432: ‘You and Frank Sinatra’, With Dan Moren

Episode Summary

Executive Summary: The conversation centers on two big Apple-related themes: the limits of AI-driven browsing and the latest Apple hardware cycle. The hosts argue that ChatGPT Atlas feels like an "anti-web" browser that adds friction, weakens normal web navigation, and serves OpenAI more than users. They then discuss the UK App Store ruling, concluding Apple’s 30% model is increasingly untenable. Finally, they assess M5 updates and the iPhone Air, describing the former as sensible speed bumps and the latter as a compelling but likely niche product.

Main Topics: ChatGPT Atlas and the case against AI browsing (Priority: 5/5): The hosts critique OpenAI’s Atlas browser as a solution in search of a problem. They argue that chat is great for questions, but a browser should remain a browser—fast, direct, and visually navigable rather than mediated by prompts and agents. Apple’s App Store legal and regulatory exposure (Priority: 5/5): They examine the UK ruling against Apple over App Store commissions, arguing this is part of a broader global backlash against Apple’s centralized control and 15-30% take rate. They discuss how Apple may eventually need to proactively redesign distribution and payments. The tension between AI convenience and web authenticity (Priority: 4/5): A broader thread through the discussion is that AI summaries and overviews often replace reading, browsing, and research with shortcuts that can be wrong or shallow. They worry this weakens the web ecosystem and encourages hallucination-prone behavior. M5 product cycle as disciplined incrementalism (Priority: 4/5): The M5 MacBook Pro, iPad Pro, and Vision Pro are framed as expected speed bumps rather than revolutionary upgrades. The hosts praise Apple’s cadence and product discipline while noting that annual chip advances are mostly incremental for existing owners. Vision Pro improvements and persona/FaceTime breakthroughs (Priority: 3/5): The M5 Vision Pro update is seen as meaningful for performance, battery life, and higher refresh rates, while the new FaceTime personas are described as much more convincing and less uncanny than earlier versions. iPhone Air as a promising but uncertain form factor (Priority: 4/5): They debate whether the iPhone Air has a real market. It is praised for thinness, lightness, and battery life, but criticized for its camera tradeoffs and awkward position between the standard and Pro models.

Key Arguments: Chat-based browsing is a poor replacement for direct web navigation because it adds steps, increases error risk, and makes simple tasks slower. Atlas appears designed more to extract user context for OpenAI than to help users efficiently browse the web. AI summaries and overviews can be wrong or hallucinated, so users should still seek links and original sources. The web and App Store ecosystems were originally beneficial breakthroughs, but the current scale and power concentration have changed the regulatory picture. Apple’s App Store model is no longer simply a developer-friendly platform tax; it is now a chokepoint over global digital commerce. Apple should proactively redesign App Store distribution and payment rules before regulators force a worse solution. M5 upgrades are mostly speed bumps, but that is the right outcome for a mature platform with fast annual cadence. Vision Pro’s software and hardware progress show Apple remains committed to the platform even if consumer uptake is limited. The iPhone Air is a physically impressive device, but it is not obvious who the target user is. Apple appears to be prioritizing the Pro lineup more heavily than the Air, suggesting doubts about the Air’s long-term role.

Data Points: Sleep score (Apple Sleep Tracking): 79 and 65 - Speaker reports Apple’s sleep scores for recent nights as relatively optimistic. Sleep score (Sleep Plus Plus): 20/100 and 35/100 - Speaker’s cold-related sleep was rated far worse by the third-party app. Apple/UK damages exposure: £1.5 billion - Potential damages Apple could face if the UK ruling stands after appeal. Estimated US dollar equivalent of UK damages: about $2 billion - Converted estimate discussed in the App Store case. App Store developer revenue split: 70/30 - Classic App Store commission structure referenced repeatedly. Small Business Program split: 85/15 - Lower commission structure for smaller developers under Apple’s program. Subscription split after year one: 85/15 - Commission reduction for subscriptions after the first year. M5 Vision Pro baseline jump: M2 to M5 - Vision Pro jumps multiple chip generations, making benchmark gains more noticeable. Apple product cadence: about 1 year - The discussion emphasizes Apple’s annual chip/product schedule. Vision Pro battery improvement: about half an hour - Noted as a meaningful gain from the M5 update. iPhone Air portion of lineup: about 8% - Estimate mentioned for the share of new iPhones represented by the Air. Mini iPhone sales estimate: 5% to 10% - Rough historical estimate for the iPhone mini’s market share. Original iPhone launch year: 2007 - Used to compare how the iPhone was once a novelty.

Pivotal Quotes: "This is worse in every conceivable way." — Dan Moore / quoted Anil Dash: Response to Atlas’s demo of searching browser history through chat instead of normal browser navigation. "I think it's more for them, not for us." — Dan Moore: Summary judgment on Atlas and similar AI browser efforts. "I think the right way to make a website is to appeal to actual humans." — Dan Moore: Used to argue against designing for search engines or AI intermediaries rather than readers.

Implications: The episode suggests AI browsers may struggle unless they beat classic browsers on speed and clarity, not just novelty. Apple faces growing pressure to open App Store economics and distribution. M5 and Vision Pro show steady platform maturity, while the iPhone Air may remain a niche experiment unless Apple rethinks the lineup.