3 ms·
I would invest more development resources into natural language processing, for example: 1. Calculate "abusiveness score" of a tweet 2. Calculate "profanity s
by r721 10y ago
I would invest more development resources into natural language processing, for example:
1. Calculate "abusiveness score" of a tweet
2. Calculate "profanity score" of a tweet (I believe that's a different thing from 1.)
3. Allow users to set thresholds for these scores in settings
4. Use these scores for sorting replies (optionally?)
5. Google News-like clustering of news stories, and then tweets linking to them without adding original text
6. Automatic summarization of news stories
- lj3 10y ago3. Allow users to set thresholds for these scores in settings Brilliant. You could have a slider for each of those things in the setting area, along with an example 'worst scenario' tweet that would be allowed with the settings you chose. Use these scores for sorting replies (optionally?) Didn't facebook get heat for attempting to do that? I'm one of those anal retentive people who want so see everything in chronological order, but I may be in the minority there. .... news I'm not sure if automatic summarization is possible, but it shouldn't be difficult to auto-tag tweets by news story and let users filter them out of their primary timeline and into a separate list. Or just ignore them completely.