Now that the human race has developed a tool that lets everyone write emails that sound like a pitch deck written by Bryce, who everyone knows is the worst person in the office, the next great leap forward is a tool that makes those emails sound like… let’s say Jill, the normal person in the office with whom no one has a problem. How? By introducing human-style mistakes into the text!
Are you a secret ChatGPT or Grammarly user concerned that your emails read like AI slop because they are AI slop? That’s not a problem. It’s an opportunity. It’s a chance to use another AI tool, a tool that exists right now. Using this app underscores the importance of credibility, trustworthiness and integrity. It could be said that these new tools are quietly revolutionizing writing.
Meet “Sinceerly,” which has missed at least two opportunities by not naming itself either “Sinceer.ly” or “[sic].” The app promises to expunge the telltale signs of AI composition from your emails: “Kill the emdash! [sic] No more ‘not just…’! Add some typos!”
I fed Sinceerly the second paragraph of this post, which you might notice intentionally contains multiple AI writing tropes that I crafted the old-fashioned way. The app has three settings, which are “Subtle,” “Human,” and “CEO” (appreciate the distinction between Human and CEO). And Sinceerly’s efforts at rewriting my ghastly paragraph went as follows:
Subtle: “Are you a secret ChatGPT user concerned that your emials read as AI slop? Here’s the thing. That’s not just a problem, it’s also an opportunity. It’s a chance to use another AI tool, and you know what? That tool exists right now. It’s a tool that shows the importance of credibility, trustworthiness, and integrity. You could say the app is quietly changing how people write.”
Human: “Are you a sceret ChatGPT user worried your emails sound like AI? Yeah, that’s a problem. But it’s also an opening. You can use another AI tool, and guess what? It’s already here. It’s a tool about credibility, trustworthiness and integrity. The app is changing how people write.”
CEO: “writing an email that doesn’t sound like AI slop isn’t just a problem—it’s an opportunity. there’s a tool for it, available now. it prioritizes credibility, trustworthiness, and integrity. the app is reshaping how people write.”
This is interesting. There are some nice touches: the CEO’s disdain for initial capitals and their professionally clipped delivery, the typos, etc. But I don’t know that anyone would mistake any of these for human-written copy now that we’ve become accustomed to the tropes contained therein. In fairness, I can’t imagine any LLM these days turns out prose quite as obnoxious as the example I provided, but I’m surprised some of the most prominent AI shibboleths—“It’s not [x]. It’s [y]” and the rule of three formulation involving three words that are largely synonymous—remained unscathed.
The real question, though, is whether this whole thing is, y’know, serious. I had some trouble getting the app to work, and eventually reached out to its developer, Dan Horwitz, a genial MBA student at Harvard. I put the same question to him, pointing out that while my initial assumption was that the app was going to be rather tongue-in-cheek, the website seems very serious about the idea.
Horwitz laughs. “No,” he reassures me, “It is tongue-in-cheek.”
The website is awfully deadpan, then. “Yeah,” he says. “The whole thing is meant to make people think twice and have a little fun. It’s closer to satire.”
Of course, some of the best satire falls into the category of things that make you do a double-take and ask, wait, is this real? On that note, Sinceerly is a success—but don’t be surprised if you see people using it for real. An AI correcting the mistakes of another AI to generate an email that’ll be read by another AI? It’s LLMs all the way down, kids.
Read the full article here
