This looks useful, someone let me know if they find an equivalent for Android!

If we plot the two together, we don’t see that much of a correlation

...That looks like a fairly strong correlation to me? Maybe I'm reading this graph wrong, but only the data point for 2018 looks substantially different.

I use Signal with anyone else who's willing, and I support advocating switching from Messenger to Signal, but I think this post is pretty poorly written.

The first thing is that the tone is far too normative. You are telling the reader what to do, and you are telling them that you are right about it being better. This is unpleasant to read and is a poor way to convince people to do something.

I don't think you made an honest attempt to check for reasons why people prefer Messenger. I have many friends who used Signal with me over several months, but eventually went back to using Messenger. There are reasons why they did this.

I also find a bunch of your arguments unconvincing. For example;

But this effect is truly marginal; we are talking about a difference of at most several seconds per connection

This is something that feels like it would be true, but if you have studied user acquisition or consumer behavior or whatever, then you start to see how the tiniest and most trivial feeling step can hugely reduce how many people actually sign up. It's not about how long it nominally takes. It's about how many people will actually bother or decide to. Everyone is already on facebook, and therefore they're already on Messenger, and you also already know it's them.

In my own experience, the problem pointed out by Habryka is the biggest one. Even if there is some way to get message history to transfer, it is a lot of effort to figure out what that method is and then do it.

if you grind up the universe and pass it through the finest sieve

igotthatreference.jpg

That diamond/carbon scenario is an excellent concrete example of the ontology problem.

What are some of your techniques for doing good research?

What is your stance on whole brain emulation as a path to a positive singularity?

What are your biggest flaws, skill gaps, areas to grow?

How does MIRI plan to interface with important AI researchers that disagree with key pieces in the argument for safety?

What are some of the most neglected sub-tasks of reducing existential risk? That is, what is no one working on which someone really, really should be?