Jurvetson has a vision of AI everywhere and other VC see DNA and microbiome breakthroughs

What new trends will emerge in the next several years with the potential for explosive growth in about five years’ time?

The Churchill Club’s most anticipated events of the year: the 19th Annual Top 10 Tech Trends debate provided some answers.

Speakers:
Mike Abbott, Partner, Kleiner Perkins Caufield and Byers
Steve Jurvetson, Partner, DFJ
Rebecca Lynn, Co-founder and General Partner, Canvas Ventures
Sarah Tavel, Partner, Benchmark
Hans Tung, Managing Partner, GGV Capital

Masters of Ceremonies:
Quentin Hardy, Head of Editorial, Google Cloud
Mike Perlis, CEO, Forbes Media

Looking ahead to 2022, Jurvetson now envisions a world that extends intelligence to the edge of the Internet of Things (IoT) and the day when voice becomes the dominant user interface enabling, “continuous persistent communications.” The majority of fellow panelists preferred to focus their big bets on biotech developments, specifically on advancements in DNA and human microbiome applications.

Jurvetson predicted voice will move from the interactive voice response being championed today by Amazon Alexa and Google Home to become the user interface for all devices in our everyday lives. New chips the size of shirt buttons that use little power will be layered on to all sorts of devices for as little as 40 cents to provide “always on, ambient listening things to respond to your voice and do what you want them to do,” he said.

Battery-powered voice devices won’t need Internet connectivity, he said. But this scenario requires improvements in deep learning, where he sees a lot of research that will bear fruit within the next two to five years. Fellow panelists tended to agree, only challenging the timing and whether voice would be on the only interface.

Jurvetson referred to his prediction on embedded inference engines as “tiny brains in everything” from appliances to cars, drones, robots and more. With 30 billion connected devices forecast in the next three years and trillions of devices to be connected to the Internet thereafter, Jurvetson said, if we don’t go to this model “the Internet would collapse under the weight of IoT traffic.”

Pushing intelligence to the edge also means, “data will stay local, and it will provide better privacy, latency, security, lower cost, better bandwidth and utilization,” he said.

All of this will be made possible by inexpensive neural networks running on a variety of sensors and other devices. Jurvetson’s perspective comes from chip investments made in companies such as Nervana and Movidius (bought by Intel) and his current investment in Mythic, which is designing an analog chip geared to dramatically lower costs and power consumption.

Nervana has hardware optimized for deep learning. The Lake Crest chip (coming in 2017) is an application specific integrated circuit (ASIC) that is custom-designed and optimized for deep learning. Lake Crest includes everything you need for deep learning and nothing you don’t, resulting in a 10x increase in training speed and ensuring that Nervana Cloud will remain the world’s fastest deep learning platform for the foreseeable future. Training deep learning networks involves moving a lot of data, and current memory technologies are simply not up to the task. The Nervana Engine uses a new memory technology called High Bandwidth Memory that is both high-capacity and high-speed, providing 32 GB of on-chip storage and a blazingly fast 8 Tera-bits per second of memory access speed.

Data will be coming from many hyper-efficient sensors. DARPA is achieving a breakthrough in reducing power consumption in IOT sensors that could reduce power requirements by 1000 in passive sensors and 20 times in active sensors.

Jurvetson agreed creating good models would be the biggest challenge, except for data rich companies like Google and Apple. Sarah Tavel of Benchmark also questioned the use cases and if the overhead required for what would typically be relatively low-powered devices would create enough value for effort.

Among other popular predictions from the panel:

* Mike Abbott saw a rise of DNA applications due to low cost sequencing.
* Hans Tung of GGV Capital said food production will be revolutionized globally.
* Rebecca Lynn of Canvas Ventures said microbiome engineering will help prevent a future plague.
* Sarah Tavel forecast a billion-dollar business will be built on understanding our microbiome.

About The Author