Atomic Fungus is coming up on its first year of operation, in just a bit less than a month. I don't know how to cope with this. AF#1 was posted on April 3, 2006, and I can't quite remember why it was that I actually decided to have a blog (or Livejournal or WTF-ever).
Hmm, haven't yet gotten around to that review of Yawara!. Well, AnimEigo is planning--finally!--to release it sometime this year. I'm going to get to see more Yawara!
Further commentary about the upcoming anniversary will have to wait until it actually occurs.
* * *
According to the UPS web site, my computer left Albuquerque, New Mexico, at 1:13 AM Tuesday morning. That was an hour after it arrived there from El Paso, Texas. Wherever it is now, it's been in transit from there for about 26 hours, now. Right now it's around 3:20 AM on Wednesday, and scheduled delivery is Thursday afternoon.
This Internet thing that Al Gore
And we never knew when the order would arrive, either. We couldn't track it; there was no infrastructure to allow such things. So every day you'd have to check the mail and expect a disappointment.
These days just about every package being shipped by anyone has a tracking number, and most shippers have internet portals into their computer systems, allowing the hoi polloi to see where their packages are in the system.
Pretty interesting stuff.
* * *
I was thinking about the technology of Singularity a bit the other day.
The brain of any robot in the story would have to be a fiendishly complex device, something beyond anything we currently have. (Since the beginning of the story doesn't occur before around 2050, that's okay.)
The robot brain, I reasoned, is--like the human brain--a massively parallel device. So I thought: what if we started with a simple silicon substrate--a single, circular wafer 10 or so centimeters in diameter--and just built up layer after layer after layer of circuitry?
Semiconductor manufacturers have been working on "stacking" integrated circuits. This takes it to an extreme; in the end I envisioned a rough hemisphere (or maybe oblate spheroid) not more than maybe 15 cm in diameter. The interior of the thing would contain literal trillions of transistors, and it would make even the fastest personal computer we have now look like an abacus.
But the Meido-san OS is pretty sizable--it would not be excessive to call it "bloatware"--but it has to be that way because the Meido-san mimics a human being in many ways. The robot brain's computing horsepower would be necessary to run such an OS.
Isaac Asimov's "positronic brain" was also spherical--and shipped in light oil, as I recall from reading I, Robot. (The original one, not the novelization of the Will Smith movie.) Like Asimov's positronic brain, the surface of the Meido-san brain would be covered in connectors. It would fit into a precision socket, of course, and be locked in place very securely.
Cooling might be an issue. The human brain dissipates about 42 watts; since there is no theoretical minimum power required for computation I could see a similar power draw for the Meido-san brain. The human brain has a very narrow operating regime and begins to fail if its operating temperature goes more than about five degrees Farenheit over nominal; the Meido-san's brain would be a little more forgiving. But it would still need some kind of active cooling, even if it was just some circulating coolant which had a radiator elsewhere in the thing's body.
I'm thinking that the Meido-san brain is fabricated in a 30nm process fab, or maybe even 40nm--at a feature size of 40nm you can fit a lot of logic into 1.7 liters of volume. (The volume of a sphere is 4/3 times pi times the cube of the radius. I get a bit more than 1700 cc.) And the program data, the "personality", and the optional software modules don't even reside in the brain; they reside in a memory pack in the chest of the robot.
Since this would be at a time when single-plane semiconductors are built using nanotechnology, it's not that much of a stretch, IMHO. God alone knows how tiny feature sizes will be in 43 years. 43 years ago they couldn't even build microprocessors, and computers were room-sized. The way to bet, as Heinlein always said, was on technology increasing exponentially; personally I think Moore's Law will get our semiconductor technology to the point of making Meido-san brains without much difficulty long before 2050. No new science is involved; just engineering--we could probably have the first prototypes coming out of semiconductor fabs within a couple years of deciding we needed them, without inventing anything really new.
Since they've developed nanotech for semiconductors it stands to reason that nanotech would be found to be useful for other things--hence the nanotech "skin" Dan Watson uses on Alyssa.
I deliberately stayed away from doing much thinking about transportation systems in that world. Fusion has become commercially viable; Dan could conceivably have a fusion-electric car, particularly if they're capable of building fusion reactors that pump out a couple kilowatts and can fit in the chest of a robot (such as Cassandra). A typical automotive fusion reactor need only generate about 15-20 kW on a continuous basis; capacitors or batteries could be charged with excess power from the reactor (such as when the car is cruising or unusued) and dump that power when extra power (for acceleration) is needed. This is another technology we already have, in fact; only the actual power source--a compact fusion reactor--is missing.
... I could go on and on about this. Maybe sometime I'll discuss the economics of the world, but right now I'm going to go have a nap.