Going dark

The NSA is worried about civilian communications “going dark”. Personally I’m hopeful they will, since I think the government will continue to invade our privacy by any means available to them, 4th Amendment be damned.

But what does it mean to “go dark”? The NSA seems to simply mean “secure encryption”, where the communication is encrypted end-to-end. The NSA has the data, but they cannot read it unless they devote meaningful resources to cracking it. This sort of “going dark” would prevent widespread data collection, but it wouldn’t prevent the NSA from listening in on any one particular conversation.

I think of “going dark” in broader terms. It applies to more than just communication, and it’s a lot more than just encryption. I think of “going dark” as matter of increasing privacy and avoiding centrality on both software and physical layers. Darkness is a matter of degree.
Let me give you a few examples.

In the communications realm, consider a simple text-message program that allows for coms between any two mobile devices somewhere in the world; there’s a number of different ways to increase privacy and avoid centrality. Here’s some of them –

  • End to end encryption. This increases privacy. And of course there’s many sub-concerns within the realm of encryption, like perfect forward secrecy or using a trustworthy random number generator.
  • Decentralized protocol. Like Bittorrent’s new chat program, there’s no central server, just peers. This decreases centrality, as there’s no one server or cloud company for the Feds to co-opt.

  • Onion routing. This increases privacy on the software layer by hiding metadata. The contents of packets are encrypted, but so is the path. Any one server only knows who to pass a packet to, not its source or final destination.

  • Forgetfulness. This is important, and often overlooked. Data should only be retained as long as necessary to serve the end-user’s intended purpose, then discarded. Worrying about today’s tyrants is bad enough; worrying about what tomorrow’s tyrants might do with what you said five years ago will make your hair grey.

  • Mesh networking. This technology avoids centrality at the physical layer. At least in theory a sufficiently large mesh network would replace the telecom companies like Verizon and Comcast entirely. However, you don’t need to a global mesh network – even local mesh networks that also employ encryption, onion routing, and forgetfulness are useful to hide within once data reaches the (presumably co-opted) main backbones.

  • Locality. Protocols that avoid backbones except when necessary would decrease centrality. Local communications remain local, avoiding compromised central nodes.

And each of the above can be employed either alone, or together. There’s synergy to adding more kinds of privacy on top of one another. A decentralized protocol using end-to-end onion encryption over a forgetful community mesh-network would be “dark” far beyond anything contemplated by an “off the record” Google Hangout.

Now let me repeat – this isn’t just about communications. “Going dark” is something that can be applied to commerce, finance, and even personal travel to some extent. One of the hot topics of 2013 is BitCoin, but by the standards outlined above, BitCoin isn’t very dark at all. It’s decentralized, but it’s not encrypted, forgetful, local, or mesh. Forked coin projects like Anoncoin and Namecoin address these issues, and I expect one of them to become more successful than BitCoin in the long run.

But also consider how “going dark” could apply to commerce. Right now much of American commerce in physical goods goes through just a couple main central networks – FedEx, UPS, and the Postal Service. The technology is coming soon that will allow the total disintermediation of FedEx, and allow much of boxed commerce to “go dark”. I will expand on that thought in another post.

For now, I will leave you with the following thought exercise: Consider as many aspects of your life as possible, and ask if there’s benefit to them going dark, and if so, just how dark you could make it.



Drones are in the news with Amazon’s announcement about its intention to use mini-copters to directly deliver products to customers by air within 30 minutes of placing the order. Ambitious. But drones are bigger, and far more revolutionary than getting a can of shaving cream on the double. Let’s just take a look at the current landscape.

My working assumptions are that every mode of transport currently existing will have drone versions within 5-10 years, but the comparative advantage of one mode over the other will not materially change. E.g., fixed wing aircraft will remain superior to helicopters in most situations.


Of course the mini-copters are drones. Full-size choppers could be converted to drones as well. Helicopters have very high maintenance and fuel costs though which are not related to occupant safety. It’s the engines. Putting an A.I. at the helm will not improve these.

(What would improve these is a dramatic increase in battery energy density, allowing electric helicopters and engines. That would be a lot cheaper. But that’s beyond the scope of this article).

My assumption is that drone helicopters will remain a niche product, just as real ones are. The primary benefit will be “eyes in the sky”, like a news chopper, for the common man. But Amazon’s delivery service will remain a novelty.

This gets interesting, but not for people. Passengers would probably get freaked out at the thought of no pilot being at the helm. But for freight, drone cargo planes are the near term future.


I consider the Google self-driving cars to be drones. Anything that moves on wheels today will be a drone-car tomorrow. Taxis, delivery vans, busses.

You think I jest? Behold the Narco Sub. Criminal elements are frequently on the cutting edge of technology. Notably, the US Coast Guard estimates they successfully intercept only 1 in 10 narcosubs.


I’m not immediately familiar with any drone boat plans, but this seems to be inevitable if we already have drone submarines. I would guess that drone boats are less intersting for commercial shipping as (1) the cost of crew is small relative to the cost of fuel on a super-carrier, and (2) a slow-moving and totally unmanned craft would be a tempting target for piracy. (Assuming that merchant shippers aren’t allowed remote controlled gun platforms)



The future is Uber, only more so. Decentralized, Point to point, and private. FedEx and UPS may continue to exist, but it won’t be duopoloy (with the USPS vestigal appendage) that we have today. Most of the market will probably be “independent”.

Imagine this – a company builds widgets in Utah, in an unmarked factory at the end of an umremarkable road. Maybe they get deliveries by FedEx, but not necessarily. They could own (or lease) their own drone-trucks for picking up supplies and delivering finished goods. Running low on feedstock A? Send a truck to get some more from the supplier.

There are ways to make the above model more efficient. You could have companies whose only job is to run trucks, and they keep fleets of them in reserve. You could move to a standard container system, where everything has to fit into one or two ISO Standard 6780 pallets, and drone trucks with drone forklifts come by every morning like the garbage man for pick-up and drop-off.

But the above isn’t necessary. If control, privacy, or other special needs take precendence, each producer could own (or lease) its own fleet of trucks. Nondescript, they’ll drive down the highways in wagon-chains (to conserve energy) with no indication of what they carry.

(Unless the Feds demand that every drone carry a transponder that announces their location, destination and cargo, and subjects them to spot checks to enforce honesty. But that sounds expensive and intrusive.)

The same goes for airplanes. Right now FedEx owns the largest fleet of planes in the world, but what about tramp cargo drones for hire? Send your truck to the nearest local airport for pick-up or drop-off. Once the planes get out over the ocean they can use wing-in-ground effect to approach the cost and efficiency of surface transport.

Within 10 years I expect cargo transport to be just as decentralized as the Internet is today. FedEx and UPS may exist, just like Google and Facebook do, but the majority of traffic will be outside the main channels.

Nothing else changes

There was an article I read in Smithsonian magazine years ago. I tried to find an online link to it, but Google failed me. This is my best recollection.

Many decades ago some fine British gentlemen were noodling around with shovels in the deserts of the Middle East. While there they found a tablet of obvious very ancient origin. The writing was completely illegible though, so they boxed it up and sent it back to the British Museum in London where it was promptly shelved and forgotten about.

Some time later radiocarbon dating (or one of its similar technologies) was discovered, so the British Museum started going through its collection to date things. This particular tablet stood out as incredibly ancient. If it wasn’t the very oldest piece of writing we had, it was up there. But the text was still illegible, so they shelved it again.

Then in the 1990s digital cameras started being a thing, and a fellow had this idea that if you took enough pictures of an object from different angles using high-contrast flash photography, perhaps a computer could analyze all the pictures tease out faded carvings and writings. This experiment was successful, so the tablet was brought out again and processed. The text was a very ancient dialect of cuneiform, and a copy of the text was sent up to the linguists for translation.

Some time later, however long the translation took, a report was released on what the tablet said. Many people were very interested, as this was the oldest writing example in the Museum’s collection by a large margin. What could we learn from the past so far removed from our current era?

As I recall the tone of the article, many scientists were disappointed by the content, but I am not. I am fascinated. The tablet was a receipt for the sale of 20 male slaves, with a money-back guarantee if the purchaser was not absolutely satisfied with the purchase.

This is the lesson I take from a writing from 10,000 years in the past – knowledge increases and technology improves, but nothing else changes.