Mark Manson is launching an AI app to help more people ‘Subtly not Give a F*ck’

Mark Manson’s 2016 book The Subtle Art of Not Giving a Fuck delivered some hard truths and prescient advice to millions of readers seeking answers. Now he’s building an AI-based application to do the same.

At that time, Manson says the self-help field was “unrealistic, not very evidence based—just designed to make you feel good,” inspiring him to write a book that offered “a more skeptical, realistic, and zero bull shit approach to personal growth and self-help.”

Nearly a decade later, Manson says he’s seeing the same pattern in the digital world, with millions turning to generic AI platforms for guidance, only to receive unrealistic, potentially harmful advice. That’s what inspired him to team up with serial tech entrepreneur Raj Singh—who most recently sold his AI hotel concierge service GoMoment—to create on-demand life coaching app Purpose, which launches this week.

The AI tool offers users life guidance and actionable steps to solve problems, from career to relationships and beyond. It also seeks to challenge, rather than validate users (something many experts have argued many AI platforms do), albeit in a polite, supportive tone. It pairs existing research into cognitive behavioral therapy with data gathered from users upon sign up, and retains conversation history to track patterns over time. According to an internal survey, 41% of early users said the app has been “life-changing.”

“We’ve actually had a number of people tell us that they’ve cried while using it,” Manson says. “We really tried to build this AI to go deep quickly, to not beat around the bush or do any fake pleasantries.”

Fast Company caught up with the author-turned-tech-founder from Los Angeles to talk about the legacy of The Subtle Art, whether we’re ready to share our deepest thoughts with an app, and why AI is the perfect tool to deliver personalized coaching at scale.

This conversation has been edited for length and clarity.

Were you surprised by the success of The Subtle Art of Not Giving a Fuck?

The scope of the success caught me off guard. My pie-in-the-sky dream at the time was a million copies, so the fact that it’s exceeded that so drastically has been a pleasant surprise.

Why do you think it was such a hit?

When I got started in this industry in the early 2010s, self-help was very touchy-feely, very woo-woo. It was, in my opinion, unrealistic, not very evidence based—just designed to make you feel good. It felt like there was a deep hunger, especially among millennials, for a skeptical, realistic, zero-bullshit approach to personal growth and self-help. That was the inspiration.

Coincidentally a lot of the cultural observations at the time—like about social media and distraction and chronically comparing ourselves to others—have aged really well. I think I identified early on the world we were heading toward.

How has your life changed since?

Financial security is probably the single biggest change, and that’s great, but being a very successful author is weird in that the book is the celebrity, not me. I’m not getting recognized on the street or anything. It’s just a bunch of numbers on emails that you’re like “wow, that’s a large number.” Then you go back to your same apartment and hang out with your same friends.

Why transition from author to tech cofounder?

When my career started to blow up, there were opportunities to do seminars and start a coaching business, but I didn’t really want to charge somebody $5,000 to spend a weekend in a hotel ballroom with me. I’ve been obsessed with this question of, how can we help more people? Like, what does a better version of this industry look like? 

One of the problems I identified well before ChatGPT came out is that the stuff that actually works doesn’t scale, and the stuff that scales doesn’t work. My book was read by millions of people, but a book has to talk in broad principles, and the reader needs to connect the dots for themselves. It might move the needle for some people, but not much. What really works is working with an excellent coach or therapist, which is extremely personalized and requires a major investment of time and energy, so it doesn’t scale.

Then ChatGPT came out, and I heard people were asking it big life questions. I tried it myself, and it does okay in terms of some questions, but not others. It felt like there was an opportunity for a properly trained AI to scale that personalization.

How did you meet your cofounder?

At a poker game. After ChatGPT came out, I was meeting with AI companies, but they didn’t really get it. And then I sat down next to Raj Singh at a poker game just as he was finishing a sabbatical in 2023, and was independently thinking about doing something with AI and mental health. We started chatting, and it turns out we were thinking about the same question.

What is Purpose all about? What makes it different from generic AI platforms?

Purpose is a personalized AI mentor designed to help you find clarity and direction in your life as soon as possible. One of the biggest issues I see with using ChatGPT for life questions is that its default approach is to validate you. If you complain about your ex-girlfriends, ChatGPT will tell you they’re terrible, and you’re a great guy, so just pick yourself up and it’ll be okay when what you might actually need to hear is like, “hey dude, you might be the problem.” The goal with Purpose is to build an AI that will challenge your assumptions, poke holes in some of your beliefs, point out blind spots, and help you reconsider how you see your own approach.

The app does take a positive tone, though.

There’s a fine line between flattery and positivity, and it’s something we’re trying to calibrate all the time. It changes based on the user, the context, the issue.

Part of our onboarding experience is designed to get an early read of the user’s personality traits. The AI speaks to people differently based on how agreeable they are. Some people want very direct advice, others want it to be a bit more encouraging and positive, and it’s something we’re working on fine tuning all the time.

The entire field of psychology is already in all these LLMs. The value is finding that calibration. ChatGPT is poorly calibrated for it, partly because of the sycophancy, partly because of the lack of personalization, partly because of the poor memory. We’re trying to fix those things.  

Are people open to receiving that advice from a bot?

When you’re talking to a human—even a trusted person, like a therapist—it’s almost impossible to not worry what that person thinks about you. With AI, I don’t care. I don’t feel shame: I just say what I feel. And what I’ve also found is that when it gives me harsh feedback, I don’t feel that social anxiety that’s attached to criticism or negative feedback from a person.

What about privacy and security concerns?

Trust is by far the most important thing in this space. We are building Purpose with all the highest security and privacy regulations, like HIPAA and GDPR. Conversations are anonymized on each device, so even if somebody puts a gun to my head or Raj’s, we can’t share your conversations, because we don’t have any way to identify individual users. It’s also entirely self-funded—which is something Raj and I feel very strongly about—to keep the proper incentives in place and not jeopardize the goal and mission of the product.

What’s next for you personally?

I am going to be strongly involved in Purpose’s product design, but I do feel like it’s probably time to write another book. It’s been about five years, and I need to get back on that horse.

Did Purpose suggest that?

Not specifically. But it probably would have if I asked. 

View the full article