The Barbell Media Diet
Read only the up-to-the-minute and the timeless — distrust everything in the middle
- Difficulty
- Easy
- Time to result
- ~ongoing to results
- Steps
- 4
- Confidence
- 85%
An information-consumption strategy that concentrates attention on two extremes — real-time signal (X) and time-tested knowledge (old books) — while treating the middle band of media (last week's news, magazines, pundit forecasts) as noise. It also prescribes going directly to practitioners over mediated coverage. Use it to allocate reading time and avoid absorbing confident-but-wrong predictions.
Origin
Andreessen describes his 'almost perfect barbell strategy' of reading X and old books, and argues the middle is untrustworthy by pointing to how badly dated predictions age — you can test this by re-reading last week's newspaper or old magazines and seeing that almost none of it played out. He adds that direct exposure to domain practitioners (via Substack, podcasts) remains dramatically underrated.
Core principles
- 01Consume either the up-to-the-minute or the timeless; be skeptical of everything in between
- 02Old newspapers and magazines are a calibration tool — they reveal how confidently the middle band predicts wrongly
- 03Practitioners explaining their own work beat mediated coverage filtered through interviews
- 04Media prizes definitive answers over open questions, which systematically rewards false confidence
How to run it
- 1
Set the two ends of the barbell
Pick a real-time source for what's happening now (Andreessen uses X) and a corpus of old, time-tested books (things written ~50 years ago that survived).
Pro tip If a book has stood the test of time, presume something timeless is in it.
- 2
Distrust the middle by default
Treat last week's news, magazines, and pundit forecasts as low-value — they make endless predictions with little information and are often out of date before publication.
- 3
Calibrate with old news
Periodically re-read an old newspaper or magazine and notice how little of what was predicted actually happened — this trains your skepticism of confident current commentary.
Pro tip Read last Friday's newspaper this Friday and check what actually came true.
- 4
Go direct to practitioners
Prioritize principals in a field explaining their own work (newsletters, podcasts) over mediated interviews, discounting a little for the fact that they 'talk their book.'
Pro tip People genuinely love explaining what they do — there's real alpha in listening to leading practitioners directly.
Watch out Adjust for self-interest: a founder explaining their work is partly selling.
In the wild
Andreessen notes nobody does this and there's no market for it, but if you re-read coverage of the internet from 1993–2010, almost all the confident predictions were badly wrong. The same holds for last week's paper.
→ A cheap, repeatable exercise that inoculates you against today's equally confident forecasts.
The Substack, newsletter, and podcast phenomenon gives direct exposure to the actual principals in a field, bypassing the old mass-media filter where everything was mediated through TV or magazine interviews.
→ Access to leading experts explaining themselves — which Andreessen calls still dramatically underrated.
Common mistakes
Living in the middle band
Consuming a steady diet of week-old news and magazine forecasts loads you with predictions that rarely come true and were often stale on arrival.
Trusting mediated over direct sources
Relying on filtered interviews instead of practitioners explaining their own work costs you the highest-signal content available today.
Is it for you?
Best for
Founders, investors, and analysts who need high-signal information and want to stop absorbing confident-but-wrong punditry
Not ideal for
People who need broad daily situational awareness for a fast-moving operational job where mid-band news genuinely matters
From the transcript
“I have like a almost a perfect barbell strategy um which is I read X and I read old books”
“Just read last week's newspaper, right? And just go back and read it and be like, "Oh my god, like none of this happened."”
“direct exposure to the people who are actually principles in the field who actually know what they're talking about is probably still dramatically underrated”
From the episode
Marc Andreessen: The real AI boom hasn’t even started yet