For the developers and third-party publishers, it's a really big deal to have a standard of some kind. It means that they don't have to spend their time making

For the developers and third-party publishers, it's a really big deal to have a standard of some kind. It means that they don't have to spend their time making the market: the console makers do that, and then they just sell into it. There is always grumbling about the various overheads of consoles, but as long as it results in unique experiences that people pay for, this state of affairs probably isn't going away either. And the differentiation of special case acceleration like what we're seeing on these architectures is less exciting on its surface than strapping a high nominal spec to some wild programming model, but it's also a more sane way of achieving next-gen performance. Cerny's PS5 talk put it this way: The PS1's "time to triangle" (i.e. the basics of the asset pipeline are proven and real development can commence) took 1-2 months, the PS2 doubled that, the PS3 doubled that again, and everyone protested so much that they got it back down to 1-2 months for PS4, and with PS5 they are aiming for under a month, which is actually really ambitious for any development platform! Also, I sense that the special-case approach also has neat goodies in store down the line because console devs always end up abusing hardware features to get results beyond the original requirements. There's stuff in the PS5 arch intended for DSP code, for raytracing, for decompression, and probably other things too, but most of these things are a little bit more open than just being an accelerated single-task function. I haven't seen what they're doing on the XBox end yet but I wouldn't be surprised if it has some good details to geek out over too. These functions are uniformly win-wins since they give an immediate result that accelerates something common, and then we find out what else they're good for later. Sometimes they aren't, but it's common that they surprise.