Beyond ACLs, Redundancy, and Zero Trust
The original provocation: Web3 security, AI agents, decentralised threats, incident response and verifiable accountability.
Open proposal →Security talks · Systems arguments · Proposals & Thoughts...
A working archive of conference proposals, research-backed talk material and candidate submissions by @unix_jim (Inclusion here does not imply acceptance, scheduling or delivery.)
The original provocation: Web3 security, AI agents, decentralised threats, incident response and verifiable accountability.
Open proposal →A bounded-autonomy argument for machine-speed defence: which security decisions can be delegated, which must remain human, and where containment becomes the critical authority boundary.
Open candidate →A systems-history argument about abstraction, agency, constraints and responsibility: what modern AI-era computing risks forgetting about the machine beneath the interface.
Open candidate →Assembly literacy as a defensive verification skill: calling conventions, unsafe boundaries, memory assumptions, compiler output and the places where source-level guarantees stop being sufficient.
Open working draft →Representation, state, interfaces and failure were always architectural choices. AI-assisted coding and autonomous agents now turn those old choices into questions of identity, authorization and consequence. The conference version concentrates the long-form essay around that security thesis.
Open working draft →We can engineer constraints, however we cannot engineer wisdom or common sense when lucre is at stake.
"Deus nobis omnibus adiuvet."
October 2026