I'll need to change the house wiring, verify if any new paint job is hiding mold or infiltration... All that is the cost of uncertainty and uncertainty grows with distrust.

This seems like a case where the tooling is compensating for unclear conventions. Collaboration tools are useful, but it may be more effective to strengthen the code specifications and documentation practices first, so intent lives in a predictable place instead of being scattered through implementation details.

But that kinda reminds me of the Chinese room thought experiment [1]. Is there anywhere in the latent space of the LLM a result that takes into account all these factors cited before? What kind of prompts or prompt orchestration would be necessary to achieve it? Would we still be able to tell the difference?

[1] https://en.wikipedia.org/wiki/Chinese_room

No more people doing crazy stuff with UX just because.

No more frontend dark magic only the most dedicated zealots can make sense of.

Democratised UI for all. And this might not even have had an AI assisted development, I'm just abusing this comment and going slightly off topic to say I'm happy for this tech, as someone that only dealt with backend and DevOps/SRE to be able to confidently create useful graphical applications for regular users and learning a ton about frontend architecture in the meantime.

Btw, a hall of fame for comments would also be interesting!

Currently I save all login tuples to Bitwarden and store OTP secrets onto Aegis. Could have been 1password and authy, it's irrelevant. The thing is, I only get pwned if both are compromised.

Now with ubiquitous passkeys in Bitwarden if someone has access to my vault unencrypted it's already endgame.

It's important to do a first pass cleaning up and preparing code before unleashing agents onto it, similar to how Working Effectively with Legacy Code advocates approaching consolidated codebases. After that the agent will actually pick up the new pattern and start propagating that to the rest of the code.