The Future of AI Hardware with James Wang
On today’s episode we welcome back our very own James Wang to talk about the future of AI hardware and his recent talk at the CogX event. With a definite move back towards fabrication and production of chips in Silicon Valley, the future of technology in both America and abroad is in just as much transition as ever. James informs us on what the industry and related fields look like right now, how this will all translate into the marketplace and the effect that it will have on companies and customers. We discuss the last year in chip fabrication and think about the years ahead; what vertical and horizontal integration can we expect? We talk about how the field is placed geographically and China’s commitment to industry and market leadership by 2025. We also get into the implications for the autonomous driving market as takes shape. Key Points From This Episode: Comparing this year’s CogX event to the one last year The current startup costs and landscape for chip fabrication Competition, opportunity, training and inference in the field of ai hardware How to think about the market size and the pricing implications AI training as a continuous process as data sets grow over time How it translates in the autonomous driving market Breaking down the geographic placement of the startups in question Horizontal and vertical players in the field and how this differs to the PC era The three main conclusions; the rise of China, vertical integration and optical computing Looking forward to the possibility of quantum computing Tweetables: “What is really interesting is that semiconductors are the new black again. Venture capitalist have once again poured money back into chip companies.” — @jwangARK [0:02:07] “For companies that are truly operating on this kind of scale, training is a continuous process because they’re trying to improve the accuracy of the networks.” — @jwangARK [0:08:20]