En kodare


Anders Hovmöller
GitHub twitter email

Sam Harris [was] wrong about encryption

2017-05-02

Update: in the later episode #30 Sam changes his mind based on arguments very similar to these. That’s great to hear! Hopefully this article can enlighten someone who doesn’t have the time (or hearing!) to listen to that podcast episode.

In the pre-show of his podcast episode Meat without misery Sam Harris makes the case that encryption should be breakable by law enforcement. All his arguments and reasons are valid and correct but he ends up with the wrong conclusion. This is common and in fact almost necessary for intelligent people to get wrong when not faced with the math.

But the math is the key thing here. You can’t just ignore the math and talk about encryption like it has nothing to do with math and everything to do with locks. That’s ignoring reality in favor of wishful thinking. The fact is that a safe and lock are very different from encryption in a way that makes the analogy break down completely.

Let’s fix the analogy

To make sense on why technologist are for absolute encryption we can simply take the analogy of the room that law enforcement can’t break into that you can have in your house (this is the thought experiment Sam Harris uses in the podcast). But we must modify the thought experiment to reflect the reality of the math! So let’s do this:

  1. The room is not actually in your house. A copy is available everywhere in the world.
  2. Anyone can try to break into it all the time without you knowing. When a third party opens the room you can’t tell.
  3. A second lock usable by law enforcement (a “back door”), once leaked or discovered means that all doors are now unlocked to anyone with that key. But you can’t tell. This goes also for any such rooms law enforcement has! Not to mention banks and hospitals.
  4. The lock and the room and the contents are in fact mixed up at the subatomic level so you can’t even melt it down to see which atoms were in it. Think about that: if the room is filled with nazi gold, if you melt it down you will get just a slurry or random elements like carbon, hydrogen, oxygen etc. This is the facts of how the math works out. I think these alterations to the analogy seems to you to be so radical as to invalidating the rest of the arguments that follow from the original analogy.

There are of course other problems with the thinking behind Sam Harris’ stance like the fact that the genie can’t be put back in the bottle but I think the above is quite enough to convince you to at least think hard about encryption and not be swayed by arguments to emotions for victims of crimes.

If we break encryption, we will all be victims, albeit to robbery, blackmail and embarrassment and not murder. But victims we will be. Count on it.

« When you assume, Gruber edition Keyword argument confusion in Python »