Alphonzo Terrell was inside Twitter when Elon Musk bought it for $44 billion and proceeded to lay off roughly 80% of staff. Now he's building Spill, a social platform whose stated organizing principle is trust. 'If trust breaks, your business breaks,' he tells Fast Company. This week, the platform he left behind announced it was replacing its 'misaligned' revenue sharing program with something called Original Content Rewards. The comparison is not flattering to X, which is essentially admitting its incentive structure actively rewarded the wrong behavior, then replacing it with a different metric it hasn't proven is better.
Incentive Design as Platform Philosophy
The original X revenue-sharing model paid creators based on impressions generated by Premium subscribers. This reliably rewarded outrage, controversy, and engagement bait, since those are the posts that Premium subscribers respond to most consistently. The new 'Original Content Rewards' system claims to favor originality, but without disclosing the specific signals it uses, it is impossible to evaluate whether it will avoid the same failure mode. This is not a technical problem. It is a philosophical one. Revenue sharing systems encode a theory of what a platform values. X's system said it valued reach. Terrell's insight at Spill is that a platform encoding trust as its primary value requires different architecture from the ground up, not a renamed reward program layered on existing infrastructure.
The Platform-Scale Theory of Trust
A 2020 paper in the Journal of Communication by Bucher found that algorithmic systems on social platforms produce what she calls 'anticipatory compliance': users modify their behavior not in response to explicit rules but in response to their model of what the algorithm rewards. This means that changing an incentive label without changing the underlying signal is worse than useless. It teaches users to game the new label while keeping the old behavior. Terrell's bet is that building trust as a structural feature, not an aspiration, requires building a community that is small enough to be coherent before it is large enough to be profitable. That is the opposite of every major platform's growth logic. As Connor Hayes, VP of Threads, has described, building at scale while maintaining a deeply personal feed is the hardest unsolved problem in social product design. Spill's answer is to defer scale entirely. It is a thesis about what trust actually costs.