What are the key takeaways from this Land of the Giants episode?

Key insights include: **Platform power asymmetry:** Twitter maintained separate moderation rules for Donald Trump throughout his presidency, with Jack Dorsey personally making final decisions on Trump's account while a small team of roughly five policy specialists manually reviewed all other high-profile cases. This two-tier system created internal resentment and delayed action on tweets that safety staff flagged as inciting violence months before January 6.; **Misinformation labeling over removal:** Twitter's 2019 deep fake policy established a tiered response framework — algorithmic demotion, contextual labels linking to verified sources, or outright removal — rather than defaulting to deletion. This model was then applied to COVID-19 misinformation in May 2020, but proved difficult to scale because official health guidance from the CDC and WHO changed continuously throughout the pandemic.; **Retaliation as a moderation deterrent:** When Twitter first labeled a Trump tweet about mail-in ballot fraud, mid-level trust and safety employee Yoel Roth was publicly named by Kellyanne Conway on Fox News, triggering personal threats. Twitter's leadership recognized that content moderation actions against Trump would produce targeted harassment campaigns against individual employees, which factored into future enforcement decisions.

What did Deray Mckesson and Jack Dorsey discuss on Land of the Giants?

This episode of Land of the Giants traces how Twitter transformed from a protest organizing tool during Ferguson in 2014 into a platform paralyzed by Donald Trump's rule-breaking tweets, culminating in Twitter's permanent suspension of Trump following the January 6, 2021 Capitol insurrection and years of escalating content moderation failures. Key topics include: **Platform power asymmetry:** Twitter maintained separate moderation rules for Donald Trump throughout his presidency, with Jack Dorsey personally making final decisions on Trump's account while a small team of roughly five policy specialists manually reviewed all other high-profile cases. This two-tier system created internal resentment and delayed action on tweets that safety staff flagged as inciting violence months before January 6.; **Misinformation labeling over removal:** Twitter's 2019 deep fake policy established a tiered response framework — algorithmic demotion, contextual labels linking to verified sources, or outright removal — rather than defaulting to deletion. This model was then applied to COVID-19 misinformation in May 2020, but proved difficult to scale because official health guidance from the CDC and WHO changed continuously throughout the pandemic..

How long is this episode of Land of the Giants?

This episode is 49 minutes long. SignalCast provides an AI-generated summary so you can get the key insights in about 3 minutes.

Land of the Giants

Havoc On the Platform, and Off

November 8, 2023

49 min episode · 2 min read

Deray Mckesson,Jack Dorsey,Maggie Haberman

Episode

49 min

Read time

2 min

Topics

Health & Wellness, Leadership, Software Development

AI-Generated Summary

Published Jul 31, 2026

Key Takeaways

✓Platform power asymmetry: Twitter maintained separate moderation rules for Donald Trump throughout his presidency, with Jack Dorsey personally making final decisions on Trump's account while a small team of roughly five policy specialists manually reviewed all other high-profile cases. This two-tier system created internal resentment and delayed action on tweets that safety staff flagged as inciting violence months before January 6.
✓Misinformation labeling over removal: Twitter's 2019 deep fake policy established a tiered response framework — algorithmic demotion, contextual labels linking to verified sources, or outright removal — rather than defaulting to deletion. This model was then applied to COVID-19 misinformation in May 2020, but proved difficult to scale because official health guidance from the CDC and WHO changed continuously throughout the pandemic.
✓Retaliation as a moderation deterrent: When Twitter first labeled a Trump tweet about mail-in ballot fraud, mid-level trust and safety employee Yoel Roth was publicly named by Kellyanne Conway on Fox News, triggering personal threats. Twitter's leadership recognized that content moderation actions against Trump would produce targeted harassment campaigns against individual employees, which factored into future enforcement decisions.
✓Russian interference panic shaped overreach: Twitter's October 2020 decision to block sharing of the New York Post's Hunter Biden laptop story — later acknowledged internally as a mistake — stemmed directly from the 2016 election trauma, when 36,000 Russian bots posted 1.4 million election-related tweets in the three months before the election. Fear of repeating that failure caused Twitter to misapply its hacked materials policy.
✓Fringe-to-mainstream radicalization timeline: Calls for violent government overthrow on Twitter began as fringe content in early 2020 around COVID mask mandates, then became mainstream by the September 2020 presidential debates. Safety policy staff tracked this escalation and flagged Trump's December 19 tweet about January 6 as a specific time, date, and location signal that high-risk accounts immediately amplified as a mobilization call.

What It Covers

This episode of Land of the Giants traces how Twitter transformed from a protest organizing tool during Ferguson in 2014 into a platform paralyzed by Donald Trump's rule-breaking tweets, culminating in Twitter's permanent suspension of Trump following the January 6, 2021 Capitol insurrection and years of escalating content moderation failures.

Key Questions Answered

•Platform power asymmetry: Twitter maintained separate moderation rules for Donald Trump throughout his presidency, with Jack Dorsey personally making final decisions on Trump's account while a small team of roughly five policy specialists manually reviewed all other high-profile cases. This two-tier system created internal resentment and delayed action on tweets that safety staff flagged as inciting violence months before January 6.
•Misinformation labeling over removal: Twitter's 2019 deep fake policy established a tiered response framework — algorithmic demotion, contextual labels linking to verified sources, or outright removal — rather than defaulting to deletion. This model was then applied to COVID-19 misinformation in May 2020, but proved difficult to scale because official health guidance from the CDC and WHO changed continuously throughout the pandemic.
•Retaliation as a moderation deterrent: When Twitter first labeled a Trump tweet about mail-in ballot fraud, mid-level trust and safety employee Yoel Roth was publicly named by Kellyanne Conway on Fox News, triggering personal threats. Twitter's leadership recognized that content moderation actions against Trump would produce targeted harassment campaigns against individual employees, which factored into future enforcement decisions.
•Russian interference panic shaped overreach: Twitter's October 2020 decision to block sharing of the New York Post's Hunter Biden laptop story — later acknowledged internally as a mistake — stemmed directly from the 2016 election trauma, when 36,000 Russian bots posted 1.4 million election-related tweets in the three months before the election. Fear of repeating that failure caused Twitter to misapply its hacked materials policy.
•Fringe-to-mainstream radicalization timeline: Calls for violent government overthrow on Twitter began as fringe content in early 2020 around COVID mask mandates, then became mainstream by the September 2020 presidential debates. Safety policy staff tracked this escalation and flagged Trump's December 19 tweet about January 6 as a specific time, date, and location signal that high-risk accounts immediately amplified as a mobilization call.

Notable Moment

A Twitter safety policy staffer described manually searching Twitter dot com for phrases like "execute Mike Pence" during the January 6 riot and deleting tweets one by one in real time, with no automated tools involved — a stark illustration of how unprepared the platform's infrastructure was for the crisis.

Know someone who'd find this useful?