Paul Graham, cofounder of Y Combinator, says he’s noticed a growing trend of founders using AI to write outreach emails—and that once he realizes it, he gives up because it feels like being lied to. His comments sparked debate among startup leaders and investo
A founder’s pitch can be the difference between “we’ll talk” and silence. But for Paul Graham, the tipping point is sometimes something far more basic than the business idea itself: the way an email sounds.
On May 25. 2026. Graham posted on X that he was seeing emails “in a hard-hitting journalistic style” from founders—and he believes many are written by AI. “I know they’re written by AI, because no founder ever wrote this way before,” he wrote. Once he figures out the message is AI-generated, Graham said, he stops engaging.
“It feels like being lied to, and who would stand for that?” he wrote.
The remark landed in a wider conversation inside tech about how much AI belongs in the earliest steps of getting attention—especially when first impressions are increasingly shaped by tools that can mimic confidence, punchy formatting, and polished tone.
Graham isn’t portraying himself as an AI opponent. A commenter had pulled an older post from April in which Graham wrote that AI gave hardworking founders the “growth they deserve.” The commenter called it “cognitive dissonance.” In response. Graham replied: “You’re supposed to use it. but in the right way. Like any technology.”.
Still, Graham’s current focus is on signals he believes reveal when AI is doing the writing. He previously said that in 2024 he looked for the word “delve” as a sign of ChatGPT’s influence. His May 25 post sparked debate—along with support from people working on large language models. Google DeepMind researcher Nataniel Ruiz wrote: “It’s so true that it’s hard not to ignore something written purely by AI.”.
But not everyone on the debate side treats the issue the same way. Boom Supersonic CEO Blake Scholl said AI writing is “usually a bad idea.” He recommended that people write themselves first, then ask AI for feedback rather than outsourcing the draft entirely.
Serial entrepreneur Zain Manji took a different tack, describing a kind of self-audit. “When I write, I will turn on my AI slop radar and check myself,” he wrote.
Even outside startup culture, the reaction to AI-style writing has been framed as a question of authenticity and craft. Steven Sinofsky. a former Microsoft Windows boss. commented that Graham’s post reminded him of when he first used a “letter quality” printer as a freshman in college. Sinofsky wrote that his professors had mixed reactions: some thought the “format” distracted from the content and felt like a ruse. while others said the formatting made them look more carefully at the work. Still others told him to use a typewriter “like a normal student.”.
Sinofsky pointed to a turning point when the Macintosh computer arrived and wrote, “everything changed.” He then wondered what the “Macintosh moment” for AI writing would be.
Graham’s critique also runs alongside practical hiring realities. In the HR world. AI cover letters and résumés are already a problem companies manage—sometimes with success. other times with applications being tossed in the trash. And broader concerns persist that overly polished writing can hurt founders’ chances.
That tension showed up again in April, when Instacart cofounder Max Mullen said he looks for founders with “dirty white sneakers.” Mullen’s logic was blunt: those founders aren’t focused on their appearance. It’s a sign, he said, they’re likely working “around the clock.”
One commenter asked Graham why an AI-written email wouldn’t actually be a positive sign—arguing that it could signal the founder is “AI native,” much like techies flaunt their “high token bills.” Graham rejected the idea. “Any teenager can do that,” he wrote.
Paul Graham Y Combinator startup pitches AI email founder mode Nataniel Ruiz Blake Scholl Zain Manji Steven Sinofsky Instacart Max Mullen AI cover letters hiring large language models
So what, we’re gatekeeping email vibes now? Lol
I don’t get it, like AI can just make it sound normal. If the product is good why does the email have to be written by a human. He’s basically mad about tone? Idk.
This is actually kind of wild because the whole “delve” thing sounds made up. Like if one word is the tell then anyone can avoid it. Also Paul Graham acts like every founder used to write in their own exact voice but… people always copy each other.
Wait so he just stops replying once he suspects AI? That’s kinda petty tbh. But I guess if it’s that “hard-hitting journalistic” style then yeah it feels fake, like those investor emails that all sound the same. Still feels unfair since plenty of founders aren’t great writers and they’re trying to get noticed. Plus people have been using templates forever, not like that’s new.