379: ‘An Extra Kick in the Nuts’, With Christian Selig

Episode Summary

Executive Summary: Christian Selig discusses Reddit's new paid API pricing and the forced shutdown of Apollo, arguing the move is abrupt, strategically misguided, and damaging to trust in Reddit's platform. He explains why simply raising user prices or relying on ads wouldn't work, details his attempts to work with Reddit, and reflects on Apollo's origins, design philosophy, and possible future beyond Reddit.

Main Topics: Reddit's API pricing and Apollo's shutdown (Priority: 5/5): The core topic is Reddit's decision to charge for API access at a level that makes Apollo economically impossible, forcing the app to shut down despite years of success and loyal users. Developer trust and platform responsibility (Priority: 5/5): The conversation emphasizes that Reddit's abrupt policy shift undermines trust for developers who built businesses on Reddit's API, echoing Twitter's earlier third-party app collapse. Why higher subscription prices or ad models won't solve it (Priority: 5/5): Selig explains that passing costs to users would still leave many subscriptions unviable, create feature throttling incentives, and fail to solve the scale problem. Reddit's business strategy, IPO pressure, and AI monetization (Priority: 4/5): The hosts and Selig explore whether Reddit is pricing APIs mainly to improve IPO optics and monetize AI training/data extraction rather than support third-party clients. Apollo's design philosophy and technical efficiency (Priority: 4/5): Selig defends Apollo as a fast, carefully engineered client that already works within published limits and includes server-side optimizations and moderator tooling. Apollo's origins, monetization, and design legacy (Priority: 3/5): The discussion looks back at Apollo's beginnings, its one-man development model, custom icons, and iPad ambitions, framing it as a polished indie app built around user delight. Possible futures beyond Reddit: Pixel Pals and other platforms (Priority: 3/5): Selig considers whether Apollo could adapt to alternatives like Mastodon, Lemmy, or Kbin, but is skeptical about long-term portability; he notes Pixel Pals as a successful side project.

Key Arguments: Reddit's API price is so high that Apollo's continued operation would cost about $20 million per year, making the app unsustainable. Thirty days' notice is far too short for an indie developer with annual subscriptions already in circulation and obligations to users who prepaid. Apollo is not an inefficient or careless API consumer; its average usage is far below Reddit's rate limits, and Selig has actively optimized it. A single paid tier cannot cover both light and heavy users without either losing money or forcing feature degradation to reduce usage. Reddit should have used separate pricing classes for consumer apps, moderator tools, and large-scale data/AI training access rather than one punitive model. Reddit's move likely reflects IPO pressure and the opportunity to monetize AI companies training on Reddit content, but it collaterally harms legitimate third-party clients. Third-party apps and moderators provide value back to Reddit and should not be treated like generic commercial data scrapers. Reddit's public handling of the situation damaged trust further, especially through accusations that Selig tried to shake them down. Apollo's success came from being a fast, polished, Apple-like Reddit client with customization and thoughtful UX, which the pricing decision now destroys. Even if Apollo cannot continue as a Reddit client, Selig has other opportunities and a strong reputation in indie software. Reddit's community-based value depends heavily on volunteer moderation and user-generated content, making reputation especially important to the company's future.

Data Points: API bill estimate: $20 million/year - Projected cost for Apollo if Reddit's usage pricing were applied at current levels Notice period: 30 days - Time from pricing announcement to when charges would begin Initial pricing decision delay: 43 days - Time Reddit took from announcement to final pricing details Apollo average requests per user per day: 345 - Selig's estimate of average Apollo usage Paid user requests per day: ~500 - Average usage among existing subscription users Heavy-use subgroup: ~6% of users - Users generating between 1,000 and 2,000 requests per day Heavy-use cohort size: almost 100,000 users - Approximate number of very high-usage Apollo users Heavy-use monthly cost: $7 to $14 per user - Estimated Reddit API cost for the heaviest Apollo users under new pricing Existing yearly subscribers: ~50,000 - Users whose subscriptions would continue after July 1 Monthly bill from lingering yearly subs: about $100,000/month - Projected ongoing cost while existing annual subscriptions lapse Apollo icon count: over 100 - Number of alternative app icons available earlier in the year Potential icon total: close to 120 - Including unreleased icon designs ARPU comparison: lower than peers - Discussion of Reddit's low average revenue per user relative to other platforms Reddit user base in Apollo: almost 90% American - Selig describes the geographic composition of his user base Apollo's side project success: 90% of Pixel Pals users don't have Apollo installed - Shows that Pixel Pals became its own business outside Apollo Apollo user base on iPad/Mac: iPad about 5% of iPhone; Mac about 1% - Why a Mac version was not prioritized

Pivotal Quotes: "I don't 100% see it happening. And I feel like getting my hopes up in that way would not be smart." — Christian Selig: On whether Reddit would reverse course before Apollo's shutdown deadline "I can literally go back through my emails and I have conversations where I've talked to Reddit and said, like, the way this API works, isn't that efficient." — Christian Selig: Responding to Reddit's claim that Apollo was inefficient or not optimized "I genuinely thought... we were going to kind of go back to the drawing board, talk to developers, figure out how we can make this right." — Christian Selig: Describing the response he expected from Reddit leadership after the controversy

Implications: The episode shows how fragile platform-dependent indie businesses are when a host company changes rules abruptly. It also highlights the need for clearer API classes, better developer communication, and diversified product strategies beyond a single platform.