Inside Medium’s mission to elevate authentic voices: a Q&A with CEO Tony Stubblebine

As a writer who has been active on Medium for nearly a decade, I’ve had a front-row seat to the platform’s evolution. In recent years, Medium has taken bold stances on issues like compensating writers, not using their content to train AI models, cracking down on low-quality AI-generated posts, and carefully considering its approach to content moderation in an increasingly polarized online landscape.

To gain a deeper insight into Medium’s philosophy and future direction, I’m excited to bring you an in-depth conversation with Tony Stubblebine, Medium’s CEO—perhaps better known by his nickname, “Coach Tony.”

Joan Westenberg: Tony, thanks so much for taking the time to chat with me today. As a writer who has been active on Medium for nearly a decade, I’m really excited to dig into the platform’s evolution and future. Before we get into it — I’d love to know the story behind your “Coach Tony” nickname.

Coach Tony: I have a huge self-improvement background. Before Medium, I ran a company called Lift that pioneered habit tracking on the iPhone. It then morphed into, a coaching platform and directory with a focus on habit coaching. I also ran the Better Humans self-improvement publication here on Medium, reaching millions of people. So, “Coach Tony” stems from my work helping people improve their lives.

JW: Let’s start with the big topic on everyone’s minds—AI. Medium has taken a strong stance against allowing the platform’s content to be used to train AI models without compensating writers. Can you expand on Medium’s position?

CT: Our stance has a few key points. One is that it’s unfair for AI companies to train on writers’ content without consent, compensation, or credit. They’ve broken the basic premise of an exchange of value. We feel our role is to negotiate on behalf of our writers since we have more clout than any individual author would.

Secondly, from the author’s perspective, if AI companies can turn their content into a tool handed to spammers, it destroys value for writers. We get flooded with 4x as much spam as we used to. So we’re focused on looking at deals with AI companies that involve compensation that we can pass back to writers. We won’t do any deal unless it’s widely supported by the Medium community.

It also feels morally right and aligned with everything else we do. Someone asked me recently if it wouldn’t be better just to pocket this AI licensing money. The way I answered is that the thing I like about Medium is that there’s a way to make it a triple bottom line — it’s a win for the readers, a win for the authors, and a win for the company. You break that virtuous cycle when you start doing stuff like selling people’s data behind their backs. I think we can build a big business without having to do that, so why start destroying trust at the very moment when we’re just getting it back?

JW: I’m glad Medium is looking out for writers on this front. Related to that, how is Medium approaching the growing issue of AI-generated content on the platform itself?

CT: We take a default “no” stance to AI-generated content. Where that content really hurts is when human curators and publication editors encounter floods of AI content. They’re doing heroic work to keep that content from reaching readers.

The recommendation systems effectively filter it out because there’s a lot of high-quality human curation. We’re not losing writers or readers over this. Our audience is still growing. But it’s a huge bummer for editors who have to wade through the poor-quality AI content.

By the way, it’s not like we’re necessarily correctly identifying everything that’s AI-generated. We’re really identifying that it’s not good. That’s why we don’t like AI-generated content, generally — it’s confidently wrong, or confidently cliché. And you know, if a human submitted the same bad writing, we wouldn’t show that either.

JW: Makes sense. Do you think AI-generated content could ever be good enough to be published somewhere like Medium?

CT: Mostly no, because Medium is about substance, not entertainment. We’re trying to get full human stories with context and emotionality, not just nuggets of information. AI-generated content is missing the humanity that makes a story powerful. Some people think factual info can be shared efficiently in a condensed AI-generated form, but they misunderstand how the human brain works. The human element is a required part of meaningful information sharing.

JW: Shifting gears a bit; in recent years, with — shall we say — “troubling” world events, I’ve felt that my role as a writer has become more about saying difficult, important things that might not be palatable on other attention-driven platforms. How do you see the balance on Medium between substantive writing versus people using the platform more as a business and promotional tool?

CT: I would split the business people into two sides — one of which is dead on what Medium’s mission is, and the other is off-mission. What we’ve had to get out of is Medium being adopted by the “creator economy.” The bulk of the creator economy is essentially people who have been sold a get-rich-quick scheme of passive income, that it won’t be that hard, and you can live comfortably and travel the world. But the reality is that content is bad on both sides — for the creator, it’s a huge grind and they’ll probably fail. Even if they don’t fail, they’ve signed up for a content treadmill that’s really hard to keep up with and can end up miserable. On the reader's side, this movement has created a lot of ill-informed content — the internet has been flooded with people who don’t really know what they’re talking about at a deep level. It’s a lot more opinion and easy summarization, and the depth has disappeared. The people with real depth have been crowded out.

On the other hand, before the creator economy existed, writing online essentially supported the “expert economy.” You write to demonstrate your expertise while simultaneously advertising yourself to people who will pay for that expertise. It’s all above board. A good example is a tech company with an engineering or design blog, where they are trying to write the smartest thing possible because it will resonate with the smartest job candidates they’re trying to reach. It’s the highest quality writing because the incentive is to impress other smart people. I see that in a lot of good professional writing on Medium. They use writing as a portfolio to advertise their services, and it’s really effective. It incentivizes them to write true, interesting and deep stuff, and to be recognized as the person who knows what they’re talking about. That’s where you get more business.

So, if you pair those people with people who are just writing to write, we end up with a great set of authors writing from a real place of authenticity. That’s what got lost in the attention economy — people are writing to get attention, but not writing anything true or with real passion behind it. We’re finding the authentic voices that got crowded out and elevating them. There are actually more of those voices out there; they just didn’t want to do the work of hustling for readers themselves. We can do that part for them.

JW: That’s an important distinction. Related to the rise of shallow content, we live in a world of increasingly short attention spans and addictive video platforms like TikTok. Do you worry about an overall decline in the attention people are willing to devote to in-depth reading and writing?

CT: I frame it in terms of two complementary things. The sweet spot of attention is entertainment and leisure—when we’re in that mode, we want to maximize the feeling of entertainment and leisure. TikTok is doing a great job at that and has replaced much of my leisure and entertainment time.

On the other hand, the complement is substance. There’s a world of people — you wish it were everyone, but it’s still enough people to build a business — who think, “My life will be more successful if I’m smarter.” That’s the mindset behind why you go to college — “I’ll be educated in a way that helps me get a job.” Then, once you get the job, you will continue learning to get ahead. All of these things are backed by the idea that being smarter will make life better.

The question for that person is not “What’s the most addictive media?”

The question is, “What’s the most effective media for making me smarter?” And that’s still reading and writing. You can see this in other places, too—it’s why Amazon is managed by memos rather than slide decks. People who write memos think better and deeper and communicate those thoughts better than people who communicate through slide decks.

For the world of people who care about the end result and about media that effectively transforms their lives, nothing has replaced writing in terms of effectiveness. It’s still the best tool for that. And since we’ve gotten out of the ad business and become a subscription business, we don’t have to compete for attention. We have to compete on the idea that we can effectively make your life better. I have no worries about macro trends like people pivoting to video. That’s not where people will spend their time if they’re looking for something truly valuable that will change their lives. Increasingly, Medium is helping more of them find exactly that.

JW: Content moderation has become a thorny issue beyond banning hate speech. We’ve seen the Substack meltdown and other instances of Content moderation issues on X, etc. What’s your overall philosophy toward content moderation in 2024?

CT: Other people’s philosophies have become so polarised that you can’t even talk about them functionally. But there are two frames I use to get to a healthier place.

One is that the content is the product. Even though we build the software, when a reader comes to your platform, what they read or view is their experience—that’s the product. A good example is the programming Q&A site Stack Overflow. It’s a programming site not because of the software but because of the content it allows and the moderation policies. You can’t post political content there. As a result, everyone understands it’s a product for programmers.

The other frame is that it amazes me that people will talk about free speech but not property rights. In the entrepreneur and hacker communities, people get worked up about free speech, but they operate in a context where investors put money into businesses expecting a return. If the platform only exists because investors thought they could build a business, you have to allow that they have the right to determine what type of product they can build a business around. It’s Stack Overflow’s right to have content moderation policies that let them build a business for programmers.

So to take this to Medium, we want to meaningfully deepen people’s understanding of the world. We think we can build a business around that — investors put a lot of money into the company based on that premise. And now we have traction that shows it’s working. If we do it effectively, readers will pay for a subscription and we’ll have a growing, functional business.

The alternative is to essentially not have the platform at all because the platform couldn’t have existed without those investors who expect you’ll build a real business someday. So, our policies are there to allow us to achieve our mission.

We think that’s alright. We come to work daily, spending investor money and our time to build something valuable. The market can accept or reject it, but it’s alright for us to determine what we want to build. And that comes down to what content we allow and elevate. Reach is a big part of it — it’s not just what you can publish, but what gets distributed.

Some of that perspective came from the folks who originally built Medium, some of whom built Twitter. The divisiveness there didn’t create a conversation that effectively deepened understanding. It made people dig into entrenched thinking. Our moderation aims to dampen those unproductive divisions. Hopefully, readers think Medium is a pretty healthy community. The worst you’ll likely encounter is some spam, not the level of hate and vitriol you see on Twitter.

We can deliver that product because of our Trust and Safety team and policies. So far, we’ve been able to talk about Medium without getting mired in debates over censorship. We see it as facilitating a certain type of community. I’m curious to get your take on that, as I haven’t explained it to many people in this way before.

JW: You raise an important distinction between distribution platforms and media companies. That’s where a lot of the content moderation debate gets thorny. Is Substack just a place for blogs, or is it a media company? I’d argue that by actively recruiting and paying writers, it behaves more like a media company.

In contrast, Medium is still an open platform — anyone can sign up and start writing. But you’re not making active decisions about who to recruit. That feels more like a distribution role. However, I think any time a company develops an editorial voice around what views it wants to elevate, it starts to take on the responsibilities of a media company, including the content moderation challenges that come with that. There are valid concerns about corporate censorship at that point.

CT: I would frame Medium’s role a bit differently. Our business model is a bundled subscription to quality content. What we recommend is fundamentally tied to the business. We have to have a point of view because we’re literally selling subscriptions to content we think is valuable.

There’s no way not to have a point of view if you want to build that kind of business. If you leave it up to algorithms, the implicit point of view is just “We’ll distribute whatever gets the most engagement.” We’re not doing that.

Now, are we going so far as deciding what the fashion trends are, or only delivering a certain political perspective? No. But I think as soon as you have any human curation and recommendation, you have to own that you’re not a purely neutral platform. You’re making decisions about what’s going to deliver value to subscribers.

It’s very different from the telephone company metaphor — if your business is supplying phone lines, who are you to dictate what people say? But if your business is based on getting people to pay for content directly, you absolutely have to have a point of view, because that’s the product.

So, right now, I’ve been describing Medium’s approach as “discriminating, not discriminatory.” We’re not the phone company. We have to proactively decide what content will grow the business.

I think a lot of the debate around this comes down to the lack of functional competition. If Medium, Substack, or Twitter were monopolies, I’d be a lot more concerned. But we’re not in a monopoly situation with newsletters or social media right now. If Substack wants to be a free-speech absolutist, and some people don’t like that, they can go elsewhere. That choice is really important.

JW: Honestly, it was a big misstep for Substack to wade into that quagmire like they did. But I think it comes from a place of arrogance among some tech leaders. They think they always know best, rather than listening to and being shaped by their communities. To Medium’s credit, you’ve taken such a community-driven approach in recent years.

CT: Thanks, I appreciate you saying that. We’re working hard to do right by our community.

JW: Last question before we wrap up — I’d love to get your take on the rise of decentralized social media. Medium has made some moves to integrate with the Fediverse. Do you plan to do more in that direction?

CT: We were early adopters of Mastodon because we could see where Twitter was heading, and short-form posting felt like an important complement to our long-form focus. Helping writers share ideas in short form is a natural fit.

But Medium is not positioned to lead the way on short-form publishing. We’ve got our own business to run, and we’re a small team. We want to focus on doing one thing really well.

So I think we’re in a waiting and supporting role. We’ll keep an eye on things like how Threads develops. I do think Threads is the big player to watch now. I suspect Twitter will continue to fade. I don’t think Bluesky, in particular, will survive, though I think the broader Fediverse will stick around as an alternative. The question is just what percentage of the market will it be — one per cent, five, or ten? I hope it’s a substantial chunk, but it’s hard to say.

Based on my conversations with folks at Meta and others trying to advance the decentralized social media space, I will share that Threads is very serious about opening up to the Fediverse. But the end results still feel murky to me.

Long-term, Medium’s focus remains on long-form writing. Our interest in short-form is about how we can amplify long-form ideas and build a bridge for short conversations to turn into deeper writing. But right now, the whole ecosystem is in flux.

@Westenberg logo
Subscribe to @Westenberg and never miss a post.