📰 Daily Digest — 2026-03-20

2 items | Business

---

📋 Quick Summary

Guillermo Rauch’s 5 Lessons for Founders Building in the AI Era

Source: a16z speedrun (Substack) · Category: Business · Link: Original

• Guillermo Rauch frames open source as a fast test for both distribution and product-market fit: if users will not adopt a free, easy-to-try product, the problem choice may be wrong.
• He argues AI lowers the cost of building so sharply that founder advantage shifts toward focus, taste, and deciding which problems are actually worth solving.
• The piece also emphasizes hiring for proof of work, packaging quality, and founder conviction, including the discipline to dismiss hype rather than chase every new tool.

Meta decides not to shut down Horizon Worlds on VR after all

Source: TechCrunch · Category: Business · Link: Original

• Meta reversed a March 19, 2026 plan to make Horizon Worlds web-and-mobile-only on June 15 and will keep VR support on Quest after all.
• The article presents the reversal as a sign of weak metaverse economics, citing Reality Labs losses of $73 billion since 2021 and Quest sales down 16% year over year from 2024 to 2025.
• Horizon’s mobile traction looks better than VR, with 45 million lifetime downloads and 1.5 million downloads so far in 2026, but consumer spend is still only about $1.1 million.

---

📝 Detailed Notes

1. Guillermo Rauch’s 5 Lessons for Founders Building in the AI Era

1. Open source is presented as an execution filter, not just a philosophy.
   • Rauch says startups have to solve two hard problems early: getting discovered and building the right thing.
   • In his framing, open source stress-tests both at once because adoption friction is low and feedback is immediate.
   • If users still do not engage when code is free and easy to try, that is a signal to reconsider the product direction.
2. Openness raises copy risk, but he treats that as a secondary concern before fit.
   • He acknowledges that agents and modern tooling make replication easier than before.
   • His counterpoint is that strong open source products often expand into broader platforms with adjacent paid problems to solve.
   • Next.js is used as the example: the framework itself is only a narrow slice of the larger enterprise cloud stack.
3. Vercel’s early history is used to argue for bounded execution under an unbounded vision.
   • Rauch describes a tension between long-term ambition and the reality of operating as a very small team.
   • Early Vercel tried to support every language, runtime, framework, and even databases.
   • The eventual focus on being excellent at the front-end layer is framed as the move that created a real business instead of a broad but weak product.
4. AI changes the constraint from building capacity to choosing capacity.
   • Rauch argues founders are now at risk of overbuilding because many tasks feel cheap once an AI system can generate them.
   • He contrasts a high-value use case, an agent improving JavaScript performance by 20% to 40%, with low-value internal busywork like rewriting HR tools.
   • That makes prioritization and problem selection more important, not less, as technical execution gets cheaper.
5. Hiring and judgment are treated as product skills as much as management skills.
   • He recommends hiring people who are actively shipping strong work, especially with modern AI tooling, rather than only optimizing for pedigree.
   • He also emphasizes how candidates package and communicate their ideas, comparing that polish to Apple’s full-stack product experience.
   • The closing lesson is conviction: experiment broadly, but only go all in on things you understand deeply enough to defend and deliberately choose over the surrounding hype.

---

2. Meta decides not to shut down Horizon Worlds on VR after all

1. Meta reversed course on Horizon Worlds VR support within the same day.
   • According to the report, Meta had indicated earlier in 2026 that Horizon Worlds would stop being supported on Quest headsets.
   • A March 19, 2026 community-forum update said the app would become web-and-mobile-only on June 15.
   • Later that day, CTO Andrew Bosworth said Meta had decided to keep Horizon Worlds working in VR, and a company spokesperson confirmed that reversal to TechCrunch.
2. The article treats the episode as evidence of how far Meta’s metaverse thesis has weakened.
   • Horizon Worlds was once central to Meta’s vision for social VR after Facebook rebranded to Meta in 2021.
   • TechCrunch argues the near-shutdown itself signals that the original VR-social strategy failed to reach meaningful demand.
   • The piece points to Reality Labs losses totaling $73 billion since 2021 as the clearest financial indicator of that failure.
3. Hardware momentum in VR still looks weak relative to mobile.
   • IDC data cited in the article says Quest headset sales fell 16% year over year from 2024 to 2025.
   • The story also notes Apple scaled back Vision Pro production, reinforcing that the category-wide demand problem is not unique to Meta.
   • Meta has already cut more than 1,500 Reality Labs roles in January 2026 and is reportedly considering broader layoffs that could affect 20% of the company.
4. Meta’s product logic now centers on mobile because that is where usage is actually growing.
   • Bosworth said on a podcast that Horizon had shifted focus to mobile because the product had stronger product-market fit there.
   • He also argued that supporting both phone and VR versions forces the team to build everything twice, which slows velocity.
   • In that framing, preserving Quest support is no longer the main growth story even if VR access remains available.
5. Mobile metrics show traction, but monetization remains tiny relative to investment.
   • Appfigures data cited by TechCrunch estimates 45 million lifetime Horizon Worlds downloads across iOS and Google Play.
   • The app reportedly reached 1.5 million downloads so far in 2026, up 53% year over year from roughly 983,000 at the same point in 2025.
   • Even so, total consumer spending is estimated at only $1.1 million, which underscores how little direct revenue exists against Meta’s metaverse spending.