[As more people use AI to compose grammatically perfect and typo-free emails, it’s easier to detect the AI and dismiss the message as computer-generated. To counter this break in medium-as-social-actor presence, and to draw attention to the problem, a Harvard student created software that uses AI to edit your AI-generated emails to make them more likely to appear to have been written by a human (!). Details are in the story below from Livemint, where the original version includes the relevant segment of a TBPN podcast (also available on YouTube) in which the creator discusses Sinceerly. See also earlier coverage in Business Insider (which is behind a paywall). –Matthew]

Meet Sinceerly: Harvard student uses Claude to build ‘anti-Grammarly’ tool to add typos to your AI emails

Harvard student has Ben Horwitz created Sinceerly, a Chrome extension that injects errors into AI-generated text to make it appear human

By Aman Gupta
April 25, 2026

Writing an email with perfect grammar and spelling was once considered a sign of professionalism and attention to detail, but with the rise of generative AI chatbots like ChatGPT and Gemini, a too-perfect message increasingly seems to be written by a machine. The impact of AI on our daily lives is perhaps most visible on social media, which at this point is flooded with AI-generated content created en masse to capture eyeballs.

In fact, a message that contains typos has seemingly become a sign that a human was actually behind the text. To emulate that effect for AI-generated writing, a Harvard student has created a Chrome extension called Sinceerly.

The tool, built by Ben Horwitz, flips the idea behind writing assistants like Grammarly. Built using Anthropic’s Claude AI, the tool takes perfectly polished AI-generated text and injects calculated mistakes to make it look genuinely human.

What is Sinceerly?

Speaking on the TBPN podcast, Horwitz described Sinceerly as “the anti-Grammarly,” designed to counter what he calls the growing problem of “AI slop” in inboxes.

“I was so sick of AI slop in my own inbox,” Horwitz explained on the TBPN podcast. “People would email me — it all looks the exact same. And that, coupled with the news about how other people have been emailing across social strata, inspired me to basically create the anti-Grammarly — something that really messes up emails”.

Horwitz noted that the project started as a joke and a form of social commentary on modern workplace communication. The extension offers a few free uses before charging $4.99, though he does not expect it to become a major business.

He also explained how humans are using AI to make the text feel more human.

“I think what’s kind of crazy is people are copy-pasting — first they’re prompting, then they’re copy-pasting their prompt into email, and then they’re using Sincerely to make their email not sound like AI. And that’s a crazy loop, right? Humans using AI to make AI more human,” Horwitz said.

Horwitz also noted that they have pushed a “few security fixes (for Sinceerly) and made sure people know what they’re getting into.”

Three modes of Sinceerly:

Sinceerly operates through three distinct modes to intentionally downgrade your writing:

Subtle: This mode streamlines the text by removing filler words, converting phrases into contractions, and typically injecting a typo directly into the first sentence.

Human: This setting takes the ‘Subtle’ approach a step further by forcing a highly conversational tone alongside the intentional mistakes.

CEO: This mode is designed to reflect the power dynamics and fast response times of corporate leadership by converting the text to all lowercase and applying intense brevity. If the email lacks a signature, the tool will sometimes automatically append “sent from my iPhone” to complete the aesthetic.