Listen to this story
|
OpenAI has made a strategic move by appointing Richard Ho as the head of its hardware division. Ho’s arrival marks a significant shift in the company’s focus on hardware optimisation and co-design, targeting data center networks, racks, and buildings.
Ho’s appointment coincides with OpenAI’s expansion plans, as the company is actively seeking talent for a deep learning hardware/software co-design engineer position. The job listing outlines a focus on optimising hardware for workloads, evaluating accelerators, and influencing the roadmaps for data center networks, racks, and buildings in collaboration with partners.
OpenAI’s commitment to hardware optimisation has seen parallel developments with Microsoft’s Azure cloud. The company is part of a significant $13 billion investment primarily in cloud credits, with plans to contribute insights to Microsoft’s AI chip development. This relationship underscores OpenAI’s pivotal role in shaping the future of AI hardware.
Rumours have been circulating about OpenAI contemplating its proprietary chip hardware and exploring potential acquisitions. CEO Sam Altman has been pursuing fundraising for a separate chip company, hinting at a deeper exploration into hardware innovation within the organisation.
Before joining OpenAI, Ho spearheaded the chip engineering division at Lightmatter, a photonic computing company. However, his illustrious career spans over nine years at Google, where he held a prominent role in the development of Google’s Cloud Tensor Processing Units (TPU) as the senior director of engineering.
The TPU chip family at Google plays a pivotal role in training and inference for large language models within Google’s ecosystem and as a part of their cloud services. Ho’s tenure at Google also involved co-inventing innovative methods using machine learning for chip architecture design, showcasing his expertise in the field.
“I aim to build teams that efficiently build novel chips attacking the hardest problems with creativity and flair,” remarked Ho in his LinkedIn profile. However, he declined to provide further details on his new role at OpenAI.
Amidst these pivotal developments, Altman’s temporary departure and reinstatement by the OpenAI board have triggered discussions about the company’s future trajectory. Ho was among the employees who expressed intentions to leave the company during Altman’s absence.
Coinciding with Ho’s arrival, OpenAI also welcomed Google’s Todd Underwood, who was entrusted with leading a new Site Reliability Engineering team focused on research and training workloads.
As OpenAI solidifies its hardware division under Ho’s leadership, the company is poised for significant strides in shaping the hardware landscape of artificial intelligence, bridging innovation and functionality in the pursuit of groundbreaking AI solutions.