On the challenge of getting entrepreneurs jazzed about these industrial problems, O'Reilly sounded an optimistic note. "I love the idea that we have the chance to tackle big problems like healthcare and energy using the talent we have here," he said. O'Reilly thinks entrepreneurs will be motivated to make real products and established business models -- like taxis -- better.

But the business model could be one of the tougher nuts to crack. Accenture's Matt Reilly, also on the panel, noted that a lot of industry Internet investments will be around "micro" problems -- a big project by one company with a $1 billion capital investment. Those are harder to "open up," and the successes also won't create the next Internet star. "The revolution will not be televised. … It's about things that happen and you don't know," Reilly said. "It'll be about flights that are on time, luggage that got there, product that are on the shelf."

Industrial company CIOs worry that their technology challenges don't represent a big enough market for startups to tackle. For example, Union Pacific CIO Lynden Tennison lives the Internet of things, as the company's sensors, spread across 32,000 miles of train track, take 20 million readings a day to provide an early warning when a wheel is at risk of failing. But Tennison ends up having his in-house IT team build a lot of UP's technology. The problem is that even if a technology vendor gets three-fourths of the railroad market, it's likely still a small business, so that startup needs to charge a lot for its products, and it's often on shaky financial footing. Not exactly your dream tech partner.

I've heard the same lament in other industries. UPS CIO Dave Barnes regularly works with startups and will even invest in the promising ones. In early 2009, Barnes told me how he was pushing vendors for faster data analysis so that UPS could be more responsive in creating driver routes. And this past year, UPS leveraged today's faster database and processing technology to offer a service that lets customers change delivery instructions online on-the-fly, for an added fee. But like Union Pacific, UPS often has to build the technology itself. Said Barnes back in 2009: "If we believe in something and can't get someone to do it, we'll pave the path. But we aggressively look to partner."

GE and Ford both opened Silicon Valley offices this year with the stated goal of getting closer to startups and Internet-driven innovations. Such moves are great, but they're not enough.

The industrial Internet will thrive only if there are ways for entrepreneurs to experiment with new products and ideas without the industries' permission. We need people outside the system who solve problems the industry doesn't even know it has, or who expand on ideas the industry rejected or ignored.

A start is for industrial companies to share more of their data. GE, for instance, is launching a "quest" in which it will give people two months of flight information -- from flight schedules to wind data -- from the National Airspace System. The challenge is to create a predictive model to increase airline efficiency. It's proposing something similar for healthcare.

Even after we have freed up more data, though, we need something else: data jockeys. EMC's Maritz noted that it still takes a highly educated and talented specialist to make sense of these huge data sets. Consulting firms that specialize in big data are touting them as "god-like creatures called data scientists, there are 1,500 of them in the world, and we've already hired 1,000," Maritz joked. He makes the great point that we need tools that can "democratize" data analysis the same way we made programming something your average clever person could tinker with.

It's a long list of problems to solve, but problems are the gas that drives Silicon Valley. Although we lack sufficient connections today between the technology startup ecosystem and industrial industries, you can bet some firm already fancies itself the Y Combinator or Andreessen Horowitz of the industrial Internet.

Which takes me back to Paul Graham's post, and his advice that the way to look for startup ideas is to look for what's missing. Writes Graham:

"Most things that are missing will take some time to see. You almost have to trick yourself into seeing the ideas around you. But you know the ideas are out there. This is not one of those problems where there might not be an answer. It's impossibly unlikely that this is the exact moment when technological progress stops. You can be sure people are going to build things in the next few years that will make you think 'What did I do before x?'"