The shitty thing is this was already a problem before AI development. This just made it 100x worse.

Not surprising, given how Sakurai has been known to involve himself in his projects. He's not the kind to delegate part of his work to the plagiarism machine.

That's not an AI issue, that's a developer issue (and one that's existed for a long time). I remember all the shitty C&C clones back in the day that just copy pasted design elements.

I think AI has made that process easier but ultimately it's "shitty devs making junk" and not "AI making junk". Even if design is being delegated to AI, it is still a human delegating that authority.

I agree with this except there is no way to use AI that doesn't push a design towards being samey junk, less AI is better but this is not as much a tool with a right way and a wrong way to use it (it is that to some extent of course) and more primarily a lie being sold to people that spits out things people then believe they or the AI made when it is simply an older human idea that was stolen and thrown into a woodchipper without any consideration, material support or recognition of the human who provided the training data that was crucial.

If you want to file off the serial numbers of other human ideas and mutilate them in an LLM type device fine but don't claim you or the AI are actually contributing anything real.

GuNs dOnT kiLL PeOpLe

bLiStErInG tAkEdOwN!!!

Insane headline, considering that this was just about the calmest, most level-headed take on AI's use in game development I've seen from any studio head. Which is on-brand for Sakurai; the most radical opinion he's ever put forth on game design is that Bayonetta having free ZTD combos out of the box was acceptable.

I disagree that everything AI generates is garbage, but I do agree with the "contributing anything real" comment. That might seem like it is at odds with each other, but I have a one word reason: perl.

Back in the day, I worked extensively in perl and really did believe in TIMTOWDI. The problem with TIMTOWDI is TIMTOWDI, though. Not every system I built needed to be some unique butterfly. I just needed it to function (preferably with high efficiency). Then I started working in group projects. Perl is awful outside of single-developer environment. Imagine 5 people all declaring their own best practices and refusing to compromise as it would "violate best practices" that they had just set for themselves.

There is value in NOT creating new ways to do things simply because you can. This becomes clear the first time someone approaches you about building our own security library for something. You have no idea the amount of code copied from stackoverflow that exists in production unattributed. It simply works and no one is going to care (usually not even the people posting these solutions - which includes myself).

Even outside of that, if I sat and wrote a sub() to do some recursive file handling, I don't own that code like some intellectual property regardless how cleverly I wrote it. I combined several documented parts into a single idea. If I don't own those parts, I don't see how I can own the idea that was built from them.

I’m glad someone else called out “blistering takedown” lol