David Singleton(@dps):One threat we’re particularly focused on is prompt injection, and we handle it in layers. The model is trained to recognize and resist it. The harness marks anything coming from an untrusted source. Deterministic code checks the result. And an ensemble of classifiers runs where the agent can't reach them.

David Singleton

@dps

Meta Superintelligence Labs @meta formerly CEO & cofounder at @dreamer, CTO @stripe. ❤️ @fjsingleton. I like to make things.

加入 March 2007

1.5K 正在关注 25.4K 粉丝

David Singleton@dps

2026.09.08 19:18

One threat we’re particularly focused on is prompt injection, and we handle it in layers. The model is trained to recognize and resist it. The harness marks anything coming from an untrusted source. Deterministic code checks the result. And an ensemble of classifiers runs where the agent can't reach them.

显示更多