In #Flancia we'll meet2d ago

"Should [[fediverse.party]] list apps whose devs use auto-generated code?" #305
https://codeberg.org/fediverse/fediparty/issues/305

I mean... why shouldn't it?

There is a "watchlist": https://codeberg.org/fediverse/fediparty/wiki/watchlist+-+fediverse+apps+created+using+generative+models.-

Slightly worrying, feels discriminatory (and no, not in a "good way".)

Should fediverse.party list apps whose devs use auto-generated code?

[Nolto](https://nolto.social/) is a project designed for professional networking on the fediverse. Its code was auto-generated using lovable.dev. This doesn't go against our listing criteria, but maybe it ought to? Until we have a robust discussion and come to consensus on this, I'm calling a mo...

Codeberg.org
Show threadKonrad Hinsen1d ago

@flancian I see the same kind of polarization in my professional environment. AI-yes-or-no is the most dividing issue I have seen in academia for 30 years.

One issue is the ethics aspect, which allows no easy compromise.

Another issue is workflow incompatibility. Collaboration between AI users and AI critics is de facto impossible. The rhythm of work is not the same, the subject of shared production is not the same (prompts vs. code).

Show threadIn #Flancia we'll meet
@khinsen yes, I know exactly what you mean, I'm seeing the same working in the industry. I was expecting a 'bipolar trap' like this to result from AI adoption, but I wasn't expecting it to happen this fast (that's another common pattern as of late...)
Apr 13 at 10:01amWeb
Show threadIn #Flancia we'll meet1d ago
@khinsen I think like with many such traps it can be avoided with extra effort and good intentions from each pole applied towards building a mutual understanding, maintaining interfaces and bridges between the groups, and in the process learning to work together. But at this point these resources don't seem super abundant, partly because some people are scared+tired and partly because it's always easy for groups to default to delegitimizing the outgroup (it seems like a bug introduced by our evolutionary path.)
Show threadKonrad Hinsen1d ago

@flancian I see the main problem in AI adoption being forced too rapidly. There are incompressible time scales of change in individuals and in social structures. You can't move faster then them without doing damage. Which is what is happening now.

One example is the copyright debate. There is no way we can productively cooperate across the AI divide before this is settled. But it can't be settled in less than a decade.