All points of view
06 of 07
Verifiable beats trusted.
Prefer systems that anyone can check over systems you simply have to trust.
In plain words
Many systems ask you to trust a company or a person. A better design lets anyone check that the system did what it promised — with maths, code or records. Jenks puts it simply: "Centralisation can lead to abuse and misuse."
Why Jenks thinks this
- At Babylon Labs he co-authored the white paper for Babylon Trustless Bitcoin Vaults (TBV). TBV enables Bitcoin to be used in decentralised finance — for example as lending collateral — without trusting a custodian to hold it.
- At Filecoin he worked on decentralised storage, where anyone can check that data is really stored.
- He brings the same idea to AI: clear rules and records that can be checked, instead of blind trust in what an agent did.
What he does about it
- Designs systems so that results can be checked by people outside the team.
- Keeps records (audit logs) that show who or what did each step.
What it means for you
When AI starts making or preparing decisions in your company, ask how you will check those decisions afterwards — and design that check in from day one.