@LanceJZ Right. Thousands out of how many?

There are _plenty_ of video games which are just gambling.
There were _plenty_ of video games which were just gambling _years ago_. And the situation has only been getting worse with genAI.

@LanceJZ @eldersea The comic is calling the video game _industry_ gambling, not video games themselves.

Although plenty of video games are now just gambling and they're getting away with it because they're merely not using standard playing cards or hiding it with a thin "Not Slot Machine" theming.

@TimWardCam @fribbledom Either the autocomplete is a custom non-LLM job, and therefore not AI, or it's some kind of LLM job in which case it can't be reliable enough to be trusted blindly.

If the autocomplete is LLM, especially AI, then you need to learn how the language works on your own just to make sure the output is properly correct and not just "it runs / compiles / looks" correct.

Even a custom made AI is too unreliable to be used for learning or to be trusted without double-checking.

@TimWardCam @fribbledom I don't know about "mad" but what you're describing sounds ill-advised.

If the language is obscure enough that there are no reliable resources you can find to learn about its for-loops, then it's very likely the AI won't produce correct for-loops either.

And that's assuming it manages to recognise the language in the first place, which is less likely the more obscure the language is.

And all that is before considering the inherent unreliability of AI.