That's why splash pages (and writeups, like the original) are important. But as mentioned in my previous comment, those splash pages/writeups also have to be effective advertisements.
without any explanation of why I should use your app,
or even how I'm supposed to use your app, I would
close it immediately.
Seeing you write this on HN, which has no splash screen either, contradicts this statement.It wasn't until many months later, after experience with other sites with similar interfaces, that I actually started using HN because I then understood how and why to use it. The HN interface is intuitive enough and similar enough to other website to understand that you're being shown comments and a reply box when you click on a link.
That's not the case with this [1]. If you linked me directly to a page that looked like this, without any explanation of what these boxes of text are telling me, where they are coming from, or what they are, I would not use it. The button next to them (a grey box with a plus symbol on it) doesn't explain to me at all what it does, and is not something you see enough in other apps to make a good enough guess as to what it does. After reading the blog post, my assumption is that clicking that button is supposed to send me an email, I think? But I would have absolutely no way of knowing that if you just linked me directly to the app.
1: https://hackernewslater.com/images/hacker_news_later_screen_...
edit: if you download the actual app, the first page it shows you actually is a splash screen that gives a brief explanation of what the app does and why. If that same splash screen content had just been shown and linked to at the top of the original writeup, I suspect he would've gotten a lot more installs.
The button next to them (a grey box with a plus symbol
on it) doesn't explain to me at all what it does
If so, maybe they should use a 'read later' button instead. Or put a sentence on top 'Click + to read a story later'. The more people bounce the stronger the inclination to try that and see if engagement increases. That's why all 17000 visitors would be valuable testers, no matter how they behave.
I would expect over 50% of users to take that first step. That is what I saw in similar scenarios.