Trip Report on Hackers Conference at Yosemite.
This past weekend--Friday to Sunday--was the Hackers Conference. This year's attendees included Gordon Bell, chief designer of the PDP-8 and PDP-11 (and later of Encore and now Microsoft Research), Don Knuth (of the books on programming), John McCarthy (inventor of LISP and pioneer in AI), Al Alcorn (co-founder of Atari and designer of Pong and the Atari 2600, etc.), Bob Frankston (VisiCalc), and so on. Nearly all facets of the computer industry (less so the chip industry) have been represented in past years.
I usually try to attend every other year or so. Here are some things I think might interest you folks on this Intel thread. Because crossposts to other threads is not possible (that I know of), some items going beyond just Intel or the "AMD vs. Intel Jihad" will be included.
First, about a third of the 200 attendees were using laptops at any given time--this is a lot, considering that attendees don't have to take notes for business reasons. The two most common brands were Dell and Apple. In particular, I counted at least 10 Apple Titanium Powerbooks (sometimes dubbed TiBooks), which surprised me, as the smaller iBook is outselling the TiBook. But this is an "upscale group," which can afford the more expensive machines.
By the way, many of the i86-based laptops were running Linux or FreeBSD or NetBSD. However, nearly every presentation was made with either a laptop running PowerPoint/Word or with a Mac (about a 60/40 split I would estimate). It looks like Microsoft Office is indeed a powerful selling point. Surprisingly, I saw no Star Office presentations (Star Office is a kind of Microsoft Office open source/Linux clone, now distributed by Sun).
Second, many of the machines being used for demos in the demo room were Pentium IIIs and 4s. Many presenters gave benchmarks in terms of Pentia. For example, Dan Ingalls (the inventor of BitBlt, the essential tool for windows environments, and the chief designer of Smalltalk) described a 1.7 GHz Pentium 4 as executing 150 million bytecodes per second for a version of Smalltalk he's working on (this is _very_ fast for running bytecodes for very high level languages).
AMD Athlons were invisible, or at least never mentioned. No mention at all of TransMeta. It looks from this conference that Intel has basically won the processor war, vis-a-vis AMD and H-P and Sun. (See next point about Sun.)
Third, a friend of mine who had 500 people under him as head of one of Sun's graphics group had his entire lab shut down and all or nearly all of the staff laid off. Whew. His group did a lot of the rendering work on recent graphics movies, even taking over from SGI when they faltered. He didn't say it's because Intel boxes are getting more of the business (I neglected to ask him!), just that Sun is tightening its belt and cutting way back on anything outside the actual business of making and selling boxes. "Service" units doing rendering work and graphics research are not their mainstream stuff.
Fourth, the Nvidea GeForce 2 was cited by several speakers (and it's in the new high-end PCs and Macs). One guy who's been using supercomputers for decades and who is now doing the Sloan Digital Sky Survey (billions of stars and galaxies...a very complete may of the entire sky) said that a $100 GeForce 2 graphics board outperforms his fastest supercomputers (U. of Pittsburgh, I think he's at).
Fifth, there was a heated discussion about whether we are seeing "the end of electronics." Not a true end, but an asymptotic flattening out, with performance increases slowing down. Someone pointed out that there once were thousands of electric motor designers, and many dozens of brands. But now there are probably only a dozen designers of electric motors in the U.S. (is this what processor design will be like?). Opinions on both sides. Some said "biotech" will be the next big thing, as many have said, but others pointed out how vague and sketchy that claim has always been.
Sixth, it's not clear what the Next Big Thing will be. At the first Hackers Conference I got invited to, in 1988, Virtual Reality (VR) was popular, with numerous head-mounted displays, HDTV displays, and folks like Jaron Lanier and John Walker (of Autodesk/AutoCAD) were giving demos of submersive VR systems. This year, none of those demos. Trends come and go.
Seventh, it was striking just how deeply the dot com crash has struck. I was not at last year's conference, but this year it was crystal clear about the impact. Many people were joking about being unemployed, and many more said their companies were limping along. This included some people with some innovative, core ideas, not just the Dilbert-style b.s. product ideas.
Eighth, wireless networks were everywhere. A bunch of 802.11b transponders were deployed throughout the lodge (mostly Apple/Airport and Lucent/Wavelan). Interestingly, a well-known hacker who is famous for "war driving" showed maps he's made of the Bay Area with "open networks" identified. He was basically the guy who used to "war dial" (as in the movie "War Games") hundreds of thousands of phone numbers to find computers hooked up with modems...and see if his computer could get in. "War driving" is driving around SF, the Peninsula, etc. with a laptop and Wavelan card and a GPS hooked up, looking for open access to 802.11b ports. These open ports are ones where any machine can make a connection, either because the users are not bothering with any encryption or are using the default-issued passwords. He found thousands of open nodes...homes, small companies, large companies, government agencies, universities. The GPS gave him a map of these open nodes. He also mounted a Yagi high-gain antenna up in the hills around the Bay Area and can pick up open nodes 25 miles away (sometimes). Someone like him (cough cough) could slurp up passwords, gigabytes of proprietary data, and of course he could access machines at his leisure. YOU HAVE BEEN WARNED!
Finally, a lot of cool stuff was demoed in early or prototype forms. We're not supposed to write about these things, so I won't. I don't think electronics and computers have run their course, not by a long shot.
--Tim May |
