[iristenteije]

Yes, you're right that interpreters also allow users to run code. And we could argue that Apple is simply being inconsistent in how it applies its policies. I think realistically the difference is in how popular these newer vibe coding apps are, but also the fact that they have a much broader scope of what can be generated.

With Pythonista or a Lua-scripted game, the reviewer can assess what's possible: this app can do everything Python-with-this-API-surface can do, and nothing more.

With LLM-driven generation, the set of possible behaviors isn't fixed. The same Replit app can produce totally different behaviors next month than it can today, without ever being resubmitted, based on model or system prompt updates.

That's what I meant with "you can't review adaptive software".

[skeledrew]

I think the main "issue" is that Apple won't be able to properly take their beholden 30% or whatever their fee is nowadays. A "wrapper" app has no real value by itself so they can't put a serious cost on it, but once it's on a user's device it can be used to create unlimited things of high value. Apple definitely doesn't want to miss out on taxing that value in their ecosystem, so it's easier to just block it and force devs to stick to apps with a appraisable value.

[iristenteije]

That too for sure

[NetMageSCW]

The set of behaviors is certainly fixed or constrained with an LLM exactly the same way as an interpreter- if neither one has an escape from the pre-programmed APIs included in the app, they can’t escape regardless of what they do. Apple allows interpreters that run entered or downloaded code like programmable calculators and even more powerful environments. There’s nothing special about an LLM creating an app.