EE Times Current · 8 Nov 2019 · 37m
Kevin Crewell of Taurus Research reports on the Linley Conference, highlighting the dominance of AI accelerator startups and Intel's new Tremont core for low-power applications. Sally Ward Foxton interviews Groq CEO Jonathan Ross about the company's secretive 'software-defined hardware' approach and their recent high-profile conference no-show. Finally, Peter Hartwell, CTO of TDK InvenSense, discusses the future of MEMS, including ultrasonic sensing, optical sensors, and the vision of invisible, companion-like IoT devices.
early-stage companies and companies scaling up
Other shows on the same ground, closest first.
Host Daniel Bogdanoff previews Season 5 of Moore's Lobby, highlighting upcoming episodes with Steve Cozier of Skywater Technology, AMD's Sam Napstigger, and Intel's Wei Li. The trailer outlines discussions on the foundry business model, implementing power management ICs via microcontroller backdoors, and optimizing AI software performance on Intel platforms.
Dr. Wei Li, Intel's Vice President and General Manager for AI and Analytics, discusses the evolution of Intel from a PC hardware company to a software-plus-hardware leader in AI. He details the importance of hardware-software co-design, citing a 200x performance improvement through software optimization and the introduction of DL Boost and VNNI instructions in CPUs. Li also explains Intel's strategy of providing open-source AI reference kits to simplify development for complex workloads like chatbots and quality inspection.
Jonathan Ross, founder and CEO of Groq, discusses his journey from designing Google's first TPU to building Groq's deterministic AI chip, which achieves petaflop performance using older process nodes through unique architecture. He explains how Groq's hardware solves latency and throughput issues in AI inference by providing deterministic timing, enabling applications in autonomous vehicles, finance, and nuclear fusion. Ross also reflects on his time at Google X, the importance of empathy in innovation, and the cultural dynamics of building a high-performance computing company.
Austin Lyons of Chip Strat joins Ben Bajarin and Jay Goldberg to discuss the 'startup graveyard' for AI accelerator companies, noting 67 US startups are struggling against NVIDIA's dominance. The hosts argue that software moves too fast for dedicated ASICs to remain relevant, making flexible GPUs the short-term winner for both training and inference. They highlight Google's TPU success due to vertical integration, while questioning the viability of startups like Groq, Cerebras, and SambaNova that lack hyperscaler backing or control over their target workloads.
Ben Bajarin and Jay Goldberg analyze TSMC's strong earnings and strategic ramp-up of N3 production to meet AI demand, while discussing the compute deficit impacting firms like Anthropic and Cerebras' S1 chip. They explore the fragmenting market for custom ASICs as Google considers design partners like Marvell and MediaTek, and weigh in on Tim Cook's departure from Apple, arguing his legacy was preserving the company's culture for John Ternus.
The summary, chapters and topics are generated from the episode’s audio by this directory, not written by EE Times Current. The episode itself belongs to its publisher.