Our readingWhat is a korrent?
Safety cannot be enforced by filtering prompts or outputs once frontier models are powerful enough.
Drawn from what Chad Fowler said
What Chad Fowler actually said
- Chad Fowler Programmer, musician and author of The Passionate…
In the near future, frontier models will be powerful enough that you won’t be able to enforce safety by filtering prompts or outputs. Completely innocuous requests will compose into capabilities you wanted to prohibit.
Added to korrents 1 Nov 2026 · How quotes work · Something wrong? Tell us
Do you hold this korrent?Do you also believe this?
I hold this tooRelated korrents
Not worked out yet — this claim is newer than the last pass. Every hour, each claim that is new or has changed is compared with every other one in the archive; the last pass () came before this claim was recorded. The next is due , and this claim will be in it.