315: Tony Stubblebine, CEO of Medium, On Human Curation, Subscription-Driven Quality, and Fixing the Internet
AI and the Future of Work
Deep Dive
Why did Medium prioritize human curation over algorithmic popularity under Tony Stubblebine's leadership?
Medium prioritized human curation to ensure the platform values thoughtful content over popularity signals. Algorithms are still used for matchmaking, but significant qualitative signals from human curators were added to enhance the recommendation system, focusing on the quality of ideas rather than just engagement metrics.
What business model shift did Medium make to focus on delivering value to readers?
Medium shifted from an ad-driven model to a subscription-based model, allowing the platform to focus on delivering content readers are happy to pay for. This change incentivizes quality over engagement, as the goal is to provide value that justifies the subscription cost.
How does Medium balance human curation and algorithms in its recommendation system?
Medium combines human curation with algorithms by using human curators to provide qualitative signals that algorithms cannot detect. While algorithms handle matchmaking, human curators ensure the content recommended is thoughtful and valuable, creating a balanced and informed reading experience.
What is Medium's policy on AI-generated content?
Medium's policy prohibits non-humans from publishing, but AI-generated content is not outright banned. Instead, human curators filter out AI-generated content that lacks value, ensuring only thoughtful, human-created content is recommended to readers. Writers are expected to self-report and self-police their use of AI tools.
How does Medium handle plagiarism?
Medium takes plagiarism seriously and removes plagiarized content when detected. Accounts found plagiarizing are suspended or banned. While Medium relies on copyright holders to report violations, it also controls distribution and payouts to authors, allowing it to revoke access for plagiarizers.
What is Tony Stubblebine's perspective on fixing the 'broken internet'?
Tony Stubblebine believes the internet is broken due to its reliance on ad-driven business models that prioritize attention-grabbing content over value. Medium aims to fix this by proving that subscription-based models focused on delivering value can be successful and sustainable, offering an alternative to the dominant ad-driven internet.
How does Medium's subscription model compare financially to ad-driven models?
Medium's subscription model is more profitable than ad-driven models. While ad-driven businesses often earn less than $10 per thousand views (RPM), Medium's RPMs are $20 and heading toward $30, demonstrating that focusing on delivering value over attention can be a better business strategy.
What training do Medium's human editors receive?
Medium's human editors are largely self-trained through experience. Many editors come from the community and learn by reading and curating content earnestly. The platform relies on their ability to spot quality and authenticity, which emerges naturally from their engagement with the writing.
How does Medium ensure the quality of its curated content?
Medium ensures quality by relying on subject matter experts who curate content based on deep knowledge of their fields. These curators identify the best versions of topics and guide readers toward thoughtful, well-researched content, ensuring the platform maintains high standards.
What is Tony Stubblebine's advice to entrepreneurs about success?
Tony Stubblebine advises entrepreneurs that success often takes time and persistence. He emphasizes that it's okay to be a late bloomer, as many entrepreneurs achieve their biggest successes in their 40s, 50s, or 60s. He encourages focusing on gradual growth and impact rather than rushing for quick wins.
- Medium's recommendation system prioritizes human curation over algorithmic popularity.
- The platform values thoughtful content over viral content.
- Human curators focus on the qualities of the content rather than popularity signals.
Shownotes Transcript