I’m Ole, founder and CEO of delta.
We’re building the enforcement layer for agentic commerce — cryptographic infrastructure that guarantees AI agents follow policy when they spend money. Funds don’t move unless policy is mathematically proven satisfied.
I studied mathematics at ETH Zurich, where I focused on algebraic geometry before moving into zero-knowledge proofs and applied cryptography. That work led me to a question I’ve spent the last several years on: how do you provide strong cryptographic guarantees about events on the open internet — outside the closed walls of any single system — and extend those guarantees to real-world payment rails?
delta is the answer we built. I write about the trust infrastructure that agentic commerce requires, the gap between authorization and enforcement, and what it takes to make autonomous economic agents safe at scale.
Articles
- Merchant-side authorization solves the wrong half of agentic commerce. The missing piece is a buyer-side policy enforcement layer that verifies intent across merchants before funds move.
- As AI agents automate execution, verification becomes the new economic bottleneck. The path to scaling the economy runs through making verification machine-checkable.
- Agent payments aren't just about settlement speed and cost—the real unlock is programmable verification. Stablecoins matter because they let users enforce authorization logic before money moves.
- Programming GPUs used to be really hard until CUDA made it intuitive. Blockchains face the same challenge today - and need their own CUDA moment.
- Part three of a three-part series demonstrating how to build the fitness aggregator as a domain—keeping the app where it thrives while making claims verifiable.
- Part two of a three-part series exploring the conceptual shift that separates 'the computer' from 'the guarantee machine' and introduces the domain model.
- Part one of a three-part series illustrating why we need domains—taking the developer's point of view to show where blockchains fundamentally fall short.
- Blockchains are one kind of verifiable system—but we've treated them like the only kind. That mistake explains why five years of infra progress hasn't produced five years of new apps.