After a major exchange froze withdrawals on me, I moved everything to self-custody and never looked back. Two years of managing my own keys, approving every transaction manually, switching networks by hand. Annoying but honest. I knew exactly what was happening to my capital at every step.
Genius Terminal was the first platform I trusted enough to change that habit. Non-custodial, keys stay with me, but the execution experience was cleaner than any CEX I had used. I did not have to choose between owning my assets and trading them properly. Or so I thought.
That assumption did not survive long enough.
Self-custody protects assets at rest. It does not protect assets in motion. On a platform designed so that assets are always moving, always being routed, bridged, authorized through layers I cannot see, my private key is the last line of defense for the one moment my capital is not actually moving. Which on Genius Terminal is almost never.
The attack surface is not my wallet. It is everything between my wallet and the outcome I see on screen. Unlike a compromised private key, where the damage is immediate and visible, a compromised execution layer is subtle. A manipulated routing decision. A bridging delay at the wrong moment. None of that touches my key. All of it touches my capital. And I would have no way to distinguish a platform-level failure from bad market conditions until the position is already settled.
This is the risk self-custody cannot price in. Counterparty risk did not disappear on Genius Terminal. It moved downstream into the execution layer, where it is harder to see, harder to monitor, and harder to act on in time.
The better Genius Terminal gets at absorbing execution complexity, the wider that gap becomes. Every UX improvement is an improvement in the distance between me and what is actually happening to my capital. Self-custody and execution safety are not the same conversation. On Genius Terminal, they never were.