His point is, Intel will keep competing just fine regardless of the silly ‘moar cores’ narrative AMD has successfully implanted into the minds of a few layers from the community. Quick Sync — with its great performance, documentation, and ISV support — is more relevant than 8 extra cores for many productivity workloads, and the same goes for a good dGPU with a proper compute/general acceleration SW ecosystem (that is, NVIDIA and soon Intel with DG* + oneAPI). Meanwhile, AMD is “competing” with flaky and broken-in-every-other-release OpenCL 2.x (Blender was unusable in the last few months), poorly documented/performing/supported VCE/AMF, and dropped Polaris support in the latest ROCm — with Navi* support still being an incomplete mess.
Guess AMD gave up on competing against Intel and NVIDIA in the productivity front (outside of Cinebench and those compilation benchmarks we devs don’t really run like that because build caches and partial compilation), huh?