Vinod Khosla: A huge opportunity is removing the humans from IT

Vinod Khosla – Founder, Khosla Ventures

One of the biggest opportunities for entrepreneurs and innovators in the enterprise is cutting out the human IT staff and going full automation, potentially with technologies like deep learning, said venture capitalist Vinod Khosla at Gigaom’s Structure conference on Wednesday. “It’s ridiculous to have humans manage the level of complexity that they do. People are a big cost in IT. Let’s take that out,” said Khosla.

Picture taking the data that these IT systems are putting out and use that data to build adaptive systems that have the capability to utilize human judgement and decision making. If these types of tools could be built and sold to enterprise customers, it could dramatically reduce the cost of needing to use human IT staff to make these complex decisions. It would be like a next-generation Splunk, said Khosla.

These types of nimble applications are enabled by the rise of everything as a service, or software defined “X”, as Khosla put it. It’s the idea that businesses and startups don’t know what will happen with their internet companies down the road and don’t want to plan their infrastructure years in advance.

Customers don’t have a clue where their businesses are going to go, so let’s build infrastructure so they can respond. Plan not to plan, and be really agile.

Khosla says he’s optimistic that there will be a lot more interest and innovation around enterprise IT, following the success of startups like Arista. Big internet players like Facebook, Google and Amazon with AWS, are doing really interesting things with IT infrastructure and that’s opening up huge opportunities for everyone else, said Khosla.

However, it’s the older IT infrastructure companies that haven’t continued to innovate. “I haven’t seen innovation from IBM or HP or Dell in 20 or 30 years,” said Khosla, adding that Facebook probably doesn’t even look at Cisco’s product lines for its architecture.

Photo by Jakub Mosur

Structure 2014 ticker

loading

Comments have been disabled for this post