Gutenberg Lab — Morning Paper

F1 is the world's largest annual sporting series with 827M viewers, yet teams operated at a loss for 70 years until Liberty Media's 2017 reforms

Featuring

David Rosenthal × Ben Gilbert (Acquired Podcast co-hosts), with guest appearance by Colin Fleming (ServiceNow CMO, former Red Bull test driver)

Bottom line

F1's success required a ruthless entrepreneur (Bernie Ecclestone) to centralize power, but then required his removal to unlock modern growth — a paradox of founder-mode strengths becoming founder-mode liabilities at scale.

Three-Line Summary

Liberty Media acquired F1 for $8B in 2017 and fired 45-year ruler Bernie Ecclestone. They introduced a $170M annual cost cap that made teams profitable for the first time, doubled US fans to 52M via Netflix's *Drive to Survive*, and shifted female audience from 7% to 40%. F1 Group now has $25B enterprise value; average team valuation hit $3.6B (up 89% in two years).

Key Takeaways

1. Bernie's "8% fee" lie enabled F1's existence

In 1972, Bernie promised teams a 2% fee to centralize race negotiations but actually took 8%. Teams stayed grateful because he quadrupled their race payments overnight ($10K→$40K in Year 1). By 1981's Concorde Agreement, he secured all TV rights; by 1992, he controlled 53%+ of revenue. The deception was justified by absolute dollar growth — a lesson in how monopoly aggregators can extract rents while growing the pie.

2. Cost cap flipped 70 years of losses to instant profitability

Pre-2020, top teams spent $400-500M/year on car development and all operated at losses. Liberty's $170M annual cap (excluding drivers, top execs, marketing, power units) combined with $430M average team revenue created overnight profitability. Mercedes now generates $800M revenue with $200M operating income (25% margin). Red Bull intentionally runs at ~1% margin to maximize marketing spend for energy drink sales.

3. *Drive to Survive* created 50M fans who never watch races

The Netflix docuseries launched in 2019 (Mercedes/Ferrari initially declined participation), became #1 show in 93 countries by Season 3. US viewership grew from 500K (2018) to 1.3M average (2024), with Miami GP hitting 3.1M. Critically, only 2.5% of US's 52M F1 fans actually watch races — the majority consume narrative content, buy merch, follow drivers on social media. Oracle CMO directly attributed their $100M/year Red Bull sponsorship to DTS.

4. F1 monetizes fans at 1/18th of NFL's rate — the growth opportunity

F1 has 830M fans generating $5.5B total revenue ($7/fan/year). NFL has 180M fans generating $23B ($127/fan/year). F1 has 4.6x the fans on 24% the revenue. Primary issues: inventory (22 races vs 285 NFL games) and US penetration (1.3M average viewers vs NFL's tens of millions). European TV rights remain undermonetized (state broadcasters, no competitive bidding). Apple won 2025 US rights at $150M/year, up from ESPN's $80-90M — but still far below potential.

5. Teams have no durable competitive advantage; F1 Group has all the power

Teams' success is pure operational excellence — innovations like Brawn GP's 2009 "double diffuser" or Lotus's 1970s ground effects last one season before competitors copy. F1 Group controls: network economies (teams + tracks), FIA-designated "pinnacle" branding, switching costs, scale economies ($3.4B costs amortized across 22 races), and the cornered resource (100-year commercial rights secured by Bernie in 2001 for $360M no-bid). Current market cap: $22B (enterprise value $25B), up 5x from Liberty's 2017 $4.4B equity investment.

Editorial Take

The F1 case study illustrates a classic pattern: founder-mode strengths in the 0→1 phase become founder-mode liabilities at scale. Bernie's opacity (no contracts, personal handshake deals), dismissal of young audiences ("they don't buy Rolexes"), and tax evasion (pleaded guilty in 2023, paid £653M) were essential to navigate F1's global complexity — 10 teams, 22 circuits, 92-country broadcasters, gangster promoters, FIA bureaucracy. Unlike NFL's Pete Rozelle (a hired executive), this required an actual owner-operator street fighter. But those same traits blocked growth in the social media era. Liberty's insight was recognizing that the transition required Bernie's *complete removal*, not just "stepping back." The irony: Bernie sold F1 four times without ever losing control or truly owning it (Eddie Jordan's quote), yet Liberty had to pay $8B to finally extract him. The lesson for platform businesses: visionary aggregators can justify monopoly rents if they grow the pie faster than they extract — but know when the founder must exit.

Source: Acquired Podcast "Formula 1" (2026-03-02)

https://www.acquired.fm/episodes/formula-1

• --

AI開示文 / AI Disclosure:

本コンテンツはAIの支援を受けて制作し、編集部が事実確認・分析の監修を行っています。 / Produced with AI assistance; facts and analysis reviewed by our editorial team.

Netflix CPTO on the "Storming Phase" of AI-Driven Role Fluidity and Organizational Strategy

Participants

Lenny Rachitsky (Host, Lenny's Podcast) × Elizabeth Stone (Chief Product and Technology Officer, Netflix)

Bottom line

The scarcest resource in the AI era isn't execution capability but systems thinking—the ability to step back and design the building blocks that enable coherence across an entire organization.

3-Line Summary

Netflix CPTO Elizabeth Stone examines how generative AI is blurring traditional PM/engineer/designer boundaries in what she calls a "storming phase" of organizational transformation. While acknowledging role fluidity, she argues craft excellence (great engineering, data science, creativity) remains scarce, and the most in-demand skill across all functions is now "systems thinking"—the ability to abstract business domains into reusable building blocks. Netflix's 20+ year AI/ML track record and culture of "excellence as an operating system" (high talent density, extreme autonomy, minimal process) now mirrors how frontier AI labs operate, making Netflix's experience a leading indicator for industry-wide evolution.

3 Key Points

1. Organizational Guardrails for Role Fluidity

As PMs ship code, designers write PRDs, and engineers do product work, Stone frames this as a natural "storming phase" with transformative technology. To prevent chaos while capturing AI benefits, Netflix maintains clarity on source-of-truth data, guardrails on shipping code to production, quality review processes, and the principle that humans remain accountable for outcomes regardless of whether an agent wrote the code.

2. Systems Thinking as the Universal Top Skill

Infrastructure engineers now build common "paved paths" instead of local solutions; designers create templates and design systems; all roles step back from specific features to think about building blocks and system coherence. This shift is driven by four factors: (1) higher velocity of work, (2) more people doing unfamiliar work, (3) AI agents operating alongside humans, and (4) inability to rely on tribal knowledge at organizational scale.

3. Talent Competition with AI Labs Is About Persona, Not Just Compensation

Netflix attracts people passionate about the *application* of technology to entertainment and consumer products at global scale—a different persona from those excited by foundational model work at frontier labs. "That sweet spot between tech and product and entertainment" serves as a distinct differentiator in talent strategy.

Editorial Take

Netflix's experience reveals a paradoxical truth about AI-era organizational design: as tools democratize, the scarcity of people who can design entire systems intensifies. As role boundaries blur, the value of craft excellence within each discipline rises. This mirrors past platform transitions—when the App Store democratized mobile app development, the value of exceptional UX design skyrocketed. The fact that Netflix's culture principles from 20 years ago (high agency, autonomy, top-of-market pay, bottom-up thinking)—codified before the Netflix Prize in 2006—have now become the standard operating model for frontier AI labs suggests Netflix's organizational culture serves as a leading indicator for how the broader tech industry will need to evolve. The company's insistence on talent density as "the non-negotiable" and resistance to adding process when things go wrong may be the most durable insight: in an era of rapid technological change, organizational resilience comes not from tighter controls but from trusting exceptional people to navigate ambiguity.

Source: Lenny's Podcast "Netflix CPTO on AI and the future of product and tech roles | Elizabeth Stone" (2026-07-19)

https://www.lennysnewsletter.com/p/netflix-cpto-on-ai-and-the-future

• --

iOS 27 Beta Siri Reviews Are Premature — Third-Party Integration Blocked Until September

Speakers

Marques Brownlee (MKBHD) × David Imel, Ellis Rovin, Andrew Manganelli, Adam Leavitt

Bottom line

Current iOS 27 beta testing only reflects Siri's capabilities with Apple's first-party apps, as Xcode beta technically prevents developers from shipping apps with new entity and intent schemas to the App Store until the official September launch.

Three-Line Summary

The Waveform podcast exposes a critical gap in iOS 27 Siri coverage: the current Xcode beta blocks App Store submissions with new entity schemas (data type definitions) and intent schemas (allowed actions), meaning all circulating beta reviews evaluate Siri's performance with Apple apps only. If Google refuses to adopt these schemas for Gmail, Calendar, and Photos—which dominate iOS usage—Siri could become "half useless." The episode also dissects Nothing's pricing crisis (Phone 4b was rebranded from CMF due to RAM cost inflation) and Meta's opt-out AI training, which hosts argue violates DMCA's "good faith effort" requirement.

Three Key Points

1. RAM Shortage Killed Budget Flagship Strategy

Nothing Phone 4b (£330/€330) was originally designed as CMF Phone 3 Pro but rebranded due to memory price increases making it unsellable under $300. Specs: Snapdragon 6 Gen 4, 8GB RAM, 5200mAh battery (outside India), AMOLED 120Hz—performance matches 2019 OnePlus 7 Pro and Tensor G2. David Imel: "It's harder than ever to have that [OnePlus-style value proposition] in the world with RAM."

2. Google Holds Veto Power Over Siri's Usefulness

iOS 27's new Siri requires third-party apps to adopt entity schemas (index cards defining data types like "recipe" or "photo") and intent schemas (buttons Siri can press). Without Google supporting these for Gmail, Calendar, and Photos, Siri cannot access user data in those apps. Matthew Cassinelli (ex-Apple Shortcuts team) argues Google is "planning for a future where ads shown in front of you are not their main revenue driver," but hosts remain skeptical.

3. Meta's Invisible Watermark Is Security Theater

Meta Muse AI lets users remix any Instagram photo/profile without permission (opt-out only via Settings > Sharing and Reuse). Invisible watermark is detectable only via separate web tool, not auto-labeled in-app. Andrew Manganelli: "When misinformation gets posted and a million people see it, then the follow-up thing is like, 'Oh, this is a fake image because we found the invisible watermark'—0.005% of the million see that."

Editorial Take

This episode crystallizes the asymmetric dependency problem in platform economics. Apple demands schema adoption from developers, but dominant players like Google lack incentives to cooperate—users may switch email *apps* but not email *services* (Gmail backend persists). This mirrors past Flash non-support and third-party cookie deprecation, but with higher stakes: an OS-level feature (AI assistant) is now hostage to competitor cooperation. As David notes, "We allowed five corporations to be so deeply embedded in our lives and they keep screwing us over." Ellis's repeated invocation of "protocol" (decentralized standards) may be the only structural exit, though hosts remain skeptical of near-term viability. The RAM shortage subplot reveals a parallel crisis: hardware constraints are forcing even budget brands to abandon "flagship killer" positioning, compressing the entire Android value ladder.

Source: Waveform Podcast "Nothing Beats Phone 4b with Ear 3a?" (2026-07-10)

(unknown)

• --