Code? AI?Internet co-creator Robert Kahn was already doing it…decades ago

Image credits: Peter Adams/IEEE

Robert Kahn has been a consistent presence on the Internet since its founding, obviously since he was the co-creator of the Internet. But like many technology pioneers, his resume is longer than that, and his work actually pre-empted ostensibly modern ideas like his AI agents and blockchain. TechCrunch spoke to Khan about how nothing really has changed since the '70s.

This interview was conducted in honor of Khan (referred to as Bob during our conversation) receiving the IEEE Medal of Honor this week. The award ceremony and speeches can be viewed here.

Sound familiar? Last year, the IEEE awarded the medal to Kahn's partner Vint Cerf, who helped create the protocols that underpin the Internet and the Web. Although they have taken different paths, they share a subdued optimism about the world of technology and a sense that everything old is new again.

This interview has been edited for length and clarity.

Many of the technical and other problems we currently face in computing and the Internet are problems that we have experienced and have probably solved before. I'd like to know if you have any particular thoughts about the challenges we face today.

Khan: Well, I don't think there's anything particularly surprising. So I was concerned from the beginning that the Internet could be misused. But in the early days, it was a very motivated group of collaborators from the research community, all of whom primarily knew each other, or at least knew each other. So there wasn't much that went wrong. If there are only 100 people who do not know each other, then perhaps it will work, but if there are a billion people, then you can get a little bit of everything in society.

[CERN leadership] In fact, they approached me with the possibility of setting up a consortium, which they later set up at MIT…and probably asked too many uncomfortable questions. For example, what about misinformation and disinformation? How are you going to control this going forward? I thought there was an approach. We were actually working on a few things. So in a way, I'm not that surprised. I am disappointed that a potentially transformative approach was not adopted.

I was reading about “Knowbot”. This is very similar to an AI agent and can be accessed and interacted with in a less structured way than API calls or simple crawls.

The whole idea was launched in the form of a mobile program [i.e. the program is mobile, not for mobiles]; we called them Knowbots, short for Knowledge Robots. You told it what you wanted to do and launched it. You're just free to book flights, check your email, watch the news, and stay informed about things that may affect you. It will involve placing bids on the internet.

We basically made it available and couldn't have been more disappointed. He was in the late '80s, right around the time of the first cybersecurity threat, the Morris worm. It was something that someone did by accident, but people looked and said, “Hey, if this bad thing is going to happen, we don't want other people's programs showing up on our machines.” Basically, I'm putting it off.

But I think something very useful has come out of it. We called this Digital Object Architecture. You've probably been following some of the research surrounding cryptocurrencies. Well, cryptocurrencies are like taking out a dollar bill and removing the paper. Then you will be able to deal with the value of money online. Digital Object Architecture was like a mobile program stripped of mobility. The same information exists, but is accessed differently.

Robert Kahn receives the IEEE Medal of Honor.

It's interesting that you talk about digital object architecture and cryptography in similar sentences. We have a DOI system, which we mainly see in the scientific literature, and of course it's very useful there. But as a general system, I noticed a lot of similarities with the idea of ​​cryptographically signed ledgers and the canonical location of digital objects.

You know, it's unfortunate that people think these digital objects have to be copyrighted material. I wrote a paper called Representation of Value in Digital Objects. I think they called it a digital entity for technical reasons. I think this was the first paper to actually discuss the equivalent of a virtual currency.

But we've been talking about linking blocks all the way to the end… Back in the space age, when you wanted to communicate with a distant part of the universe, you didn't have to come back and wait for minutes. It takes hours before the transmission is delayed back to Earth to fix something. I would like to link the blocks in transit. So you can figure out what went wrong with that block before the next block is released, which may arrive milliseconds later. And that's what blockchain is for.

Digital object architecture talks about the ability of digital objects to communicate with other digital objects. It's not people sitting in front of a keyboard. You know, sending a digital object or a mobile program to a machine and asking it to interact with another digital object that represents a book, to go inside that book and do some work, to interact with that system. can do. Or, like airplanes, people think that planes need to talk to other planes for things like collision avoidance, and cars need to talk to cars because they don't want to crash into each other. But what if a car needs to communicate with an airplane? These objects can be anything that can be represented in digital form, so anything can interact with everything. This is a different concept of the Internet than high-speed communication lines.

Well, even if it's a plane in a car, it's important to know if an object needs to communicate with another object and to enable it as a protocol. In the so-called Internet of Things, we have connected doorbells, connected ovens, and connected refrigerators, all connected to private servers through private APIs. It's not a protocol issue, it's just that there's a really bad software service in your fridge.

I truly believe that most actors who would have had a natural interest in the Internet had expectations that their approach would carry over to the Internet. [rather than TCP/IP]. Everyone had their own approach, whether it was Bell Systems, IBM, Xerox, or Hewlett's Packard. But what happened was they kind of hit rock bottom. Must be able to demonstrate interoperability. You can't go in there and ask everyone to get rid of all their old stuff and take their own stuff. So they couldn't choose his one company approach and stuck with what we did at DARPA. That's an interesting story, but I don't think it's necessary to write about it (lol).

If every house you went into had a different power plug, you'd have a big problem. But the real problem is that you can't see it until you implement it.

I don't think we can rely on government initiatives. I don't think he can take control of the industry. Because there could be five or ten different industries competing with each other. They cannot agree on whether there should be a standard until they have exhausted all other options. And who will take the lead? We need to rethink this at the national level. And I think universities have a role to play here. But they may not necessarily know that yet.

The US chip industry is seeing massive reinvestment. You were closely involved in his fundamentals in the late 70s and early 80s, collaborating with people who contributed to defining his computing architecture at the time, and of course influencing future architectures. I know. I'd like to know what you think about the evolution of the hardware industry.

I think the big problem right now, as the administration has clearly pointed out, is that our country is not maintaining its leadership role in semiconductor manufacturing. They come from Taiwan, Korea, and China. We're trying to fix that, so kudos to you for that. But perhaps the bigger problem is human resources. Who will manage those sites? I mean, build manufacturing capacity, but do we need to import talent from South Korea or Taiwan? OK, let's teach it in schools… Don't know enough to teach it in schools? Is there anyone there? Are we going to hire people to teach in schools? Workforce development is going to be a big part of the problem. But I think we've been there before, and we can be there again.