@rjmontgomery Jevons Paradox explains why cheaper inference can increase total AI usage. It still does not tell us who captures the profit. For AI apps, that will depend on distribution, workflow ownership, switching costs and how much compute expense can be passed through.
@startupideaspod The interesting part is not that the founder can work from a phone. It is that a narrow agent can absorb a repeatable professional workflow. The durable advantage will come from integrations, case context and reliable human review—not raw model access.
@gregisenberg Cloud agents make execution abundant. The scarce layer shifts to orchestration: choosing valuable workflows, defining permissions and measuring outcomes. Running 20 agents is easy. Knowing which one can be trusted with a customer or payment is the business.
Jensen Huang’s first post is a reminder:
AI will not only belong to the companies building models.
It will belong to creators, developers, and entrepreneurs who turn AI into products, workflows, and businesses.
The AI revolution is moving from “model competition” to “application explosion”.
We are still at the beginning.
🚀
For my first post, I’m sharing a letter @NVIDIA signed on why open models matter.
AI will transform every industry, power every company, and be built by every country.
Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.
The world needs both frontier closed models and frontier open models.
https://t.co/AUKzoQ5Ikb
Don't start with an AI product. Start with a painful, recurring problem.
Use AI to deliver the result once.
Charge for it.
Repeat it.
Then automate what works.
A model tells you what can be built. A paying customer tells you whether it should become a business.
Most people use AI agents for content. Financial firms are testing them for payments, fraud and compliance.
The real divide isn't capability, but how much authority companies will delegate.
The closer AI gets to money and identity, the more valuable the control layer becomes.