They used Jim Keller as
conduit for communicating future Zen performance (AMD Core Innovation Summit May 2014). They used his reputation to lend credence to their future performance claims. In hindsight what they were doing was even more obvious as Mark Papermaster behaved like a talk-show host, not the CTO of AMD. That should have been a dialogue between two senior AMD officials, not a simulated interview.
The information provided to us in that recording was instrumental in discussing future AMD processor performance. I'm surprised you don't remember it, since parts of the interview were very important in proving that AMD was actually developing a high performance core with Zen. There were a number of voices on the forum suggesting Zen had more in common with Bobcat cores, that AMD simply did not have the resources to design and produce a competitive high performance product.
So do yourself a favor and remove that
strike-through in my quote, AMD started what Intel continued.
[Later Edit] Here's
the entire presentation. AMD delivering their initial promise on x86 Zen performance kinda' made us forget they were actually promising a lot more with developing a high performance ARM product as well.