From 'Long Boom' to days of doom and gloom

Bill Joy, deep thinker and Sun's chief scientist, speaks about the dangers of technology, the future of Jini, and Sun's stewardship of Java

Bill Joy, chief scientist and cofounder of Sun Microsystems, recently threw a cold bucket of water on the red-hot high-technology industry. Joy's view is that the prosperity of the high-tech industry can be realized only if its participants become better stewards of potentially dangerous technologies, notably robotics, nanotechnology, and genetic engineering -- all of which are enabled by computer science. Joy, who outlined his thesis at InfoWorld's CTO Forum conference in San Francisco, believes that the virtual and the physical worlds could be catastrophically affected by unmarshalled technologies and that high-tech leaders need to be more conscientious and proactive in order to pre-empt world disaster in the future. Conversely, fellow futurist Peter Schwartz, the author of the highly optimistic book The Long Boom and also a keynote speaker at the CTO Forum, believes that the world is headed for unprecedented prosperity due largely to the productivity garnered from IT investments and advances in other technologies. Joy sat down with InfoWorld editors Katherine Bull and Ed Scannell to further explain his views on the future, as well as Jini, his latest project at Sun.

InfoWorld: Peter Schwartz took a contrarian view to the piece you wrote in Wired on the dangers to humanity posed by technology. Why is that? Bill Joy: It is clear that within the next century there will be very grave dangers from certain technologies because they are basically military-class in their ability to cause harm [and] because they are based on information. In other words, you can design them just using a computer. We have no policy constraints to prevent individuals from doing that. The same is true of the Internet and personal computers, given the way Microsoft has done their software, which is poorly. Look at the Visual Basic viruses. The problem we have over the longer period is that these technologies are so powerful that they can do damage to the physical world. It is very easy to have optimistic, naive, and uninformed opinions -- not saying that Peter's opinion is; but unless you have read the article or have studied the stuff, then it's easy to say, "OK, it is all going to work out." InfoWorld: What kinds of damage to the physical world do you think could occur? Bill Joy: In the case of nuclear weapons, we have had them a long time, and we have been lucky in that nothing has happened yet; but cleverness will not make a good defense against them. We can't make a missile-defense shield that will work, because you can simply overwhelm the shield with decoys, and the cost of any miss is huge. Similarly, if you have a Microsoft mailer with no defense -- because of stupidity in this case, because they allow any user to run any macro -- you are going to have catastrophes like the ILOVEYOU bug. So the question is, are we going to allow unrestricted use of these powerful technologies throughout the course of the century, putting ourselves in danger? We could easily do that. I guess we could say "oops," if we are all around to say, "oops." InfoWorld: Who will be responsible for regulating and controlling the problems that could arise? Will it be the government? Bill Joy: Government is really us. If we can agree to that, then it is quite clearly the case that people can develop things so powerful [that] it could be dangerous. And this stuff could be information that could be just passed around. Look at the ILOVEYOU virus. Worse than the virus itself, everyone now knows what the virus did. So everyone has the source code for that thing, and you can build on that knowledge. That seems to me to be exposing a very basic weakness in the system. And so millions of people are going to have to repair the security hole that Microsoft left in their mailer. InfoWorld: What are the odds that people are going to do that? Bill Joy: Right, people barely understand what they are doing when they use a computer. Some barely know how to turn it on and they will do [only] what they have been taught. It is a situation we should not have got ourselves into. There were plenty of things in the press in 1995 where we were telling people that ActiveX does not work. Well, Microsoft didn't say to turn off all attachments to mail. They could have used Java and had attachments because it was [built] with the proper security. In handling these technologies, which are potentially destructive because of the way they couple together and replicate, you have to take added precautions. InfoWorld: So do you think we need to take social responsibility? Bill Joy: We can't assume that everyone on the planet is going to be socially responsible. So the question remains: what size weapon are you going to give to psychos out there? And there are psychos out there. Ted Kaczynski, for instance, was clearly the case of a psycho. And the psychos are not just young people; they can be older people with a lot more knowledge. So maybe what we should do is invest more in mental health. I don't think we can have a civilized society if we give individuals an unlimited amount of power. Look at the stories written in 1995, '96 about Java and ActiveX and security issues. People were just lying. [Microsoft] shipped millions of copies of the software with holes that people could drive a truck through, and so someone did. [Microsoft was] completely unrealistic about the security flaws in the system. When you allow people to run unfiltered attachments, that's dangerous. There were plenty of discussions in 1995, '96, and we should not be surprised [that these viruses are now occurring]. InfoWorld: So what do you think we need to do in order to control the damage that can be inflicted? Bill Joy: Well, in terms of computer viruses, we can do a better job of engineering the systems. The state of the art is far beyond what these systems are actually doing. It is good to see that Microsoft is using a patch now. It is several years too late because you have all these copies [of Windows] out there. The truth is that even well-designed systems will have security flaws. We need more genetic diversity in our systems and that is a good thing. If the systems are all identical, then they all have the same flaws. That is the perfect situation for a pathogen, whether is it a virus in the computer world or the real world. InfoWorld: When is Sun's Jini going to stop being slideware? Bill Joy: It is a Metcalfe's Law kind of phenomenon, which means it is an infrastructure play, which means it starts slow. The other thing about Jini is that it's about connecting devices which are connected to persistent and pervasive networks. And we seem to have a problem in that the architecture for persistent and pervasive networks has been slow to develop. You can look at wired nets and wireless networks. In the case of the wired networks, it is hard to get DSL, and [then there's] how slow cable deployment has been and the difficulties of getting it to the home. Even ISDN is hard [to get]. On the wireless side, we have digital wireless networks in this country which are in shambles because there are so many different ones. So the trouble is, the hit products in the pervasive and persistently connected world do not have the infrastructure yet. We have seen [this problem] in places like Japan where we have put Java in phones. The hit products will drive these markets, [but] the hit products can only emerge when the network infrastructure is in place. We see Jini and Java being those hit products, but they are being held up by the lack of a network. InfoWorld: Jini takes up a lot of hardware resources. How are you going to get it onto pervasive devices such as Nokia phones? Bill Joy: We are working on making a smaller-footprint version of Jini, and we have already made a smaller-footprint version of Java. InfoWorld: And your goal is to get it down below 1 MB? Bill Joy: Yeah, we can clearly do that. I think where there is a need for Jini as a small footprint is in the space of sensors. We are putting a lot of these sensors everywhere. Some of these sensors are wireless or are in a one-chip device like a radio or an airbag sensor in a car. You could have sensors in the engine of a car or sensors in a jet engine. Those are the perfect devices to be included in a Jini-like network. But they transmit only, so they don't need to have the ability to receive data. We hope to put this [capability] into the first version of Jini. That is what I hope to roll out this year as a less resource-intensive version for the sensor world. We have been working with companies like Echelon on technologies like that. InfoWorld: There is a concern about Java as a standard fracturing. Why should we believe that it will remain intact? Bill Joy: Well, the reason there is continuing discussion of this is because Java is so important. You can see with the ILOVEYOU virus that Java is even more important. It is the only way you can have safe attachments. So this gives a chance to go back and rediscover that. My view is, the industry has become balkanized. Everyone is feeling their oats, and everyone wants to believe they are going to be the next IBM or Microsoft. What we were trying to do with Java is [to] focus on the customers. The reason we are so firm on holding onto the control of the process is for the benefit of the customer. The promise we made to customers was [to keep Java] cross-platform. Frankly, I don't trust these other companies who don't seem to care that much. I don't think they have that cross-platform vision as strongly in place as we do. It wasn't Microsoft's intent. Clearly they wanted to take it and put it on a single platform. It is only natural that people want to make it work best on their platform and break the rules. We have tried to act as a steward for the sake of the customer. We are willing to go into these other processes and participate in good faith if other people are entering into it with good faith.

This story, "From 'Long Boom' to days of doom and gloom" was originally published by InfoWorld.