Wednesday, October 26, 2016


Otto, the self-driving semi-trailer company acquired by Uber, has made its first commercial delivery. Meanwhile Uber is launching its software product UberFreight, which seeks to match cargo with trucks and drivers (just as regular Uber matches drivers and fares).

I don't know if these particular efforts will be successful. Uber has the resources to deliver here, but they might be beaten by other companies. Nevertheless, these are signals of the world that is slowly emerging and I'm still trying to wrap my head around it all.

Consider also-
That's just a taste, the trend is pretty clear. Every form of vehicle is being converted to self-driving variants. Not just Google's car, but planes, ships, and cars, and new categories (such as that little grocery delivery bot, as well as small flying drones) are emerging too. And trains will too, I'm sure. Further, the regulatory and liability hurdles that are faced by UAVs (by which I mean everything from an unmanned autonomous 5lb drone to an unmanned autonomous ultra-large container vessel) with human passengers are much lessened for freight vehicles, so they will probably emerge into commercial applicability first. Commercial operators also have fewer emotional connections with their operations, so they won't hold onto human-operated vehicles out of nostalgia. As soon the cost-benefit numbers are there the transition will be swift.

It's hard to overstate how huge this is. Freight shipping defines the global economy. Anything that lowers the costs of shipping, reduces time, and/or increases volume will have knock-on effects throughout the economy as markets expand in scope and local specialization deepens. The Roman Empire was built on its roads and its control of the Mediterranean. The British Empire was built on its control of the oceans, and consequent trade via tall ship (and later steam ship) between Europe, the New World, and Asia. Container shipping is an ongoing revolution as container ships get ever larger. These changes were all inflection points in the integration of the global economy, and UAVs are going to be another one.

Tuesday, October 25, 2016

The 10,000,000 Year Ship

I've seen a lot of discussions and potential designs around how to build starships capable of reaching other star systems. Everything from huge generation ships with hundreds of crew to wafer-thin "ships on a chip" sailing on starlight. But I was thinking today about how you could send a ship to another galaxy.

Without getting all mathy, the distances between galaxies is freaking huge. The Andromeda Galaxy is 2.5 million light years away. At even 0.25c that's a 10,000,000 year trip. How could you possibly design a ship to last that long? What would it look like?

First, a couple assumptions. I'm not assuming any new physics or "magic" technologies. That's actually probably unrealistic over these time frames, but I don't care to speculate in that way. So we are talking about ships that have, at best, anti-matter acceleration and they cannot reach the relativistic speeds where time-dilation slows down the observed journey to within human lifetimes. If the EM Drive works, and/or the Alcubierre warp drive works, then we can revisit. For now I'm assuming a top speed of about 0.25c.

With that said ...

My first thought is that we can rule out sending physical human bodies. I'm not sure there's a level of cryostasis that would keep a human brain from succumbing to entropy over that timespan. So we are talking about a seedship that will grow humans once it gets there. Human DNA would also be subject to entropy, but if you have enough copies (easy to do with DNA), you should be able to peace back together a working genome at the other end.

In fact, entropy is your main enemy over these distances (more so than crashing into things - intergalactic space is pretty empty). That probably rules out any sort of active nuclear isotope battery. Anything that has a usable level of radiative energy would probably have radiated away by the time you get there. You could probably bring some Thorium-232 with you as long as it was kept shielded from neutrons. Thus you could have the parts for a thorium nuclear reactor on board, but what would you use to put that into operation? You'd need the equivalent of a battery and spark plugs to turn the engine over.

I'd say the key technology you'd need are transistors that don't age. No growing new crystals, no deterioration in the electrical properties, over 10,000,000 years. Super, super stable chemistry. Once you have that you make solar panels, batteries, and on-board compute and sensor systems.

From a design perspective you want to expose as little of the front edge of the ship as possible to oncoming H atoms you might crash into on your long trip. But you also want a lot of surface area for solar panels. I'd say the best shape is a hollow cylinder with a narrow leading edge with a stable shielding material, maybe water ice, to act as a bumper. The outside surface of the cylinder is solar panels all the way around. There could also be a central solid cylinder where all your key hardware is, connected to the outer cylinder with thin spokes.

The ship would be designed such that the light from a distant galaxy cannot power anything, but merely being within a galaxy provides enough solar power to turn on some basic senor systems. Once you pass beyond our galaxy's light the ship will lose power and go into a fully suspended operation mode, just coasting through the intergalactic space on momentum. Once you enter into a new galaxy the solar panels will start collecting solar energy again and turn on basic systems.

Now here's the real question, what kind of maneuvering ability would you have once you got there? I'm thinking two kinds.

First, we rule out anti-matter. It's possibly you use anti-matter to accelerate up to cruising speed, but I doubt you could bottle anti-matter for 10,000,000 years. More likely the bottle would fail at some point and the ship would destroy itself in deep space. No, you need a chemically and "nuclearly" inert system.

Second, you have to realize that we cannot aim for a particular star from the launch point in the Milky Way, so the computer will need to be able to wake up at the destination galaxy and pick a star, and hopefully have time to maneuver before it crashes into something. (It's probably best if you send multiple ships on slightly different trajectories)

The two best possibilities are probably storing a solar sail for the long journey and an inert gas ejected through ion thrusters. Mostly a solar sail, because it doesn't require bringing any reaction mass with you over the 2.5Mly trip. You then maneuver for a close encounter with a star and use gravity assists to slow down. Possibly more than one, making for a very long deceleration phase - but if you survived a 10My journey, a couple thousand more years here or there isn't a big deal.

One of my assumptions is that no civilization would try this until they've already mastered how to send Von Neumann probes throughout their own galaxy, so once you manage to get into a stable orbit around a destination star in your new galaxy, you'll just initiate that program out of the box. That's the easy part. It's designing a system that can coast for 10,000,000 and then wake itself up and maneuver at the destination galaxy that's hard.