... and remove the login prompt, which admittedly is one of the worst features of Twitter right now. Perhaps I'm letting my cynical professionalism cloud my judgment, but it seems to me that one of these problems takes about 4 hours to fix, while the other would be a 3-12 month research problem.
That's a neat way to guarantee partial success, though!
I was wondering, he did a poll on twitter though that suggested the most mentioned problems with search may be fixed just by adding some specific filters/subpages, which does beg the question though, what is broken about Twitter's search?
"Searching within "liked" and "seen" tweets with a modifier should be table stakes.
‣ The modifiers and advanced search should be surfaced a lot better. Discord does this well.
‣ Move away from exact text search and toward embeddings."
User-approval isn't the only metric by which features are judged. No one liking it was obvious as there's no user benefit. If it results in increased revenue/recurring users/valuable ad displays, that's different.
This is a non-dismissal login prompt. I doubt the benefit of user logging in and signing up is high.
It is a straight up bad UX and not user first, and the owning team defends it because they don't rely on common sense.
Of course, if you put login prompt anywhere, you would get at least some logins or signups (at least 1).
Twitter had this issue before where they didn't allow non-logged in users to browse twitter home at all by showing a non-dismissable login prompt. It took Jack Dorsey to come back and change that in 2016.
But yeah sure the non-dimissable login prompt on the search page is great. Hahaha
Initially I thought hiring this guy would be mostly hot air (e.g. spewing out grand vision non sense). He immediately found a concrete high-impact and low-hanging-fruit problem to solve on his first week.
This login prompt is universally hated by everyone. You'd have to wonder why twitter never thought or be bold enough to remove it.
That's a neat way to guarantee partial success, though!