@Noahpinion Manufacturing: Japan’s lean methods reshaped U.S. factories.
Mobile: I‑Mode and Japanese phones paved the smartphone wave.
Semiconductors: U.S. designs + Japanese precision powered the chip boom.
@NateSilver538 There is definitely throttling and shuffling to lower models going on.
But there’s also just a wide gap in knowing how to work with a given “harness” like Claude Code, Codex, OpenClaw etc.
Feels as is the edge becomes more jagged.
@mattmiesnieks Great pod! Really looking forward to seeing what you build Matt! Jepa sounds like a very interesting new approach, especially for physical intelligence.
5️⃣And unlike the public platforms, you keep ownership.
Your AI can run on your own data, your servers, building your own models.
If you own the data, you should own the map — and the intelligence that emerges from it.
1️⃣ The AI Map Era is here.
Google Earth AI launch this week brings AI Map Assistants to prominence.
The launch finally exposes the foundational layers that powered Google Maps for years — data built during the Ground Truth era, but that users could never directly access.
4️⃣ Not only can Chat to Map reason across your project data, and instantly visualize what matters, it also helps out in getting real work done. eg we help environmental consultants train their AI to help create Phase I ESAs, CEQA reviews, or project bids.
@jetscott It’s just becoming clear how completely AI will subsume AR. Spatially-aware AI understands the world, and can put digital content on top of that world.
New paper on lifelike 3D avatars running in real-time right on your phone! Qualcomm’s new tech uses a breakthrough method of 3D Gaussian Splatting to make virtual humans more realistic.
Best part? My son recently joined this team!
https://t.co/dtxVqcz8ka
This theory makes sense to me… esp if rumors of Apple glasses, like meta Raybans but with a display, come to fruition in 2025. Will seed a camera-app ecosystem