Product design question

How would you improve Twitter?

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Product sense for a public conversation platform, finding a real pain point beyond generic feature requests.

Pick a segment, for example casual readers who follow topics versus active posters who reply and build threads.
Identify a pain point for readers: low quality replies and spam crowd out substantive conversation on popular tweets.
Prioritize reply quality over new posting features, since a noisy reply section is the most cited complaint about the platform.
Propose surfacing 'best replies' ranked by a mix of author engagement and community signals, shown before chronological replies.
Consider abuse: pair this with better detection of coordinated low quality replies to protect the ranking's integrity.
Define success as time spent reading replies and reduction in users reporting reply spam.

Targets a specific, widely cited pain point, low quality reply threads, instead of a scattershot feature list.
Explains the ranking mechanism concretely, engagement plus community signals, rather than saying 'better algorithm'.
Gives an illustrative number, e.g. assumes 30 percent of top tweet replies are currently low value or spam.
Considers the abuse vector, since any ranking change creates new incentives for spam actors to game it.

Proposing unrelated new features like editing tweets instead of a focused, high impact improvement.
Ignoring how bad actors would try to game a new ranking system.
No success metric tied to the actual pain point being solved.

Chapters of the AI PM course that teach what this question tests.