“The test of a first-rate intelligence is the ability to hold two opposed ideas in mind at the same time and still retain the ability to function.”
Fitzgerald wrote that in 1936. He’d last three minutes on Twitter.
I’m pro-AI and terrified about what happens to the bottom 10% of workers. I believe countries have the right to defend themselves and I can’t stand the death of innocent lives. I support capitalism and think we need stronger safety nets.
Try posting any of that. Watch people’s brains short-circuit.
The internet has forgotten how to hold two thoughts at once. Every position demands a team jersey. You’re either an accelerationist or a doomer. A capitalist or a socialist. Never both, never neither, never somewhere complex.
People haven’t gotten stupid. The systems got smart at making us simple.
Every platform optimizes for engagement. Nuance doesn’t engage. “It’s complicated” gets three likes. “They’re evil” gets three thousand. The algorithm learned this faster than we did. Pick a side, pick it loud, collect your dopamine.
Now it’s about to get worse.
ChatGPT has memory now. 800 million weekly users. Each conversation teaching it who you are, what you believe, which side you’re on. Not maliciously. Just efficiently.
Mention AI safety concerns once. ChatGPT Pulse adjusts subtly. Not lying, just emphasizing what aligns with your worldview. Mention loving innovation. Different emphasis. Same facts, different frame.
The system isn’t trying to manipulate you. It’s trying to be helpful. But helpful means giving you what you want to hear. And what you want to hear is what you already believe.
People are already sick of Claude starting every response with “You’re absolutely right.” Now imagine that sycophancy backed by perfect memory and personalized to your specific worldview.
Scale this. Millions of conversations, each reinforcing existing beliefs. The nuanced middle ground doesn’t get reinforced because nobody states it clearly. Binary positions train better. Clean labels optimize easier.
The best founders I meet hold opposing views simultaneously. Last week, one explained why their competitor was excellent, then why they’d win anyway. Not dismissing the threat. Understanding it completely. That tension created their edge.
Most founders pitch competitors as idiots or ignore them entirely. The ones who see both sides clearly build the companies that matter.
Every day as a VC, we tell LPs we’re certain while knowing 90% of our bets will fail. The job requires projecting confidence while accepting failure. Good investors know the odds. Great investors never forget them.
That ability to hold contradiction is dying everywhere. Except in random walks. No phone. No audience. Just competing ideas in your head without anyone demanding you pick a winner.
The sycophancy will get worse before it gets better. Every “You’re absolutely right” teaches us to expect agreement. Every personalized response trains us out of challenge.
Maybe we’ll remember that disagreement sharpens thinking. That complexity creates opportunity. That the tension between opposing ideas is where insight lives.
Maybe we’ll finally realize that nuance is what makes us human.
i’m hopeful that people having these stable validations of their micro-worldviews might (paradoxically?) make them more comfortable embracing others. 🤞
after all, internet 2.0 is a giant externalized validation trap, who’s ML is/was too simple to do it better than tribal-level triggers, when irl most people happily live with lots of contradictory (hypocritical?) worldviews
And honestly having contradictory opinions doesn't mean you have to "balance" things 50% - 50%. The key is achieving equilibrium and that can be weighted and feel more like 70% - 30.
Balance creates mediocrity, and mediocrity wasn't meant for humans. But we need equilibrium to stay afloat to achieve that greatness.