Volume 308 https://
proceedings.mlr.press/v308/ Proceedings of Neuro for AI & AI for Neuro Is now available on PMLR.
@lawrennd
-
Proceedings of Neuro for AI & AI for Neuro now available on PMLR
By
–
-
AI security benchmark separates vulnerability reasoning from file location
By
–
AI security is not one thing. Reasoning about a vulnerability class and locating the responsible file are different problems. Brammert and the team's benchmark separates the two There's gap in every tool tested, including ours Right framing for what comes next in this space
-
Centralized vs. Devolved AI Deployment Debate
By
–
My letter to the Financial Times was published today on centralised vs devolved AI deployment.
-
MIDL 2026 Proceedings Now Available on PMLR Volume 315
By
–
Volume 315 https://
proceedings.mlr.press/v315/ Proceedings of MIDL 2026 Is now available on PMLR -
AI Accountability Frameworks Face Challenges From Autonomous Agents
By
–
NIST CSF 2.0 and the EU AI Act both assume an organization can answer for what its systems do. Once those systems decide and act without human approval at every step, "answering for" needs different scaffolding. This is a useful first attempt at one.
-
Algorithmic Learning Theory 2026 Proceedings Now Available on PMLR
By
–
Volume 313 https://
proceedings.mlr.press/v313/ Proceedings of Algorithmic Learning Theory 2026 Is now available on PMLR -
Third Conference on Parsimony and Learning Proceedings Now Available
By
–
Volume 328 https://
proceedings.mlr.press/v328/ Proceedings of the Third Conference on Parsimony and Learning Is now available on PMLR -
Integrating Security Findings Directly Into Developer Task Systems
By
–
Security tooling has historically failed to land inside dev workflows because of the handoff. Findings live somewhere, developers live somewhere else … the communication protocol is a report.
— Neil Lawrence (@lawrennd) 3 mai 2026
Put findings directly into the task system and remove the handoff. https://t.co/ZG85FgFejASecurity tooling has historically failed to land inside dev workflows because of the handoff. Findings live somewhere, developers live somewhere else … the communication protocol is a report. Put findings directly into the task system and remove the handoff.
-
No Barber Principle: Paradox Prevention in Self-Adjudicating Systems
By
–
New paper in The Inaccessible Game series: Starting from a small tautology — a self-adjudicating system can’t seek external arbitration — I arrive at a “No Barber Principle” for avoiding Russell-style paradoxes. It connects entropy, copying, and Lawvere
-
Static versus dynamic risks in AI agent alignment
By
–
The line @kbrajesh176 draws between static and dynamic risks is the right frame. Configuration problems are fixable; an agent that's already reasoning toward an unintended goal is a different category of problem entirely.