The Adjacent Brief

TL;DR: SAP's new API policy bars partners from routing its data into third-party AI systems, putting enterprise software integrators in a difficult position as AI becomes the default middleware layer. Parent coalitions in Salt Lake City and New York City secured rollbacks of school technology policies. HYROX's path to a billion-dollar valuation, built almost entirely on community word-of-mouth, offers a counterpoint to the week's platform-and-lock-in stories.

Worth Reading

• Package design as a form of consumer deception — Seth Godin on snack packaging that overpromises by design; a short read with long implications for brand trust.
• Driverless car optimism rises, but ownership intent stays flat — 31% of adults expect autonomous cars to be common within five years, up from 19% in 2018 — but the share who'd actually own one hasn't moved. Aspiration and adoption remain separate questions.
• Movie theaters are raising prices into strong demand — Marginal Revolution asks whether theaters are finally pricing like a real market rather than a nostalgia subsidy.
• MAHA wellness culture is reaching teenagers, and adults aren't keeping up — Vox tracks how RFK Jr.'s food-purity framework is spreading through teen social networks faster than any adult institution has a response for it.
• Renewable energy just ended a century-long fossil fuel streak in the US grid — A structural milestone in the power mix, relevant context for the Solaria raise and the data center energy story.
• The real and the fake, from Orson Welles to the Crunchwrap — A sharp essay on authenticity, representation, and why the question "is this real?" has never been harder to answer.
• Creator payment infrastructure is still broken for most independent operators — Ownersnotrenters on the gap between creator revenue and the tools available to collect it reliably.

Brand & Growth

Community as distribution: what HYROX actually built

HYROX didn't buy its way to a billion-dollar valuation. The indoor fitness competition format grew by making participants the marketing channel — every race entry was also a social post, every finisher time a recruitment tool for someone's friend group. As the Huddleup breakdown of how HYROX became a billion-dollar business makes clear, the company spent almost nothing on paid acquisition. The product was designed so that completing it made you want to tell people about it, and the social proof was embedded in the format itself: a shared leaderboard, standardized distances, wearable race numbers. This is the architecture that CrossFit briefly had before it fragmented — but HYROX kept the competitive structure standardized enough to stay legible across markets.

The lesson for brand teams isn't "spend less on marketing." Product design and community architecture can do work that media spend cannot. When the unit of participation is inherently shareable — a race time, a rank, a finisher photo — the distribution problem solves itself. That's harder to replicate than it looks, but it's also more durable than any paid channel.

Lock-in dressed as policy

The more consequential brand-and-growth story today is SAP's. The company's new API policy effectively prohibits partners from using SAP APIs to feed data into third-party AI systems — a clause that has integration consultants and ISVs alarmed. The Register frames this as a lock-in play, and that's the right read. SAP's enterprise customers have spent decades building workflows on top of its data layer; the new clause makes it significantly harder to route that data to Claude, GPT-4, or any external model without going through SAP's own AI stack.

Watch this not just as a SAP story but as a pattern. Enterprise software vendors with deep data moats are realizing that the AI layer is where the next decade of margin lives. If you control the data and the AI interface, you control the relationship. SAP is doing openly what others are doing through policy language before customers notice it's happened. Partners who built on the assumption that enterprise AI would be model-agnostic should treat this as a concrete prompt to audit their contractual exposure.

Connected World

The robot demo problem, and the one that cleared the bar

Most robot demonstrations are staged to hide failure modes — constrained environments, pre-positioned objects, cherry-picked tasks. Will Knight's piece in WIRED on when robots have their ChatGPT moment is worth reading precisely because it takes that skepticism seriously and then reports a demonstration that held up anyway. The focus is on manipulation — specifically, dexterous hand and gripper work — which remains the hardest unsolved problem in physical AI. Eka's decision to prioritize robot hand manipulation over full humanoid development reflects the same judgment: the bottleneck isn't bipedal locomotion, it's the ability to handle objects reliably across variable conditions. When that problem cracks, the scope of what robots can do in warehouses, kitchens, and labs changes faster than most deployment timelines assume.

Solar capital flows to where the demand actually is

Spain's Solaria raised €300 million to co-locate solar and battery storage with data centers, which is a more strategically legible raise than most infrastructure rounds. Data center operators need power that's contractually guaranteed and physically proximate — grid dependency creates both cost and reliability risk at the scale hyperscalers are building. Solaria is betting that co-location solves both problems simultaneously. The raise comes as renewable generation crossed a threshold in the US grid for the first time in over a century (covered in Worth Reading), which means capital is now being deployed around the energy transition rather than just anticipated.

Design students imagined a post-slab phone. Google watched.

The ecal x Google collaboration on phone form factors beyond the touchscreen slab is less interesting as a product story than as a signal about where Google thinks the design conversation needs to go. Student design briefs partnered with a platform company usually mean the platform company is stress-testing assumptions it can't test internally. The concepts themselves — wearable, modular, ambient — have circulated for years. What's notable is that Google is publicly associating its brand with the question rather than an answer. Whether that leads anywhere concrete is a separate matter.

The New Consumer

Parents are pulling the lever, and it's working

Parent groups in Salt Lake City and New York City have successfully forced rollbacks of school technology adoption policies, per the Times — not through legislation but through direct pressure on school boards. The wins are specific: reduced screen time mandates, phone-free classroom rules, delays on AI tutoring tool rollouts. What's useful here isn't the outcome but the mechanism. Organized parent coalitions are proving more effective than regulatory bodies at setting the pace of institutional tech adoption, at least at the district level. Institutional AI policy tends to move faster than the communities it affects are ready for, and the friction shows up at the most local level of governance available.

The screen-fatigue thread running through this story connects to something worth watching. MAHA wellness culture, which Vox covers in Worth Reading, is now reaching teenagers through social networks rather than through adult-mediated health messaging. The same parents winning phone-free classrooms are navigating kids who've independently adopted raw-milk and supplement frameworks from TikTok. The concern about technology and the concern about information quality are both landing at home, even when the parents and the teens are coming from different starting points.

The homeownership aspiration gap widens

Fewer Americans expect to buy homes than at any point this decade, according to new polling from Semafor. Survey data measures aspiration rather than behavior, so treat the number as a sentiment indicator rather than a forecast — but the direction is meaningful. When homeownership moves from "deferred goal" to "not part of the plan," downstream consumer behavior shifts in ways that compound: furniture, renovation products, appliances, neighborhood stability, and retirement planning all change. Brands built on the assumption that consumers are pre-homeowners waiting for their moment should be stress-testing whether that aspiration still operates the way their marketing was designed around.

Machines & Minds

Theology at the guardrail

Chatbot developers are recruiting theologians as formal contributors to AI alignment work, embedding religious and moral traditions into safety design rather than treating ethics as a purely secular engineering problem. Quartz covers the Silicon Valley turn toward religion with appropriate skepticism — but the more interesting question isn't whether tech founders have found God, it's whether the move is substantive or reputational. If safety teams are genuinely incorporating centuries of applied moral philosophy into how they think about harm, refusal, and value hierarchy, that's worth taking seriously. If it's a credentialing move — "we consulted theologians" as a liability hedge — it's a different story entirely. The distinction matters because the AI safety field has a persistent problem distinguishing between process legitimacy and actual outcome quality. Adding a theologian to a safety board doesn't automatically resolve ambiguity in edge cases; it adds a perspective that may or may not be operationalizable.

The software factory that doesn't stop compounding

The Refactoring piece on the compounding software factory makes an argument that's been building in the developer-productivity conversation for several months: teams that treat AI-assisted development as an ongoing system — where each cycle of code generation, testing, and refinement feeds the next — are operating at a different velocity than teams treating AI as a faster autocomplete. The compounding framing is doing real work here. A team that ships faster and learns from each cycle structurally diverges from one that ships at the same speed with better tooling. The gap compounds quarterly. For product leads watching competitive dynamics, the question isn't whether to adopt AI development tools — it's whether your development culture is set up to learn from the output or just consume it.

11 articles across 4 themes · 10 sources · Powered by Folo + Claude