Thesis · Episode 013
Marketing, Compute, & Intelligence
we look at the hardware plateau, the software race on top of it, and why the transaction model for entertainment robots may stay cleaner than the one for the home.
October 27, 2026 · 46 min
The Argument
Treat Marketing, Compute, & Intelligence as a brief, not a news item. We look at the hardware plateau, the software race on top of it, and why the transaction model for entertainment robots may stay cleaner than the one for the home. Compute gains are slowing.
Physics and manufacturing capacity are limiting the next five to ten years to marginal improvements rather than generational leaps. The bigger near-term gains look like distillation, optimization, and more efficient software running on hardware people already own.
Photonics still has to clear the conversion layer before it delivers usable speed. Open architectures like RISC-V are appealing but face the practical problem of who will fund the fabs.
Listen to this as the next house rule. Mechanism is allowed to be miraculous only after the performance has something true to stand on.
In This Conversation
00:00
Studio
We look at the hardware plateau, the software race on top of it, and why the transaction model for entertainment robots may stay cleaner th…
10:13
The claim
Physics and manufacturing capacity are limiting the next five to ten years to marginal improvements rather than generational leaps. The big…
22:18
The mechanism
Photonics still has to clear the conversion layer before it delivers usable speed. Open architectures like RISC-V are appealing but face th…
33:27
The stake
We also touch the rumor of a fifth form of compute that sits outside quantum, photonics, bio, and traditional silicon. That hardware pictur…
Takeaways
- We look at the hardware plateau, the software race on top of it, and why the transaction model for entertainment robots may stay cleaner than the one for the h…
- Physics and manufacturing capacity are limiting the next five to ten years to marginal improvements rather than generational leaps.
- Open architectures like RISC-V are appealing but face the practical problem of who will fund the fabs.
Transcript
The conversation as recorded. Speakers are unlabeled on purpose.
Open the full transcript
More
Also In The Conversation
- Thesis · Episode 012Noveltywe examine novelty in the age of generative tools — what gets built quickly, what actually gets used, and why character robots are starting to leave the lab.
- Thesis · Episode 011The Attention & Entertainment Economywe look at the attention and entertainment economy — live events, shared experience, and why people are leaving the house again.
- Thesis · Episode 010Immersive ExperiencesMost immersive rooms still work like a gallery: lights happen, nothing asks you back. Story is what makes a place worth returning to.
If you are in this work, write us.