Quoting Paul Graham

Origin: A Sharp Observation from a Tech Godfather

A brief observation by Y Combinator founder Paul Graham has resonated widely in the tech community. He noted that many founder emails he receives now adopt a “hard-hitting journalistic style,” which he can instantly tell are AI-written. Crucially, once he realizes it’s AI-generated, he finds it hard to finish reading and feels “like being lied to.” This isn’t just a personal preference; it touches on a deep-seated issue emerging from the widespread adoption of AI: as machine-generated content becomes ubiquitous, how will trust and genuine human communication be maintained?

Breakdown: Why Are AI-Written Emails “Instantly Recognizable”?

Graham’s description of the “hard-hitting journalistic style” is a precise observation. Current mainstream large language models are trained on vast amounts of news, PR copy, and formal documents. Consequently, when generating formal communication, they unconsciously gravitate toward a structured, formal, and slightly promotional “safe” style. This style lacks the casualness, subtle emotional nuances, and unique personal rhythm common in personal emails. For seasoned individuals like Graham who read countless emails daily, this stylistic homogeneity and “inhuman feel” acts as a clear signal. The core issue isn’t whether the AI writes well, but that it erases personal voice, turning communication into standardized information delivery, which triggers the recipient’s psychological defense mechanism—“You’re dealing with me with a formula.”

Trend Insight: AI is Creating an “Uncanny Valley of Communication”

This reveals a subtler trend than “AI replacing human writing”: we are entering an “uncanny valley of communication” phase. Just as humanoid robots that are too human-like yet not quite right can cause discomfort, AI-generated text that is too much like “good writing” but not like “real human speech” can provoke aversion. Graham’s reaction represents a widespread sentiment: we value intent and effort in communication. An email with rough writing but traces of thought can build connection better than a rhetorically polished but obviously template-generated one. When AI drives the cost of “generating perfect text” toward zero, the “informational value” of the text itself declines, while the value of the “human attention invested” and “unique perspective” behind it soars. This foreshadows that effective future communication may not be about whose report is prettier, but whose voice is more authentic and whose thinking is more distinctive.

Practical Value: What Does This Mean for Developers and Professionals?

First, beware of “AI-speak.” When using AI to assist in drafting emails, documents, or social content, deep personalization editing is essential. The goal isn’t for AI to help you “finish,” but to help you “draft” or “expand your thinking.” The final product must bear the marks of your personal expression habits and thought process. Second, reassess communication strategies. In critical business communications (e.g., fundraising, key partnerships), authenticity and personality may be more important than perfect grammar and structure. Sometimes, a hand-written, slightly flawed but passionate message can cut through the noise. Finally, consider product design. For developers, this suggests a product direction: future writing assistance tools should aim not just to generate “better text,” but to help users discover and amplify their own voice. The tool should act more like a coach than a ghostwriter.

Counterintuitive/Unexpected: Using AI as a Ghostwriter Exposes Anxiety About Ability

There’s an easily overlooked sharp point in Graham’s comment: “It means they can’t write well unaided (or feel they can’t), and that they’re trying to trick me.” This points to a deep-seated psychology: many people use AI ghostwriters not out of efficiency considerations, but源于对自身表达能力的不自信. They attempt to use AI’s “perfection” to掩盖自己的 “imperfection.” However, Graham’s reaction shows that this strategy often backfires, exposing anxiety and losing trust. The real solution may not be finding more隐蔽 ghostwriting tools, but embracing imperfection, improving one’s own expressive skills through practice, or honestly using AI as assistance rather than a replacement. In the AI era, daring to show an imperfect thought process may itself become a new form of competitiveness.