Closed labs can patch guardrails to block exploits; open weights can't
AlexBarry4 · x · 2026-10-22
In the ongoing debate on open-model cyber capabilities, xeophon notes that publicly accessible models already match closed ones at finding CVEs and reading code (php2shell), but closed models can — and did — update their guardrails/filters to block exploit generation for the broad public, a key asymmetry when assessing open-weight risk.