Ashlee Vance on the Meta-Anduril reconciliation, Palmer Lucky's VR toybox returned, and Core Memory's global expansion

Summary

Ashlee Vance's Palmer Lucky interview — conducted the morning of this segment — is the centerpiece. Lucky appears to have come in straight from his meeting with Zuckerberg, wearing the same clothes visible in the photo the two posted together, marking a public reconciliation after what Vance describes as a nine-year war between Lucky and Meta.

The Anduril-Meta deal

The partnership pairs Anduril with Meta on military VR — an extension of the IVAS contract Anduril took over from Microsoft. The official press release and the Wall Street Journal's coverage were both thin on technical detail; Lucky went deeper on the Core Memory podcast. Beyond the display and AR data-feed layer, Lucky wants to reinvent the helmet itself: lighter materials, improved ballistics protection. The ambition is a genuine next-generation warfighter helmet, not just a screen bolted onto existing hardware.

On the personal falling-out, Lucky told Vance he doesn't place the blame squarely on Zuckerberg. His read is that Zuck was responding to pressure from within Facebook and doing what he felt he had to do to protect the company. The sharper structural mistake, in Vance's view, was that Meta never gave Lucky a direct line to Zuckerberg after the Oculus acquisition. Multiple layers of management sat between them at the exact moment Meta was betting heavily on VR through Reality Labs. Vance's inference — not stated by Lucky directly — is that Lucky probably stays at Meta working on VR if that reporting structure had been different.

Meta's broader product position

Vance flags Meta's Control Labs acquisition as underreported. Control Labs reads motor neurons in the wrist as a brain-computer interface proxy, and Meta has demoed body-based navigation. With OpenAI and Jony Ive moving toward a new device category, Vance sees Meta's wrist-based interface work as a serious, if difficult, bet on what comes after the smartphone.

On the hardware side, the expectation is that the next Quest headset will benefit from display technology Apple pioneered for Vision Pro — technology that Apple developed but failed to scale to mass-market volumes. Meta could absorb those advances at consumer price points two years later.

Llama 4, by contrast, has slipped on benchmarks. The read is that Meta leaned too hard into pre-training and underinvested in reinforcement learning.

Core Memory's expansion

Vance's media company, Core Memory, has moved beyond its Silicon Valley base. The team just finished several weeks of filming in Switzerland, covering university robotics labs at EPFL and ETH — both heavily backed by the Swiss government — and a handful of Swiss startups. Vance's observation from those labs matches a pattern he sees across Europe: strong research output, weaker translation into funded startups with the ambition levels common in the US.

The bigger Core Memory project in progress is a feature-length documentary on brain-computer interfaces, with Neuralink at its center. Vance is tracking that story over months, potentially years.

Europe and the AI infrastructure gap

Vance visited Abilene, Texas with Sam Altman and Greg a few weeks prior — touring what he describes as buildings costing roughly $50 billion each, of which there are many. He gave a talk in Poland shortly after, and found European audiences largely unaware of the scale of US AI infrastructure investment. His read from conversations with Altman is that France is currently the only European country positioned to participate in Stargate at a meaningful level, with those negotiations ongoing. Nvidia, Crusoe, and other Stargate supply-chain companies are actively seeking to extend the buildout internationally.