Let’s face it – although the design of new devices and operating systems and apps is going to be critically important, it’s also going to matter what kinds of environments tomorrow systems run on.
When we look at the past, we get incredible clues for the future. We see how Moore’s law and other advances started to create this “hockey-stick” environment, and we see what we’re likely to experience in the next 10, or 20, or 30 years, as new things build on what we’ve already established, with big data, and cloud, and AI/ML – and hopefully, we get just a little more confident about being able to navigate that world.
In light of that, when David Clark gives us a disclaimer early in his talk that ‘it’s not about AI,’ you could say by way of debate that it is critically related to what we’re doing in AI/ML and related areas. You just have to see the connection.
Clark has obviously been in the business for a long time – one of the high points of the talk is when he unrolls an old paper data feed from his old file system, dated August 11 of 1976.
That prop comedy leads into a better understanding of how systems have evolved in the last half-century and change
In talking about building the Internet, Clark repeatedly asks the audience: “are we done yet?”
The answer is: ‘no’.
Think about how Clark characterizes the Internet of Things in the video – elaborating on the goal of helping to democratize data, noting that it will involve a lot of work, and re-estimating the number of devices that will be connected to the trillions. (Lots of us remember seeing early estimates of something like 50 billion devices for the IoT, but that was a while ago.)
So: Are we done yet?
Clark gives us a good example of criteria for philosophizing about technology systems: who are these things going to be talking to? What will they be saying, and in what language? To what extent can they be trusted?
He also gives a concrete infrastructure example – traffic lights, of which an estimated 330,000, he says, exist across the U.S.
How do we know they’re all properly hooked up and safe from tampering?
In other words, security is paramount.
You also get a rare insight one of Clark’s quick asides as he talks about evaluating new “gee-whiz” technologies – of which we’ve seen many over the past few years!
Clark asks another question: do we want our systems to be open ecosystems, or stovepipe apps?
What are the interface specifications that are going to win out?
“Every set of things is going to have to come to grips with these questions,” Clark says. He advocates for open structures.
There’s another interesting point where Clark contends that early HTML was ‘fairly harmless’ until people wanted to inject active code to the client/server interaction. He invites us to think about how the budding tech capabilities eventually get used.
We concur that the open systems will fare better in tomorrow’s tech world, where we, by any account, will have to wrestle with pretty powerful AI systems. Look for more on this blog to get up to speed on what’s going on around you with all kinds of ‘smart machines’.
Read the full article here
