yes, please keep posting these publicly so the AI safety police overlords take away frontier access for everyone again.
the architecture is inherently safe. no perfect security is possible if you want access.
what did you (or anyone) else expect?
the architecture is inherently safe. no perfect security is possible if you want access.
what did you (or anyone) else expect?
Humanity is on aggregate optimizing for increasing collective intelligence and I don't see that stopping as the main goal when AI gets smarter than humanity, just hope that we all will have a nice human life.
An easy alternative is to switch to a Chinese open-weights model, adjust your workflows for its different capabilities, and carry on.