I don't ever recall intel saying that when they had the lead five-six years ago. Intel (and pretty much everybody else) has always said that. AMD is currently saying that about neural nets, AI, and machine learning. Pathfinding is expensive.
Intel is doing just fine with Surface and yoga competing with tablets. The original tablet market (galaxy and ipad) is shrinking. OTOH, phones were more of a marketing or management decision not to build an ARM product. It wasn't a technical R&D issue. I think this points directly at what Swan was getting at. The concern over protecting X86 kept them from building a competitive ARM based cell phone applications processor. The direct result of this was TSMC getting as big as Intel in fabrication capacity. I think they are trying to avoid this same mistake in the accelerator market. Alan
Market Data 
Markets 
AI
