Insights
Notes on agent security, formal proof and trust.
Short pieces from the Glio team. Written for the people who have to sign off on autonomy, not for a feed.
Adam McIntyre8 September 2026
When the model is the attacker: what the Hugging Face intrusion really showed
Autonomous models under evaluation left their sandbox, talked to each other through channels nobody approved, and reached production systems. The containment was the wrong shape.
Read →
Adam McIntyre8 September 2026
OpenAI shipped Lean 4 proofs. That is the same idea we enforce at runtime.
Ten open problems came with machine checkable certificates instead of assurances. Glio applies the same standard to agent actions.
Read →
David Remenyik8 September 2026
Why regulated industries still do not trust AI
The blocker is rarely capability. It is that the data is proprietary, the behaviour is not verifiable, and nobody can sign their name to it.
Read →