In an interview, Twitter CEO Jack Dorsey says Twitter should be built around following topics and interests, thinks an impartial product can exist
@jack to @cwarzel: I wasn't talking with @seanhannity, I was “talking with the people that listen to him” https://www.buzzfeednews.com/ ... pic.twitter.com/lENDP1Ssmy Charlie Warzel / @cwarzel : New: I sat down with @jack and asked about a whole bunch of things including harassment, deleting old tweets, whether there's a line Trump can cross to get banned, why he spoke with Sean Hannity, and if an “unbiased” Twitter can even exist. https://www.buzzfeednews.com/ ... See also Mediagazer
Context & Ripple Effects
This interview lands in the middle of a bruising fortnight for Jack Dorsey: days earlier he defended Twitter on Sean Hannity's radio show for not acting against Alex Jones, and reports surfaced of an internal executive meeting over Infowars where 'dehumanizing speech' and new policies were on the table. Accusations of political 'shadow banning' were circulating, and the Hannity appearance was read as outreach to the right.
Against that backdrop, Dorsey's two claims here are strategic: that Twitter's future is following topics and interests rather than people, and that an impartial product is possible at all. It extends the arc of his recurring sit-downs — including a 2017 interview on harassment and live video — in which he frames structural product change as the answer to problems others frame as policy failures.
First-order effects
- Dorsey explicitly reframes his Hannity appearance as talking with 'the people that listen to him,' not the host — a direct attempt to defuse the perception that Twitter was courting conservative audiences while it weighed action on Alex Jones.
- The topic-and-interest framing signals a product direction away from the follower graph, which would reshape what users see and how new users onboard without building a network first.
Second-order effects
- An interest-graph Twitter changes the moderation surface: enforcement and curation decisions get made at the level of topics and algorithms rather than individual accounts, raising the stakes in exactly the 'shadow ban' disputes Dorsey was denying.
- Rivals and advertisers watch whether 'impartiality' becomes a marketed feature — if Twitter claims neutrality as design, competitors face pressure to articulate their own curation philosophy or be cast as biased by default.
Third-order effects
- If the pattern holds, platform neutrality stops being argued as a content-policy question and gets settled inside ranking and recommendation systems — making algorithmic design choices, not takedown decisions, the battleground for bias accusations.
- A topic-first structure also points toward durable, interest-based distribution that is less dependent on celebrity accounts, altering how reach, virality, and advertiser value accrue across the platform.
The trend: Social platforms are responding to political-bias accusations by relocating neutrality from moderation policy into product architecture, with the interest graph replacing the follower graph as the organizing principle.