How to win — and lose — Decoder

Nilay joins as the guest to discuss our AI coverage, controversial episodes, and what it takes to succeed or fail on Decoder.

This interview has been lightly edited for length and clarity.

Hello and welcome to Decoder , Nilay's show about big ideas and other problems. This is Nick Statt, senior producer, and I'm joined by host and very occasional guest, Nilay Patel. Nilay, welcome back to your own show.

Hello. I dislike being the guest.

Now, you have expressed this in the past, but there's also a side of you that believes this is the ideal format for the show, where you simply participate in discussions without any hosting responsibilities. It seems you hold two perspectives regarding what the ideal version of Decoder entails.

Achieving the status of being a permanent guest is a significant accomplishment—one where others invite you simply because they find you intriguing. I aspire to reach that level of success. However, being a guest means maintaining a consistent standard of interest, whereas hosting allows me to guide the conversation. As a host, my role is to prompt, "Can you engage and be interesting for an hour?" Then we observe the outcomes.

So that is my responsibility today. A few months ago, we conducted another mailbag episode, which we considered an annual tradition around the holidays, aimed at responding to listener questions, feedback, criticism, and suggestions. However, we've decided to conduct these more frequently due to the volume of insightful feedback we receive, and we genuinely consider every email. Therefore, we are here once again. Nilay, are you prepared to dive into this?

Yes, let's proceed.

Our most popular episode of this year also turned out to be the most contentious. It was your interview with Superhuman CEO, Shishir Mehrotra, which primarily addressed the Grammarly expert review controversy. We received an overwhelming amount of feedback regarding that episode, with the majority being very positive. Several compelling emails, comments, and pieces of feedback stood out to us.

Comments included sentiments like, "Wow, Nilay's questions made me nervous," which ranked among our top responses. Another listener commented, "Tech CEOs need to experience this level of discomfort more often." A Verge subscriber mentioned, "This episode was extremely uncomfortable to listen to, which is exactly why I became a subscriber less than a week ago." To start, Nilay, how did you perceive the reception of the Superhuman episode? Were there any aspects of the reactions that surprised you?

Some reactions did catch me off guard. As Nick mentioned, Shishir was scheduled to appear long before the controversy arose, and I was genuinely looking forward to the conversation. His background includes roles as both the chief product officer and the chief technology officer at YouTube, and he serves on Spotify's board. His consideration of distributing AI via Grammarly presents notable challenges, especially in competing against tech giants like Google and Apple, which are gradually incorporating AI into their systems. Consequently, there was a wealth of discussion points related to the creator economy and the future trajectory of AI.

Then the controversy unfolded. I commend Shishir for participating in the interview; he was aware of the likely nature of the discussion. While we do not provide guests with our questions in advance, it was evident what topics I intended to pursue. I believed he could handle the tough questions given his extensive history with major companies. My intention was never to place young entrepreneurs on trial for the industry's issues, but Shishir's depth of knowledge and vast network made him an ideal candidate for this discussion.

Given Shishir's stature, I felt it was appropriate to ask him about the specific issues as a representation of larger AI challenges. Many listeners seemed to resonate with that approach. What surprised me, however, were responses that suggested, "You don’t understand AI. This is simply how it operates. You lack insight into the builder's perspective."

While I can appreciate that viewpoint, my counterargument is that this is precisely what Decoder seeks to explore. What are the implications of creating these products? How do they function in reality? How should they be designed to align with our expectations? Furthermore, if we fail to relentlessly question these matters, we will not compel product creators to contemplate the appropriate answers. That was truly my aim.

I held confidence in Shishir's thoughtfulness and his willingness to engage with challenging conversations, which is why I felt empowered to ask my questions as straightforwardly as possible. This may have induced discomfort, but I believe it served the audience well. The tension we addressed mirrors the current landscape of dialogues surrounding AI.

Are companies excessively appropriating our contributions? Are they disregarding legal protections pertinent to creativity, likeness, and the compensation of authors and creators for their work? We appear to be advancing without adequately resolving any of these critical questions. I am pleased with the outcome of that episode, and while the reactions were expected, I was somewhat surprised that our mission at Decoder generated discontent. Engaging with our 'products' should be embraced, and I hope we continue this journey.