Because not accepting it would break backwards compatibility and the general expectation that in HTTP the sender can use any non-obscure encoding and the receive has to deal with it. Of course there is no need for the receiver to implement special uppercasing for each encoding - more sensible would be to convert it to UTF-8 at the earliest opportunity.
really glad to see _this_ kind of analysis of these models. These type of analysis is much more practically useful than some context-less (especially with regard to time) number of "how good it is".
The point is he didn't behave rationally on one big really important point from a shareholder value standpoint. He didn't just sleep with someone and his arrangements in this regard are irrelevant
Viewing people like this is unhealthy. Nvidia could have failed instead of 3dfx, and that it didn't involved some luck, not just the magical clairvoyance of one individual.
And anyway, whether Nvidia created CUDA and deep learning are all moot at this point.
There's a nearly 20 year lead up to the current CUDA dominance including working with developers, but on the gaming side there's also a parallel with the dominance of nvidia's gaming tech versus how much ATi/AMD were able to get their tech to stick. Nv invested not just in capable hardware then throw it over the fence to leave developers to figure out, but software solutions easily integrated by others and engineer support to show off their hardware in the best light.
More directly on topic, I can't help thinking this is also a question of "what have you done for me lately" versus "what are you going to do for me tomorrow". It's hard to doubt he and the company as a whole have executed a vision and reaped the rewards, but have they been planning and doing the groundwork for another act.
> planning and doing the groundwork for another act.
A nontrivial part of which is pumping up an infrastructure bubble with NVIDIA money. It wouldn't shock me if that very aggressive and increasingly questioned strategy plays a role in whatever happens to Jensen.
The design of cuda is basically just brookGPU from stanford (supported by IBM, Sony, ATI, NVIDIA and DARPA), CUDA isn't so important beyond the fact that a hardware company can use it as a moat if they want to be anti-competitive. Google bought PeakStream, intel bought RapidMind.
The situation only exists because regulators in the 80s-20s believe that hardware should be a sanctioned monopoly. Intel, AMD, NVIDIA, Sun Microsystems, IBM, HP, and Microsoft were all involved. If it's not Jensen it'd be someone else, there is a good reason NVIDIA is called NVIDIA and not JENSEN. Jensen's leadership is as much resonsible for CUDA as G. W. Bush's.
reply