Try to say it again as it helps me to put things on one page to see it.

CH


"Reduced integration risk: With its sensor hub, QUIK has positioned itself on the
right side of Moore’s Law. Sensor hubs are emerging as a permanent element of
mobile/IoT architectures that should not only resist integration, but be able to
integrate other functions. 


snip on embedded AI.

specialized inference engines will appear as accelerator subsystems like video codecs and audio DSPs in an SoC,” he added.

XIlinx on inference


Once training is complete, the mode changes to “inference,” where the system applies its knowledge to situations in the real world. According to Xilinx, the big potential market in AI is inference,

Embedded inference will run in a inference engine, a heterogeneous core on a SoC, will that core be on FPGA /eFPGA...
the adjacent possible says that they will sure try it.... and IF it works the value of this IP


UP.

It has also been Dr Saxe adjacent possible since the fall of '15. BUT NOBODY saw the extent of the rise of AI/neural networks/inference/FPGA.

Is this factored in today?

Nada, QUIK has its plate full of important items and HAS NOT said anything. No hype, but it is out there.


For me one key take away is the power of reading and using S Johnsons lens of the adjacent possible as it sort of tells you way ahead of time

Open question the high performance articpro2...what is it aimed at...a competitor to XIlinx, ie data center, or do they extrapolate to high function on the device..say the embedded inference engines of tomorrow.

My vote is........