Why the New York Times told reporters to stop using Twitter

Executive overview

Twitter became quasi-mandatory for New York Times journalists around 2014–2015. By 2022 the Times reversed course, citing harassment, distraction, and editorial echo chambers. A fourth, unspoken reason: reclaiming power from individual reporter brands.

Constant Twitter use degrades journalism quality, psychological health, and editorial independence — and the Times knew it before it admitted it.

The three official reasons the Times cited

• Harassment: Twitter exposes reporters to psychologically damaging attacks, creating chronic anxiety and stress
• Distraction: heavy emotional engagement on the platform degrades writing quality
• Echo chamber effect: reporters optimise for Twitter approval, not for readers or the country at large
• A tiny, non-representative slice of users — ~15 million active US accounts — was driving editorial decisions at the most-read paper in America

The unspoken fourth reason: brand power shifting from paper to reporter

• Social-media-first reporters build personal audiences independent of their employer
• The Times benefits when reporters prioritise the paper over personal brand-building
• Barry Weiss's Substack success after leaving the Times illustrated the risk: reporters with large Twitter followings can leave and take their audience
• The Taylor Lawrence vs. Maggie Haberman dispute embodied this tension: individual brand vs. institutional reporting

Cal's 2016 prediction and what he got right (and wrong)

• Predicted in Deep Work that pressuring reporters onto Twitter would backfire — focused mainly on distraction
• Underestimated the psychological harm of routine Twitter interaction (not just targeted harassment)
• Did not anticipate the editorial influence problem that Barry Weiss's resignation letter crystallised: "Twitter is not on the masthead of the New York Times, but it might as well be"

When the hyperactive hive mind works — and when it doesn't

• Hyperactive hive mind: Cal's term for unscheduled, on-demand back-and-forth messaging as the default coordination method
• Works for small groups (three or fewer people) where flexibility and speed matter more than focus
• Fails at scale: constant monitoring required, context-shifting becomes the work, cognitive output collapses
• Naturalistic basis: small-group, ad hoc coordination predates written language — Palaeolithic hunters used it
• Ant and bee colony behaviour appears efficient but required millions of years of evolutionary optimisation; randomly deploying "just use email" is the equivalent of inserting an arbitrary algorithm into an ant brain

Managing a heavy workload with frequent context-switching

• Work on one thing at a time to a stopping point — no email or Slack during focused blocks
• Busy weeks mean less inbox time; that will annoy people, and that is acceptable
• When a new project lands, map the full execution process upfront: scheduled check-ins, handoff points, feedback windows
• Reduce dependence on the hive mind during busy periods — the ideal is executing against a clear plan with no reactive messaging
• Most people calibrate acceptable workload to best-case scenarios; calibrate instead to expected or worst-case loads, then push back sooner

Transitioning from a side passion to a viable career

• Use money as a neutral signal: earn from the passion first, then make structural decisions based on real data
• Do not write one hour a day for three years — increase to 15–20 hours a week to compress the feedback loop
• Take a two-week leave of absence to push a near-finished project over the line
• Part-time consulting work (e.g., fractional CFO) can bridge income while writing time scales up
• Avoid quitting entirely before data exists; manageable risk (leave of absence, paused promotion activities) is the right bet size

Teaching vs. scholarship: reducing time without reducing quality

• A SUNY Stony Brook study found reduced class-prep time can improve student evaluations
• Reuse well-crafted lectures with a 10–20% improvement rule each iteration rather than rewriting from scratch
• Design assignments that assess material effectively and are efficient to grade
• Students do not need constant availability — they need extreme clarity about when and how to reach you
• Over-preparation often signals uncertainty; knowing your field deeply and presenting it clearly is what students respond to

On complexity and decision-making

• For personal decisions: pick a couple of satisfying criteria, decide, move on — do not optimise infinitely
• For intellectual complexity: marinate in the material, apply a simplifying framework, then trust intuition
• "Grit in the gears" is a reliable signal that a framework needs revision
• Taste for good arguments is trained by exposure to strong writing and thinking, not by exhaustive rationalist analysis