The rich are getting younger...

Rohit Khare (rohit@uci.edu)
Fri, 02 Oct 1998 03:33:43 -0700


1. I've been reading a lot in different corners on the poverty statistics in
the US and I'm getting a very nagging feeling about what Greg Bolcer raised
this week: I'm poor, and I shouldn't be. I have to believe that this isn't
distorting the national statistics -- that is to say the ~11% overall rate
(including 3% of all blacks, 7.6% white from Forbes last week) actually
represents poor people. It's OK, I tell myself, there aren't more than a few
hundred thousand, maybe a million temporary unemployed (like gradstudents)
or dividend-collectors (probably responsible for the skew and dramatic end
of the 'poor have marble bathrooms in the US!' effect).

2. The world is simultaneously getting richer for all and lags behind
pitifully. The Julian Simon school of
increasing-returns-through-many-hands-and-many-minds is absolutely right: we
are MORE prosperous as we have more educated, active people around the
planet searching for profits. That's why poverty encompasses wristwatches
and radios instead of abject survival. At the same time, the absolute scale
of some injustices is such small change money -- and the extravagances at
the top are inexplicable. The top quintile consuming 86% while the bottom
quintile gets 1%... tough to stomach as well.

3. I was reading an advertising insert for business aviation and one of the
profiles was for Wayfarer, which was founded as the Rockefeller family's
flight department. Aside from the notion of a family airline, I find it
morally difficult to square my lust for Concordes and personal Boeing
Business Jets with societal efficiency. Yes, the numbers crunch out
correctly, but is any human's time worth $1m an hour, like bg's has at
times, or $2M/day like Milken in the 80's? Gawd, did I just say such
socialist things?!

4. It's actually that I'm completely in favor of the system aiming rewards
at those who find the arbitrage gap and take a risk (although LTCM's failure
still irks me -- should have caught those guys sooner...). I just wish we
remembered, as a society, that class was supposed to come with wealth. Not
just philanthropy, which is hitting record-low $age income (1.7% or so vs
long-term avg of 2%) in the hope that someday the boomers will retire and
really give it away. It's a sense of engagement and 'high-mindedness.' It's
the sense that the richer we are, the more paranoid we get and the more we'd
like to cut ourselves off from society and ignore it. It's easy to be
paranoid about that vision in Orange County. (Aspen Optical Shop, actually
of Aliso Viejo, just opened a store in Newport Beach with $18,000 frames)

5. But don't forget, money is worth a lot less now. Peter Drucker's essay in
the current Forbes (well worth the purchase price!) describes how he passed
up a Depression-era investment banking job to stay a professor "so I could
write" -- a $250,000 job, inflation adjusted put aside for a $2500 one. The
entire Forbes ASAP digital 100 wealth list is filled with names history will
long forget.

6. What's really lasting about these software fortunes, anyway? I always
feel like a heretic in CS whenever I argue that there's something far more
tangible about *stuff*! JimW and I have gone back and forth about how we're
all actually addicted to software, and more particularly, the myth of the
heroic infrastructure builder, when the real money is in using
infrastructure. Jim could consult on DAV tool architecture 'til the cows
come home, and I suspect it may be worth less than good marketing of Les
Concierges or any of our relationship-marketing ideas for the travel
industry. Generalizing: there is more money in business than in the business
of software, as there should be.

7. C'mon: does Microsoft in any way deserve to be the largest-capitalized
corporation in the history of mankind? Is it really more irreplacable and
permanent than GE? The whole shell game is based on future earnings: if
people really do expect monopoly profits someday, why do they expect a
monopoly to survive? I'll never have the chance to be in a similar position,
so I'll never know if I'd take my own medicine: give the money away and
establish a legacy while I still can. It's not going to be worth $58 billion
in thirty years.

8. The youngest guy on the ASAP 100 is a now-25 year old who dropped out of
Georgia Tech with ISS, internet security scanner, then shareware, now
$100M+. Yes, I envy him. Heck, I envy yang and filo, at $1B a piece and how
they kicked Excite's asses ($150M total, late entrant, friends of Cringely,
you draw the connection). Eric Brewer also came in the top half, with a grad
student in the latter half. How's that for a dream thesis? And yet, is it?
CobraBoy's major beef with us eggheads is that we'll stick our noses up to
the simples moneymaking ideas. Hotmail, for instance, at $400M, is a tricky
bit of engineering, but merely engineering.

9, But I wouldn't have built it. I was approached to be in on Juno, when I
was flabbergasted anyone would do freemail. I toyed with a web interface,
but was imprisoned by installed-base thinking: if you were *already* an
email user, you'd want to check your current primary account, right? and it
grows more complext from there. Now, today's freemail services do that, but
they began with a simpler premise: they're the user's entire mailverse.
Similarly, an ebook startup is marketing a $299 tablet that can download
books. Hardest job is keeping features out: Silicon Valley breeds features.
One they threw out was web-page caching and serving. It reminds me of a
sci-fi device in Striber & Kunetka's _Nature's End_, where off-net societal
dropouts could lobotomize their net connection with a hotwired 50GB 'rion.'
I wonder what it will be like in another few years to really just archive
away every damn last bit I ever read in a tablet like that.

10. Speaking of which, check out the HP CapShare. I want one!
http://capshare.hp.com/ . It's like digital silly putty: just swipe it,
freeform, across any matte surface and it will stitch together a page and
store up to 50. Displays thumbnails on a built-in lcd and can beam to nearby
laptops or printers. I wonder what their stitching algorithms are.. whether
they use physical inputs from scanner momentum.

11. Physics is still hard. Differential GPS, which offers the inch-by-inch
accuracy to finally free aircraft instrument approaches from ILS or MLS
corridors and straight lines, is vastly cheaper (considering the
constellation already amortized). But DGPS receivers are still $140k/plane.
Is that right, for a little software trick, with all the physical
realignment at the beacon to broadcast an error-adjustment? Gilder also gets
credit for hitting me over the head with the physical fact that crazy-high
frequencies (LMDS @ 38GHz, for example) is a *good* idea for microcells. It
won't penetrate buildings or corners well, but cubic decay functions define
small cell borders easily.

12. The surest sign of a bust is the number of players in the sillyvalley
stage that seem to be doing it for the money rather than to change the
world. Maybe I'm an old idealistic fogey on that score. But ASAP's point
about card tables is absolutely true: the winning ideas are operating on a
shoestring, not on Avedon photo shoots.

13. So if a notification infrastructure isn't the way to change the world,
what is? Start with you you want your world changed and move from there. One
good friend of mine believes it really is agents. Not the anthropomorphic
ones, but the kind enabled by structured data (XML) and a rich transport
fabric (servers connected by an OORPC of sorts). With those two ubiquitous
paradigm shifts, real delegation and mechanization may be possible. Lord
knows, I have a laundry list of them waiting.

[sidebar: I'm still looking for the 'expect' of the Web. Writing little
client-side macros is a pain, perl and all. For example, I have a
combinatorial explosion everytime I use travelocity, but want price quotes
from all three LA airports instead. Or want to quickly slurp a
table-template into a spreadsheet and summarize lowest-airfare bands by
carrier. Is there a simpler solution than XML? I think XSL could get us very
close, on the infrastructure side, but a new tool would be neat. One that
really learned from browsing patterns. I'm a heavy user, but I'd like to
think that makes me a trend-setter: >> 100 sites/day]

14. My biggest daily problem is still very mundane: a distributed data
presence. Sure, few people have multiple accounts and platforms today, but
I've got a real problem mere network filesytems can't solve. I have bits and
pieces of my message history for the last seven years scattered across 10 or
so accounts. I'd like a message switch, a depot for all that traffic. A
personal information warehouse for all my email and web pages and news. So I
keep returning to the technology I wanted in 94: an intelligent object
store. Except this time it has a protocol twist: a fun intermediate format
and protocol that slurps up all my *TP traffic. Then I can do traffic
analysis, contact management, time management, indexing, proactive agents,
on that.

15. But that's still damn technology-driven. Too much of arguing from the
desired solution-backwards. Yes, I know I can argue for it as an
intermediate step in my Munchkin vision, but I'd really hope each step along
the way is independently valuable. A personal trust manager is a similar
kind of fallacy. Nice technology, but it wouldn't quite spin -- and parts of
it are already commercially imminent.

[Not that I rescind my hope for personal-scale info-application integration.
I have too many different applications that manage the same data on
different, partially-connected platforms.]

16, "What if files weren't just bitstreams..." what an old chestnut! So I
want a web-like repository down to the desktop, with XML structured storage.
Welcome back to the Lisp Machine. Connolly had an interesting exchange with
McCarthy to that effect, and while angle-brackets are much more marketable,
I'm not going to defend them as new science.

17. Other than that, I'm kind of lost. I was once fired up to do networked
hypermedia, but that was a fad, and now the implications are already playing
out. Where's the next playgorund?

18. But *something's* gotta give if the entire planet became digital. If
computers eventually approach the scale of bacteria -- even just postage
stamps -- then *some* rules have to break, right? I don't know if we'd ever
deploy it, but why not have smart dust that clumps into formation as
whatever we desire (instant 3D lithography). Transporter beams and holodecks
oh my... but if even a sliver of that were possible, would it happen over
IP?

thinking linearly,
Rohit Khare