ALL INPredictions
Last updated Nov 29, 2025
Episodes / E10: Twitter & Facebook botch censorship (again), the publisher vs. distributor debate & more
Back
Chamath @ 00:17:06Ambiguous
techmarkets
If a major social network like Facebook or Twitter were to replace its algorithmic feed with a purely reverse‑chronological feed, its advertising revenue monetization on a per-page/per-impression basis would decrease by approximately 90% relative to the algorithmic-feed baseline.
if you go from an algorithmic feed to a reverse chronological feed, only I can tell you what will happen in my opinion, which is that the revenue monetization on a per page per impression basis will go off by 90%.View on YouTube
Explanation

There is no clear empirical test of Chamath’s quantitative claim, so its correctness can’t be determined from available evidence.

  • No major ad‑supported social network has actually done what the prediction describes. Since 2020, Facebook, Instagram, and X/Twitter have all kept algorithmic feeds as the default, while offering optional reverse‑chronological modes (e.g., Facebook/Instagram’s “Most Recent” / “Following,” X/Twitter’s “Following” vs. “For You”). None has permanently replaced the algorithmic feed with a purely reverse‑chronological feed for all users, which is the scenario the prediction is about. (wired.com) TikTok remains entirely algorithmic, and alternative platforms that are chronological by default (e.g., Mastodon) are not large, ad‑driven networks comparable to Facebook or Twitter. (en.wikipedia.org)

  • Internal and academic experiments compare algorithmic vs. chronological feeds but don’t publish the key metric Chamath specifies (revenue per page/impression). Twitter ran a large randomized experiment committing nearly 2 million users to a reverse‑chronological feed to study political amplification, not monetization. (arxiv.org) Meta’s 2020 election‑period experiment forced some users on Facebook and Instagram into chronological feeds for months; results published in Science and summarized by Wired show users on chronological feeds encountered more political and untrustworthy content and spent less time on Meta apps, often shifting that time to TikTok/YouTube. (wired.com) Earlier internal Facebook tests in 2014 and 2018 similarly reported lower engagement when ranking was removed, but again, not a quantified 90% drop in per‑impression revenue. (wired.com)

  • One leaked Facebook test even suggests ad impressions (and thus total ad revenue) increased when feeds were chronological‑only, because people initially scrolled more and saw more ads, according to whistleblower‑derived reporting cited by the Chamber of Progress. (progresschamber.org) That cuts against the idea of an automatic catastrophic revenue collapse, but it still doesn’t tell us whether revenue per page/per impression fell by anything close to 90%.

Because (1) no major platform has fully and durably made the switch described, and (2) the experiments that do exist don’t disclose the specific monetization metric Chamath predicts (90% per‑page/per‑impression decline), the prediction remains untestable with public data. Hence it must be classified as ambiguous, not clearly right or wrong.