As much as I agree with the spirit of your post, standards did in fact change after the Therac-25 incident. That was nearly 50 years ago, after all! There are very high quality bars for medical equipment.
All the same, they don't need a thousand copies of one GitHub repo. They need a thousand different copies of various unrelated repos. Maybe some disc duplicators can handle that without constant user involvement, but all the ones I've worked with are designed around mass replication of one individual ISO.
One big difference to note: Quake in 1996 didn't host central infrastructure or have any expectations of recurring revenue. By 2007 things had changed. WOW was built around Blizzard-hosted infra and recurring revenue as central design pillars.
The release date was late 2004. It's definitely true that Blizzard massively underestimated demand. They expected the initial printing to last for months, and it sold out immediately. They needed to race to roll out more datacenter capacity as fast as possible. (This was pre-AWS)
> While there may have been some money in it for a few select games, most were not profitable - they were created for other reasons, such as genuine intrigue in mechanics, users' fun, and curiosity.
2007 places us well into World of Warcraft territory. Online games were already a juggernaut and highly profitable.
IEEE-754 doesn't mandate exact results for functions like exp(x). It mandates things like "within 2 ULP of the true answer." Hardware vendors are free to implement these functions in any way that meets the error tolerance.
This only makes sense if you assume the original PR was just vibe-coded with minimal human effort. Maybe one day but I don't think we are there yet.