I think the idea behind a movement like this is good. This implementation worries me. When the plan is to have the AI do the “heavy lifting” from the get-go, this smells a lot like vibes. Even if we look past the ethics, this seems risky for bedrock systems like IAM. And even then, the master’s tools will never dismantle the master’s house. I wish them luck.

I wonder what these devs think of OpenStack

There is no clean room implementation of a discrete title that would not run afoul of Western copyright and trademark law. If you want to release Super Mario 64, as it was made, you will not have a free license to do so until it enters the public domain. You can release Supra Wario 63, with all it’s not quite the same assets, and argue that you can get away with it. The value of these projects is not the licensing.

Now, worse? This scenario would actually be improved if the AI were to do the whole implementation? Because it would apply a FOSS license where a human could not? This is nonsensical.

while* hearing about it.

Think we just have a different take on the meaning of ‘expect’ is all.

nothing about finally being in a playable state

“… appears to work on Linux with Proton 11 as captured in my screenshot above”

copy–pasted

heh

I know it’s five days old and made heavy use of AI. My comment is that before hearing about it, it’s deeply cynical to expect it to be AI driven.

As expected? There’s been a few high profile RE efforts recently of games of the era without AI tooling. The Mario 64 decompile effort from a few years ago being the most notable. While I’m not surprised to see an AI disclosure here (I would assume agents are a very useful tool for this particular work), this just as readily have been an entirely human passion project.