Oh well. People said that every year. However, if you take a look at AMD's book. Even if they continue to lose money at this rate, they can last till early 2019.
He is the lead of the HSA spec. HSA goes v.1.0 FINAL early this year and with APP SDK v.3.0 & Carrizo (first HSA 1.0 APU) out, it is very likely that he feels that he sees no more challenges in AMD. Just like Jim Keller left AMD when Zen's design has been finalized.
When I say, finish. I dont mean now, I mean it has no future. And I can tell you for sure the software for HSA is NO where near finish.
Unable to keep talents is a important sign for company. Are there no more innovation or challenging position within the company?
With the Split of Radeon Technology Group, it is looking more likely they are preparing for being acquired. Broad dont wait until 2019 when they have nothing more left to act. They are in debt, and with no sign / future of clear turn over. The market for Client Computing ( That is Intel's term for Desktop and Laptop ) has been in decline with no sign of improvement. By the time Zen comes out in 2017, capturing even 10% of a declining market is meaningless. It could stay on for longer, but it doesn't change course for the company.
On Server Side AMD is lacking everything Intel is good at, Network Controller, Very Good ECC Memory Controller and I/O. ARM Servers Chips are delayed again, with competition from Qualcomm, ( Not to mention not many want a ARM server chip, at least not yet, not now. )
Radeon has been rapidly losing market share to Nvidia.
AMD, as far as product or IPs goes, doesn't even have a single thing that is leading the market or worth of any value. ( Besides x64 )
Not having a product that is good right now may not matter, at least there is hope with the talents they have their may be more things to come. But with all these talents leaving, ( and especially to their competitors ), it is likely AMD have its own deeper problem that no body knows from outside.