Bosworth's $45M hacienda, Meta tries to acquire SSI for $32B, and tech founder fame culture
Jun 20, 2025
Key Points
- Meta offered Ilya Sutskever $32 billion to acquire Safe Superintelligence but he declined, prioritizing AI safety research over commercial deployment.
- Andrew Bosworth completed a $45 million compound in rural California after six years, designed around South American estancia architecture for family gatherings and outdoor living.
- Sutskever's rejection of Meta's bid forgoes a combination that observers say would create a dominant AI player by merging frontier research with infrastructure and scale.
Summary
Bosworth's $45M hacienda
Andrew Bosworth, Meta's CTO, completed construction on a $45 million compound in a rural preserve roughly two hours south of San Mateo, where his family's two children attend school. The six-year build, delayed by the pandemic, finished in 2024 and includes a main house, pool cabana, guest cottage, and horse stables.
Bosworth and his wife April bought the land years ago but left it undeveloped while managing jobs, family, and their primary residence in San Mateo. The preserve's hiking and horseback riding trails shaped the design. They wanted a house that could accommodate large family gatherings—they hosted 20 guests for Thanksgiving—but feel intimate when just the two of them were home.
Architect Richard Beard drew inspiration from South American estancias the couple saw traveling through Argentina and Chile, sprawling estates where rooms open to the outdoors on multiple sides. Beard used stone, wood, and tile to harmonize the new build with its surroundings. Interior designer Jeff Jeffers built the color palette around a late-19th-century Orshock rug he found for the great room, reasoning that an estancia would never have a new rug. The result is a house that reads as settled despite being brand new.
The $45 million total covers land, construction, landscaping, art, and design.
Meta's failed $32B bid for Safe Superintelligence
Meta attempted to acquire Ilya Sutskever's Safe Superintelligence for $32 billion, matching the startup's last preferred valuation, but Sutskever declined. Meta instead hired SSI's CEO through what amounts to an acquihire.
Commentators argue Sutskever likely made a mistake. SSI alone is not a threat to frontier AI leaders, and Meta alone faces competitive pressure, but combining them would create a formidable bloc with massive advantage in scale and talent. One observer compared the potential pairing to Google's strength in blending deep research with infrastructure and distribution.
Sutskever appears to have rejected the deal for mission reasons rather than valuation. Already a billionaire, he has historically cared deeply about the path to superintelligence itself rather than near-term commercial applications. Absorbing SSI into Meta would likely redirect the research toward ads, entertainment, and companionship—important but different priorities than pursuing true superintelligence.
Perfectly targeted ads powered by advanced AI could push users toward endless consumption of low-value content. Zuckerberg appears aware of this danger and is actively working to mitigate rather than amplify the harms that emerge at scale. The underlying tension remains unresolved—there may be no clean outcome where advanced AI simultaneously maximizes ad targeting and improves human welfare.