Monday last week I had dinner with Douglas Hofstadter and Daniel Dennett and had a swell time. We went to The Elephant Walk, which was quite deluxe even though the waiter was a tad stretched thin, and consequently the promptitude of service was sometimes lacking. I had some kind of spicy tofu thing. Also a really rich and handsome chocolate tartish desert.
I actually felt smart for most of the evening, although somewhat self-conscious about the hole in my mouth where a crown had fallen out a few days before.
There were eight in our party: Dennett, Hofstadter, me; two graduate students whose names I don’t recall, a woman from Brookline whose name I don’t recall who is a friend of Hofstadter and who makes documentary films for a living (she was quite pleasant), professor Deb Roy from the MIT Media Lab, and finally Geoff Arnold, a Distinguished Engineer at Sun Microsystems. Geoff is a student in Dennett’s “Philosophy of Mind” class; coincidentally enough he and I were coworkers at Sun back in the day (late 80’s).
Apologies for not remembering the names; I’ll see if I can track them down and post in a comment just so as not to diss anybody.
As reported here, I had contacted Dr. Hofstadter by email as a shot-in-the-dark exploratory gesture on the occasion of my putting together a mashup of my novella Cheap Complex Devices, which is basically an homage to DH (among others . . .) and shot through with allusions to his Goedel Escher Bach. He responded that he probably wasn’t interested in the mashup but would be happy to chat with me by phone about AI storytellers, and we did so for about an hour some three weeks ago. That’s where the dinner invitation came from. Seemed that DH was going to be in Boston-town for two days, and would be giving two talks. One to the Philosophy of Mind class taught by Dennett at Tufts University, and the other at MIT. After the Tufts class a few people would be going out to dinner, and Hofstadter invited me to join them, subject to Dennett’s approval.
I had met Hofstadter once before, briefly, in 1980, and we had had a short email conversation in 1999. Essentially, however, we did not know each other before our phone chat in late March of this year.
I had had an email conversation with Dr. Dennett in the course of researching my Salon story about the Loebner Prize for progress in artificial intelligence. I was, frankly, a little nervous about the prospect of hanging out with Dr. Dennett. He had been very helpful to me in responding to my email questions, generous with his time, but I found him somewhat pompous and some of his opinions silly –and I said as much in my Salon story. He, contrariwise, hinted (in his email to me) that I was perhaps a bit pushy and possibly too dimwitted to understand what he was trying to explain to me.
I figured that the chances were pretty good that he had entirely forgotten our prior conversation, which would have been fine with me. But there was also some possibility that he would remember me and not want me along to spoil his evening. So I sent him a note telling of Hofstadter’s invitation, reminding him of our earlier chat, and asked him if I might join them for dinner — in the capacity of normal human being, not journalist. He responded with a gracious note saying yes, please join them.
Well I had planned to try to catch DH’s Tufts lecture but it was One of Those Days, stress off the Richter, the least problematical part of which was traffic from hell from Bourne to Boston, detours and accidents, and I arrived late at the lecture hall, nervous, sweaty, out of breath, just as the last stragglers were filing out. Hofstadter was standing at the front of the class talking with (the man who turned out to be) Prof. Roy. I recognized Hofstadter instantly. He still looks boyish (“Peter Panish” is how I described him in Cheap Complex Devices) but his hair is all grey now, and he appeared small, almost frail.
A big white-bearded fellow at the far side of the lecture hall boomed out, as I walked in, “Perhaps this is Mr. Sundman now!” I hurried over to introduce myself and nearly tripped over Geoff Arnold on the way. Which was a happy but disconcerting circumstance, as I had not seen Geoff more than thrice in the preceding fifteen years, most recently in 2000, and certainly was not expecting to see him there in the front row. So I said hello to everybody and apologized for missing the lecture and stood there feeling awkward but quite relieved that I had not missed the dinner party and flattered that Dennett knew my name unprompted.
We walked to Dennett’s office. En route I strolled with Geoff and asked him how he came to be there, and found out that, at the age of fifty something or other, Geoff has gone back to school to study philosophy. DH and DD went into Dennett’s office to make dinner arrangements and a few phone calls, I borrowed Geoff’s cell to call Dear Wife and tell her that the temperatures were predicted to be well below freezing that night (after we had spent hours on Sunday bringing her house plants outdoors for the summer) and then we all kinda hung out for fifteen or twenty minutes–I believe we were waiting for Doug’s friend from Brookline. I stood in the vestibule with Geoff & the other non-profs while inside Dennett and Hofstadter and Roy were in the office chatting.
I wasn’t exactly eavesdropping, but I did pick up some bits of their convesation — about Rubick’s cubes, and about various obnoxious Nobel laureates each of them had met — and other idle talk. (There was some good catty gossip that I shan’t repeat here.) Geoff talked with me about some of his work at Sun (see below). Then we all left en block for the parking garage. Roy was talking with Hofstadter. He mentioned that the first time he had seen a Rubic’s cube he had solved it in half an hour.
“No! Really?” Hofstadter said.
“Yes,” Roy said.
“YOU should get a Nobel Prize,” Hofstadter said.
Geoff rode in my car. Like Daniel Boone, I did not get lost en route, but I do admit to being confused a bit. Eventually we got there. Along the way Geoff & I further caught up on personal histories, and I experienced a mini nostalgia trip, because we were tooling around (confused, but not lost) the very neighborhood in which Betty and I had our first apartment nearly 25 years ago in the months just before and after our wedding.
At the Restaurant
I sat at a corner spot opposite Dennett, adjacent to filmmaker-woman-whose-name-I-cannot-remember, who sat next to DH.
Here are some of the main things we talked about:
- Creationism and pseudo-creationism (‘intelligent design’) and metaphysicians at Notre Dame U and a recent debate there
in which Dennett took on Michael Behe, author of Darwin’s Black Box
- My Salon article about the Loebner Contest, and that contest in general
- Whether enlightenment or fundamentalism will prevail in USA; also Dennett’s forthcoming book proposing the scientific study of the social phenomenon of religion
- That we all want Hoftstadter to hurry up and finish the book he’s working on, although I cannot now recall what it’s to be about
- There was some talk of “machine-written” books, and I handed out a few copies of Cheap Complex Devices.
- Marvin Minsky
My main impression of the evening was that Dennett and especially Hofstadter were very nice people. Thoughtful listeners, good conversationalists. Deb Roy is very nice but he was quiet that evening. It was also quite gratifying to learn that Hofstadter really liked my Salon article and has been spamming his friends to get them to read it.
In terms of net minutes spent talking at our end of the table, I think Dennett won, mostly talking about pseudocreationism and his invitation to Notre Dame. We spoke of Calvinsim, of a Calvinist college in the midwest– another memory blank– of the circumstance of being a professional metaphysician in the year 2005. Dennett said he had gone to Notre Dame expecting a real debate and come cogent analysis, but what he found was essentially religious preconception dressed up in fancy talk. However, he had nice things to say about Notre Dame in general.
Dennett also told some stories about visits to the home of Marvin Minsky, legendary genius and eccentric. (As I’ve mentioned elsewhere on Wetmachine, Oliver Steele, my friend and boss, is married to Margaret Minsky, Marvin’s daughter. Henry Minsky, Marvin’s son, is a colleague of mine. Sometimes Oliver, Henry and I have work meetings at Marvin’s house, and I’ve met Marvin a few times. But I don’t really know him and would be astonished if he remembered me.)
1) Marvin likes nicotine chewing gum. He’s chewed it for years and has no desire to cut back.
2) Marvin has a trapeze over his dining room table.
3) Marvin has a lamp made out of a device with an intricate mechanical assembly. He challenges people to try to figure out what it was before it was made into a lamp. Dennett examined it for an hour before giving up. The answer is, nothing. Marvin had the lamp made to order by a sculptor who does such things. The mechanical assembly has no purpose other than to baffle Marvin’s guests.
Hofstadter said some complimentary things about my article, as did others, and we talked about that, and about Hugh Loebner, and about the experience of writing for Salon. For example, for my first article I was paid fifty bucks. For my second, fifty bucks and a free year’s subscription. For my third, five hundred bucks. Pay rates for all of them averaging out to about a penny an hour, but all undertaken in the hope (partially realized) of selling more of my books.
I spoke with my neighbor (…) about the films she has made (with her husband), and about the world travels they have undertaken to make them.
For me the most interesting part of the evening came when we talked about whether fundamentalism or enlightenment would prevail in the USA. I asked Dennett if he was optimistic. He said he was. I said that I had been an optimist all my life until Bush’s reelection, but that events since, culminating in the Terri Schaivo freak show, made me fear for our very republic, and indeed for the human prospect. Dennett was quiet for a while and then said, “ask me again a year from now.”
I don’t recall Doug Hofstadter saying much that evening, at least not much directed to my end of the table. But understand that he had just given a lecture and entertained a bunch of questions thereafter and presumably was somewhat talked out. He was not withdrawn, he simply was not the dominant voice at the table. Everybody in the party was articulate and courteous and thoughtful. Sometimes they were funny, too.
When the check was presented, Dennett announced “I’m paying for half of this.” Whereupon Geoff Arnold instantly replied “and I’m paying for the other half.”
It was a grand evening.
Subsequently I had an email exchange with Dennett picking up where we left off on the whole Loebner story. He had not read the Salon story before that evening, but had asked me to bring with me a printed copy, which I did. A few days later he wrote me that it was “very funny, very well written,” but that I had not been quite fair to his arguments, which, he said, he had perhaps not explained well enough — and explained again. I replied that he had already explained his point quite well both in his email to me and in his book, which I had purchased and read, but that I simply disagreed with him. Actually in my email I asserted myself quite forcefully on a few points and accused him of being “anti-scientific.” I’ve not heard back from him on that! I would like to share that email exchange with y’all but (a) I haven’t secured his permission and (b) rather than pollute this account with an old squabble, I think I’ll leave that for another wetmachine entry some time hence.
In January 1986 Sun Microsystems was still, by some definitions, a “startup”. The mother-ship home office was in Mountain View California, and there was an East Coast operation that was just getting rolling, a startup within a startup. Geoff was the first engineer in that East Coast operation. I was the seventh. Geoff was the architect of PC-NFS, the network file system, and I wrote the manual. Within a few years Geoff was a Distinguished Engineer and I was managing a bicoastal group of 45 people. I was at Sun for nine years, and my professional path did not cross Geoff’s very often after the first three years or so– although we worked in the same building and ate in the same cafeteria.
Geoff now works, as I understand it, as some kind of “visionary in residence” reporting to Greg P, Sun’s chief Technology Officer. I asked Geoff what he was doing these days.
“Oh, blue sky stuff. Futuristic. Very large scale clusters.”
“What’s large scale mean?” I asked.
“Let’s say 10,000 nodes at the smallest. Now, at that scale, you can’t talk of an operating system. You cannot even monitor the system; it has to monitor itself. At that scale, with the cpu speeds on the horizon, relativistic effects start to happen. So how do you devise applications to work in this complex? How do you even know what an application is?”
“Well, how do you?”
“Here’s a starting point: all applications must be able to evolve without human intervention.”
“Self-modifying code, like DNA?”
“Sounds like “Colossus, the Forbin Project. War Games. Let’s play global thermonuclear war,” I said.
“All applications must also be self-defensing,” Geoff said.
“You mean like William Gibson, Neuromancer?”
“Hey, it’s happening. I was talking with the CTO of [some big securities company] and he said that their operating assumption is that every node in their network is compromised in some way. They just assume that. So the challenge is to engineer software that can be trustworthy in such an environment. It has to be able to defend itself.”
The next day I asked Geoff in email if he could answer a few of my particular questions about all this stuff. Alas, he said no. It’s all highly speculative and confidential. Oh well.
I did attend Hofstadter’s talk at MIT, which was good, informal, thought provoking. The general subject was analogies. The room was packed. But let that talk be a topic for another day. I chatted briefly with DH after his talk. He was being mobbed and I did not really want to impose on him. Besides, I needed to get back to work. I told him a few tidbits related to his talk that I thought might amuse him. He said thanks and that we should stay in touch.
I have since written one brief email to DH and two long emails to Dennett, with no answer from either of them.
Kinda wish I _had_ co-authored that Loebner article!
There’s a nice article on "The Global Computer" at http://www.hpl.hp.com/perso…
Karp also points out that you have to assume that nodes will fail (he’s the guy who taught me this philosophy), but makes it work without going into a whole metaphysical trip.