Joshua Gans

I have had my iPhone Duo for a week and I can say that I am very happy with it. It works for me, but I can see how it may not work for everyone.

Some context: I am not a person who lives on their phone because I mainly live on my computer. The phone is what I use when I am away from my computer. A single screen is useful for some things but not for others.

Second, I am getting older and frankly like to have the option of seeing more than a single-screen phone can give.

For those reasons, having an easy-to-carry iPad mini in my pocket is very, very useful. This is especially the case when I want to continue using Claude or Codex away from my computer. But also just great for reading.

So, you may ask, what about that small front phone screen? I was worried about that, but it turns out I just love it. It gives more focus than I knew I wanted. I just see information I need to interact with and do that. It sits easily in my hand. Basically, it is what I wanted the Apple Watch to be but it was too small and limited. It turns out that most of the time, a small screen phone is what I need. Now I have that with a larger option. Fully hedged experience.

What about stuff like touch ID? I would prefer Face ID -- it will surely come but it is clear that Apple have to put in two FaceID cameras which is technically more challenging. They will get there. The iPhone Pro 18 has an under the screen FaceID camera so they are close. However, touch ID has improved a ton from 10 years ago when we had to rely on it. I barely notice it. I have three fingers memorised for all the ways I pick the phone up plus my watch handles most opening (that could be more reliable though).

What about the crease? I don't notice it. The nano texture layout is so nice there just isn't an issue.

What about battery life? After the first day when it indexed for Siri, it has been great. At 50% by the end of the day.

What about app layouts? That can be improved but that is a patience game. What I like is that some apps are taking advantage of the fold itself which is opening up possibilities. I have switched to Flit from Reeder for my RSS feeds for example.

This is an expensive phone. I think by next year it will be undeniably worth it. But I am very happy with it right now.

This is really quite good.

Somehow I got ratioed for this and only one commenter actually said the true thing.

How is that mathematically possible?

I was thinking about the polls and elections (and to be clear pol sci folks may regard what I am about to write as obvious) and wonder if the issue is that polls are precisely bad at the one thing they are supposed to be able to do: forecast the behaviour of swing voters.

Here is my simple theory. In the US at least, election outcomes are determined by (a) turnout of non-swing voters; (b) turnout of swing voters and (c) how swing voters will vote.

A little bit of choice theory, whereby you assume people vote if they think their vote matters, tells you that swing voters are more likely to turn out than non-swing voters and non-swing voters when the election is close are more likely to turn out for either party.

That means that in elections where there are indications that swing voters are going one way, turnout will be lower amongst non-swing voters (but that has a neutral effect) and so the election outcome is over determined by swing voters.

Thus, if you don't adjust your interpretation of the polls as a forecast for this effect you end up getting situations where polls underpredict the extent of the swing voter effect and so underpredict the performance of the winning party.

You don't need any shy voter theory (which when you think about doesn't make sense for pollsters even if it makes sense for social situations) or any other long-term bias.