This title problem is even worse as an author where you get one-chance for people to notice/read your book, but if the blurb or the cover picture is even slightly misleading or sub-par to the readers expectation they are likely to review it poorly and then the algorithm kicks it down the listings. I seriously miscategorised my first book and it did not do it any favors.
> DeArrow is an open source browser extension for crowdsourcing better titles and thumbnails on YouTube. The goal is to make titles accurate and reduce sensationalism. No more arrows, ridiculous faces, and no more clickbait.
> SponsorBlock is an open-source crowdsourced browser extension and open API for skipping sponsor segments in YouTube videos. Users submit when a sponsor happens from the extension, and the extension automatically skips sponsors it knows about using a privacy preserving query system. It also supports skipping other categories, such as intros, outros and reminders to subscribe, and skipping to the point with highlight.
It would be cool something like a llm based link title classifier that hide click-bait links or something like that.
And give a score based on how interesting it will likely be to you.
Say eg. some building / construction / architecture article could be titled "square shapes considered harmful". And then discuss architects known for buildings with rounded corners everywhere.
Personally, I don't see "audience likes it" as #1 priority. I prefer to make audience think, learn something, provide a new angle on something, or put out something that didn't exist before. Kind of like a movie that may not have a happy end, but viewers remember for the story, atmosphere, instant classic-potential, etc.
If I try to actually educate someone or do my research fully, either someone will know more than me, and an expert will weigh in to invalidate some section of my posting- or people will pretend to be an expert- and you’ll spend a day trying to discuss why what they’re saying is incorrect. Both will cause the discussion for other people to die.
The best has been tangents that are tangentially related to the topic presented. There can be multiple of these subthreads and they always make for interesting reading.
Personally, I love the debates that stress-test my posts, they're the most interesting part for me. If I put effort into writing something, might as well defend it, and wouldn't want any punches pulled. Oftentimes people's attempts at debunking my message end up doing quite the opposite to what I'd expect — further validating what I wrote. Other times I need to clarify something in the post.
I'm not at all sure it would cause a revolt. Most people probably wouldn't notice at this point.
https://norvig.com/21-days.html
Entitled: "Teach Yourself Programming in Ten Years"!
People who just see then click on the URL must be really disappointed when they read the actual title.
Doesn't hurt that it's a great article, too!
When I first started moderating titles, users took it personally, so I had to back off manual post-moderation and built a dozen pre-moderation filters that forced people to write proper titles. I blocked long sequences of uppercase letters, obscenities, too short and too long titles, duplicate postings, greetings, improper use of punctuation, series of exclamation and question marks and god knows what. It worked, but it drove away some users.
A few years later, I relaxed the pre-moderation filters and reintroduced post-moderation. This time, I sent automatic notifications to users as titles of their posts changed. I kept receiving complaints, I so I developed a few tricks that would reduce the number of complaints. Instead of rewriting a title, I took user text that represented the essence of the post and put it as the title, keeping the original spelling and even case, so that the user clearly sees that the titles comes from his own post.
Later on, I joined a media company and observed editors rewriting titles of journalists, detecting patterns in the changes. I followed Huffington Post’s research on A/B testing of titles and read blogs of their admirers. I even did some A/B testing on titles myself.
At some point, I even shook the dust off my machine learning skills and searched for correlations between titles and the votes on comments below.
It’s been a few years since I adopted a new approach to title moderation.
I removed titles. Entirely.
Users are presented with a generously sized textarea to write their post or comment, and the title is generated from the first few words of it. There was a bit of magic first, like skipping the greeting, but I ended up removing almost all of it. New users are confused by the apparent lack of title, but this only forces them to think through the text. Oldtimers know exactly what to expect in the title and adapt accordingly.
Personally, I would care (much) more about making a good thing over doing something many people likes.
For better or worse, my process is:
1. Write something
2. Create a title that is sometimes literal or sometimes a theme if the post covers multiple topics (I know, I know)
3. Rely on a one-sentence rss/html description to provide a clear preview of the content
