Paul Graham recently warned founders not to send him AI-written emails. The instinct behind that is understandable — he’s trying to read something deeper than language: urgency, judgment, originality.
But there’s a tension here.
Because the same world now actively encourages AI to help write, refine, and “improve” those exact emails.
So the real question is no longer whether AI was used.
It’s whether the thinking still belongs to the writer — or whether it has been quietly pre-shaped by the tool.
This episode isn’t about style. It’s about ownership.
AI can smooth language, but it can also smooth away signals: friction, intent, and edge.
And once everything looks polished, the harder question becomes unavoidable:
What part of this is still yours?
Watch here:
https://chihchienliu.substack.com/p/ep-183-ai-writing-human-voice-and
#AIWriting #HumanVoice #Writing #Founders #Substack #FutureOfWork #Communication
下一則: Episode 235: Force in Transition: Intelligence, Control, and Uncertainty
- How Do We Stay Human in an Automated World?
- 人文素養英文閱讀系列-18
- EP 184: From Teamwork to Human-AI Collaboration
- Episode 237: AI, Veto Power, and the Structural Split in Global Governance
- Episode 236: The Drift from Dialogue to Default Thinking
- Episode 235: Force in Transition: Intelligence, Control, and Uncertainty


