Leo Lin

Hardware trust is in freefall. @VitalikButerin is right — you shouldn’t have to hack your TV or disassemble your hardware wallet just to feel safe. We need open, verifiable silicon.

For a small robotics team, the host emulator could make code review less dependent on access to a shared Jetson. I'd use recorded sensor inputs in repeatable CI tests, then check timing and integration on-device.

As AI moves into the physical world, Ethereum’s openness and commitment could matter more than ever.

Enjoyed exchanging ideas at @shanhaiwoo with @VitalikButerin this week about bringing cryptography and privacy technologies into intelligent hardware.

Two directions I’m especially excited about:

-Financing physical infrastructure onchain. Open protocols could let communities fund and co-own energy systems, robotics and autonomous machines, with auditable operating data connecting capital to productive use.
-Training vertical foundation models from real-world device data. Privacy-preserving computation could help independent operators learn together, while retaining control over sensitive data and being rewarded for their contributions.

These could reinforce each other: capital deploys hardware; hardware generates useful data; better models make that infrastructure more productive.

Ethereum’s culture of openness, credible neutrality and verifiability feels like fertile ground for this next generation of protocols.

The design question I keep coming back to: how do we let people pool capital, data and learning without concentrating ownership and control?

A claim on future machine earnings only makes sense when it comes with a continuing contribution and a clear answer on the bad days.

Otherwise it’s just a permanent charge.

The useful precedent isn’t “share of revenue.” It’s risk transfer + ongoing obligation — the way Rolls-Royce structured TotalCare.