The latest post that finally impelled me to ask this question:

This isn't specific to that post and, having noticed that it is a cross-post, I can appreciate that this is (at least somewhat) tricky to improve, but a lot of the titles of posts are very un-informative. Even qualifying that the above post is specific to AI or AI alignment still doesn't seem like a significant improvement, or an improvement that's 'good enough'.

I notice that I often completely ignore posts with sufficiently vague titles. Upon (very shallow) introspection, I feel like the titles are pretty literally clickbait, i.e. 'click me to discover what this is about'. That seems like behavior that warrants 'punishing' (to some degree anyways).

One potential solution (mitigation) would be to include (one of) the tags in the post title, or maybe at least in the RSS feed for posts, à la Stack Overflow, e.g.:

In the above, the 'actual' title of the question is just "navigator.clipboard is undefined" and it seems like the 'first' tag is automatically included in the page title.

New Answer
New Comment

2 Answers sorted by

ChristianKl

50

When it comes to the particular post, it's written by Paul Christiano. I expect that the target audience for the post is people who want to read everything that Paul Christiano writes on AI. 

I don't think it makes sense to conceptualize not reading the post if you are not in the target audience as punishing it. 

If you do notice a post that seems to could have a better title, how about writing the author a message suggesting a better title for it?

Yes, I noticed that Paul Christiano was the author, and immediately had a much clearer idea about what the title meant – but only after clicking-thru. The RSS feed, or maybe my feed reader (Feedly). doesn't show the author of posts.

My problem is that I regularly can't tell if I'm in the target audience just based on the titles.

I also don't think Christiano's post is the best example because it was a cross-post. I'm assuming that cross-posting to LessWrong is not particularly important to authors that do so. That's a big reason why I suggested an automatic ... (read more)

1Kenny
It does seem like the ("All Posts") RSS feed includes the author as a child element of the feed item elements, ex.: <dc:creator><![CDATA[Steven Byrnes]]></dc:creator> That doesn't seem to be a part of standard RSS tho so I imagine it'd be difficult to convince feed reader authors/maintainers to support it.

Vanilla_cabs

40

You can hover over the title to have a preview showing the first lines of the article. It's usually sufficient to dissipate ambiguity, and even ignoring uninformative titles, I use it often to get a feel of an article when I'm not sure whether it's worth reading.

Sorry! My post (question) title could have been clearer! (Oh the irony!)

This problem is particular to me scanning post titles in another app, a feed reader. I can actually see some of the first bit of the post body too (at least sometimes, e.g. when I'm using the desktop web version of the app). I don't usually 'browse' LessWrong itself; just scan new posts by title usually.

I just wish the titles themselves 'stood out' better as it's much easier to scan them, especially interspersed with items from many other feeds.

Thanks for the feedback tho! I'm definitely thinking this is more of a minor issue mostly specific to myself and how I'm following new posts here.

3areiamus
I also use Feedly and have the exact same issue.
4 comments, sorted by Click to highlight new comments since:

(I should probably have given a more informative title and will go edit it now. But definitely the main issue is that it's written for ai-alignment.com and cross-posting can drop context.)

Thanks! I'm sorry I singled you out :)

There have definitely been posts I've skipped over and only read after browsing back through older posts and noticing a high karma/comment total. There have also been a fair number of posts I start reading and wish I had skipped. More informative titles would help with both problems, but my total annoyance is low.

Thanks for the data!