I am guessing that means he already has the preliminary Zen 3 performance benchmarks. ( Partners would have it under NDAs but these things leaks back and forth within the industry )
And Zen 3 would actually be something Intel cant / dont want to benchmark against.
Considering 10nm won't be any better than 14nm in pushing Clockspeed ( They used that one up too much with 14nm ), and their IPC are known, ( Willow Cove, ) I am guessing Intel has used up all their tricks until Alderlake in 2022.
Meanwhile Zen 3 should give at least 5% higher Clockspeed and 15% higher IPC. All while maintaining MultiCore dominance.
Would not be surprised If Intel's only answer were to lower Desktop price by 30 - 50%.
Much more interested in the Mobile side of things.