Unfollowing toxic social media influencers makes people less hostile! In 2 large field experiments, unfollowing a few extreme influencers: -Reduced partisan animosity by 24% -Increased satisfaction with Twitter/X -Led people to share higher-quality news Effects persisting for 6+ months and most people choose not to refollow the influencers after ... See more

The ideal user experience for a network-based product is often a moving target, based on the density of the network. If your app’s value improves with each additional user (e.g., from more content or more users interacting with you), then you should likely sacrifice the user… Show more

the best relationships come out of creating something together

How to de-risk building a social network The components that have the most engineering scope tend to be wholly reusable—such as the friendfinder & invite system. The fastest way to arrive at an app that resonates is to run rapid experiments by only changing the interaction model

Great products take off by targeting a specific life inflection point, when the urgency to solve a problem is most acute. Facebook ➝ Starting at a school Linkedin ➝ Getting your 1st job Slack ➝ Starting a company

It takes enormous discipline to NOT try to scale something that's working. But some things just don't scale -- and trying to do so can destroy the golden goose in the process. Many entrepreneurially minded people can't help themselves from trying.

Making internet hang out spaces shouldn't require scribes or high priests — & it definitely shouldn't req. mathematical prowess and/or a seed round

The natural end state of marketplaces and social media is the eventual shift from user generated supply to professionalized supply/content. They can fight it for some time, but eventually there is no other way to keep scaling (or prevent power users from peeling off)