Interview-Vernor Vinge pg2
From FutureNovo - Anticipating things to come
August 10, 2008
Richard Yonck, FutureNovo
FN: In 1981, you wrote "True Names", one of the earliest stories to describe a recognizable, if highly advanced version of cyberspace. The virtual world you called "The Other Plane" influenced dozens of authors and anticipated virtual worlds such as Second Life by over two decades. Given where computing technology was in the early eighties, that's quite a conceptual leap. Which trends were you seeing at the time that led to the ideas behind the story?
Vinge: There's a truism among science-fiction writers that science fiction may look like it's talking about the future but really it's just a mirror of the present. There is a large grain of truth in that, but it would be more precise to say that each SF story is a reflection of the present of the story's author. And my present from the early 1950s was someone immersed in science and tech speculation. From the late 1950s on, this included computers. I was surrounded by things that contained elements of the "True Names" environment. For one thing, I'd been exposed to fantasies in which "true names" were very important — Ursula LeGuin's Earthsea stories, for instance. That gave me the notion that if everything had a serial number, knowing the serial number would be power. Then just before I wrote "True Names", we got dial-in access to the university computer. (An awesome machine with something like a quarter megabyte of RAM!) I could check out a portable teletype-style terminal and modem. I could log in from home! This wasn't really the Internet, just a very ad hoc phone link. However, that was enough that I could imagine consequences. I remember being logged in. Some other user ran the talk program on me and we had this anonymous real-time chat. After I finished the conversation, I realized that by my standards I had just lived a science fiction story. It was very easy to elaborate on the environment. So "True Names" was actually very easy to write.
FN: Amazing. One of the features in the story I liked was the idea of the high-level programming paradigm that you used that's since been referred to elsewhere as intention-based programming. Though there have been attempts at developing something like this, do you think it's feasible that we'll ever implement a programming paradigm as seemingly intuitive as you described in the story?
Vinge: I think we're going to get tools — and we already have tools — that make it easy to chunk very large pieces of functionality. On the other hand, my suspicion — and I think a lot of people have this suspicion — is that the way massive multi-core systems will solve problems that aren't naturally accessible to parallel processing will be via bioscience paradigms more than engineering paradigms. Exploitation of massive parallelism will probably involve cognitive biomemetics. This is scary to some people (sometimes including me!) because we grew up with the ideal of determinism in programs, especially software for life critical systems. On the other hand, before computers, people routinely depended on such unreliable systems (horses, dogs, etc.).
FN: Something that has it's own volition.
Vinge: Yes, an animal has its volition, even if it's below human parity. We know that in life, there are important gotchas we haven't figure out, or that are beyond our control.
FN: You mean in a highly complex system?
Vinge: In life. Leave aside classical dynamics and linear algebra and other stuff where parallel works fine in a deterministic way. Let's talk about the things nobody right now can figure out how to do. If it turns out that the biological paradigm is the way to go in that, that's very scary to some people. I can remember a year or two ago, I was at Hot Chips (a leading conference on high-performance microprocessors) and this came up in a slightly different context: It may well be that as things get faster and faster and smaller and smaller that we'll just have to go for non-deterministic computing at the transistor level. That raised the issue: "but what then about life-critical systems?" There was one guy on the panel who raised exactly the point I just made at the very high integrated level that you and I are talking about — the dogs and horses example.
Consider a cavalryman riding along on his horse. What if the horse sees another horse that it wants to associate with? It might just turn aside. So I don't mean to pooh-pooh life-critical system risks. I'm just saying this is nothing new. As usual, the question comes down to levels of risk and levels of benefit.
FN: Ubiquitous computing and augmented reality were very prominent in "Fast Times at Fairmont High" and "Rainbows End". In many ways, they are going to directly and indirectly enhance human intelligence, but at the same time bombard us with a tremendous amount of input. How do you see us managing this sensory overload?
Vinge: If the Singularity does not happen, then one of the symptoms we would see is that hardware progress makes the data glut problem worse and worse. In some areas this actually would be tolerable, although it might produce a form of science such as Chris Anderson was talking about recently in Wired. ("The End of Theory: The Data Deluge Makes the Scientific Method Obsolete", Wired Magazine, 16.07) In other areas, it would just mean we have to give up on many projects, not being able to cope with the amount of data (much less create the associated information and knowledge). This would then decrease the demand for hardware improvements. So if the Singularity doesn't happen and it's 2050 and somebody's writing that hypothetical essay about why it was obvious that the Singularity never happened, the data glut would be part of the story arc.
On the other hand, the thing to watch as data collection increases is the extent to which automation is capable of dealing with it. Maybe 99.9999% of the data that's being collected is never seen by a human. But that data informs automation that ultimately is useful for decisions made. If this happens, one could argue that it is evidence of a shift in the cognitive-creative center of gravity, away from where it's been within humanity, off into our extended mentation. The sensory overload question is one of the really cool things to track.
FN: A signpost, essentially.
Vinge: Yes.
FN: Education is a recurring topic and theme in your stories. In several of them, students have assignments and tests that prepare them for a very different world, one where extreme change is expected and anticipated. There are Nobel-level scientists going back to school, struggling to keep up and fit in with the new fields and technologies while teenage students quickly take up techniques and concepts that were once totally cutting edge, even unheard of. How do you see our real-world education system having to transform itself to deal with the reality of accelerating change?
Vinge: That was one of the most intimidating things about writing Rainbows End; I'm not full of answers. First of all, while I think that children can be empowered, you recall that the children in both "Fast Times" and Rainbows End felt very intimidated. They were capable of boggling achievements — by our standards. For instance, the boy in Rainbows End could do real-time physical simulations for which a 2008 person would need years of hands-on experience and/or be very knowledgeable in physics and mechanics. But he didn't feel empowered. Basically he felt that any kid could do that and there were other things going on that were constantly moving away from him. Education was a real stress point in Rainbows End. I had the junior high school teacher, Ms. Chumlig, make the point that "You have to know something about something." It was one of her mantras. I think it's correct as long as humans are significant players. One perversely important category of knowledge is knowing what the glitches are in the various automation and filters that you're using. I used to teach numerical analysis and we'd deliberately hand out ordinary looking problems that would break calculators and computers — that is, cause automatic calculation to give wildly wrong answers. Mainly we exploited discretization errors. The goal was to teach what the traps and threats are. However, similar issues exist with other packages: computer algebra systems have their own world of glitches. These are not classical bugs, but they're issues you have to understand to be successful. Mastering such problems is very domain-related and in some cases it may change as the package changes. How this sort of thing interferes with our exploitation of software is going to be interesting to track — until we have packages so smart that they understand the psychology of the human user.
FN: That would certainly be helpful.
Vinge: True. Of course, when the apps get that clever, you've slipped over the edge. It's Singularity city.
FN: There's a real emphasis in your fictional schools on collaboration and affiliations or "affiliances" as you termed it in one story. I take it you see this as an important means of dealing with working in a changing world?
Vinge: Yes, empowering the group mind, so to speak. It's one of the paths to the superintelligence. Just as we're getting the digital Gaia, in a different arena we're using the Internet to empower humanity. The ensemble of educated, connected humanity is much more powerful than all of the experts any national government has ever employed. For the most part, humanity is good-hearted, but in the past that fact has often been trumped by fear and ignorance and poor communication. Now, I could imagine a new populism that is smarter than anything that went before and is so broadly based that it doesn't suffer from the short-sightedness of populisms past. This trend fits very well with the possibility of soft takeoffs. It's also a plausible counter to the way technology empowers tyrannies. This doesn't mean the end of governments or the end of nations, but given time, the empowerment of the people could be something very smart and wise that has a chance of managing the probably still larger changes that are coming as machines and the environment as a whole wake up.
FN: So you think that could actually overcome some of the potential issues of privacy and government and corporate abuse?
Vinge: Yes. A good example is David Brin's nonfiction book The Transparent Society, which takes the problem of government intrusion and personal privacy and just kind of turns it on its head — as a result getting both freedom and information. Are you familiar with that book?
FN: You know, I haven't read it, but that definitely makes me want to pick it up. My first reaction is that it sounds like the transition period could be very tumultuous.
Vinge: Oh, yes! I'm more privacy oriented than most people and the prospect chills my blood: the transition years would be awful. But my guess is that after that transition, in an era where passive surveillance was generally available to individuals and the data often pooled for integrated analysis, that the world would be pretty much like it is now — except that there would be fewer secret villains and the most hypocritical laws would have gone away.
