Cover Image for ML Sudo and Quintus Kilbourn | Sovereign Hardware: Hardware is broken. We are trying to fix it.
Cover Image for ML Sudo and Quintus Kilbourn | Sovereign Hardware: Hardware is broken. We are trying to fix it.
Hosted By
3 Going

ML Sudo and Quintus Kilbourn | Sovereign Hardware: Hardware is broken. We are trying to fix it.

Zoom
Get Tickets
Approval Required
Your registration is subject to host approval.
Suggested Donation
$10.00
Pay what you want
Welcome! To join the event, please get your ticket below.
About Event

Foresight Institute’s Intelligent Cooperation Group

Sovereign Hardware: Hardware is broken. We are trying to fix it.

Abstract: Every layer of the stack we're told to trust has been broken at some point: the phone in your pocket, the router in the rack, the fibre between data centres. In almost every case the public learned about it years later, from a leak or a court filing, never from the vendor. The TEE attacks of 2025 are the same story on current hardware. A passive DDR5 interposer costing under a thousand dollars pulled an attestation key out of a production part and forged the attestation that confidential AI workloads rely on.

The first half of the hour is diagnosis. What physical attacks look like, what side channels are and why they are hard to close, and why hardware assurance now matters to anyone holding model weights, operating across hostile jurisdictions, or making promises about user privacy.

In the second half, we lay out what a trustworthy hardware base requires: which pieces exist today, what primitives we are trying to conjure up, and what remains to be done.

Speaker Bio: Trusted execution environments are not as secure as we assume, and AI increasingly runs on hardware no one can independently verify. Sudo ML and Quintus Kilbourn from Project Sovereign are building secure, open, and verifiable silicon that lets users confirm a chip runs exactly as intended – no backdoors, no blind trust in manufacturers or cloud providers – reducing trust assumptions across AI coordination, health systems, safety research, BCIs, and critical infrastructure.

Foresight Institute’s Intelligent Cooperation Group

A group of scientists, engineers, and entrepreneurs in computer science, ML, cryptocommerce, and related fields who leverage those technologies to improve voluntary cooperation across humans, and ultimately AIs.

Hosted By
3 Going