Company that makes chips wants to sell more chips, shocking.
Almost as shocking as “company that makes AI processors says AI is next big thing”, or “Arms dealer says war isn’t that bad”
Hey man, no need to get a chip on your shoulder.
I may not have a choice for much longer.
It doesn’t seem to make much sense otherwise that a sandbox for software needs to be dedicated hardware
Are they copying Microsoft’s “erm akshually, you need a TPM version 1.3 or higher chip to run our newest stuff” strategy?
Who watchdogs the watchdog?
Well they’ll have a chip for that in 2 years. It’s chips all the way down, you see.
Dead man’s snitch.
We have created a problem but don’t worry we also build the solution!
… can we maybe promote Huang to the new King of Comedy now that Rogan has been buried in an elephant’s graveyard?
Well, thing is, Nvidia doesn’t have to ask. They don’t need permission to ship it, and once it’s there anyway this system has a much higher chance of being adopted.
This is basically what they’ve done with past standards, like certain compute functions. They ship it, and then people end up using it.
A simple example is low precision data formats. The machine learning world went from largely working with 32 bit floating point to certain specifications of FP16, then BF16/tensor int8, then FP8, then FP4.
Why?
Because that’s what Nvidia shipped! Alternatives were proposed and entertained, yet they faded away because they didn’t have native hardware support. To be specific, FP16 was Pascal (GTX 1000 cards), int8 was sort of Turing/Volta (RTX 2000), BF16 was Ampere (RTX 3000), FP8 was Ada (RTX 4000), FP4 was Blackwell (RTX 5000).
