But there's a world of difference about lying about capability today and working with customers to deliver what they need.
My favourite example is how ARM ended up in Nokia phones. They had a 32 bit RISC CPU core which met the performance requirements but Nokia turned them down due to the memory density of the instructions meaning it required more expensive storage than their existing solutions. On the way back from the meeting they came up with the idea of Thumb, worked out how to do it, did it, and Nokia bought it. It then took over the industry.
That conversation in SV today would involve Nokia going "We love what you folks are doing" but just keeping the existing system without ever saying why, and nothing would improve.
Interesting to think about that. I remember the SuperH had 16 bit instructions (well, SH-5 was variable length,) I wonder whether that would have become 'our future' if ARM hadn't done Thumb back then.
IIRC ARM did end up licensing the SuperH patents about that. I tend to think SuperH could have been much more of a contender had Hitachi been more willing to license it for integration by others.
The one that mystified me was why anyone thought MIPS was viable in this space, other than the fact that it existed. Ultimately MIPS utterly failed to take advantage of their very big wins.
My favourite example is how ARM ended up in Nokia phones. They had a 32 bit RISC CPU core which met the performance requirements but Nokia turned them down due to the memory density of the instructions meaning it required more expensive storage than their existing solutions. On the way back from the meeting they came up with the idea of Thumb, worked out how to do it, did it, and Nokia bought it. It then took over the industry.
That conversation in SV today would involve Nokia going "We love what you folks are doing" but just keeping the existing system without ever saying why, and nothing would improve.