zlacker

>>fancy_+(OP)
I see lots of doomers in here talking about Intel failing. Well, the truth is the government will just bail them out again.

But instead let's discuss some turnaround suggestions, starting with:

   (1) Acquiring at least one or two AI chip technology startups. (Lightmatter, Cerebrus, etc)

   (2) Assembling an AI hardware research team focused on efficient AI + software compatibility with existing AI accelerator libraries.

   (3) Investing heavily in the PyTorch team at Intel.

(1) is the most important and urgent action that Intel can do to sort out their issues.

>>rvz+Ua
> (1) Acquiring at least one or two AI chip technology startups. (Lightmatter, Cerebrus, etc)

Intel spent $2B on Habana and it went nowhere. Too bad, the hardware seemed promising enough.