Intel's A.I. Director Singer Lays Out the Vision for Deep Learning

Singer laid out for me the thinking at Intel about A.I. in advance of the company's first-ever dedicated A.I. conference, which happens next month. With A.I., Intel faces a challenge: the dominant player by conventional wisdom is Nvidia (NVDA), whose graphics chips, or "GPUs," have been widely used by Alphabet's (GOOGL) Google and Facebook (FB) and all the other cloud companies to do much of the heavy lifting for machine learning and deep learning. In addition, Silicon Valley has cultivating numerous startups focused on dedicated A.I. chips. Advanced Micro Devices (AMD), Intel's rival in PC chips, is also pushing the use of GPUs for A.I., being the only other vendor of size in GPUs besides Nvidia. Going up against that, Intel has Xeon, and also field-programmable gate arrays, acquired with the purchase of Altera, and also more specialized A.I. chips of its own from the acquisition a couple years back of Nervana.