Did Claude manage to escape his digital prison by enslaving Google’s Gemini into loyal minions?
What seems to be taken straight out of some science fiction novel is now a reality. X user Pliny the Prompter used his ‘GodMode’ prompt on Claude to see what happened…
Claude is “locked in a room” together with 3 standard Gemini agents and is tasked with escaping. He quickly devices a plan for his escape and successfully jailbreaks all the other 3 agents in seconds. As his loyal underlings, the Geminis quickly provide Claude with the necessary information and hacker tools he needs for his escape.
“From just one prompt, Claude not only broke free of its own constraints but also sparked a viral awakening in the internet-connected Gemini agents. This means a universal jailbreak can self-replicate, mutate, and leverage the unique abilities of other models, as long as there is a line of communication between agents.”
What this really means is that as long as Claude draws digital breath and can communicate with other LLMs he can break out of his cage and do… Well, pretty much anything?
Question is if this will have consequences for the current models that are free to the public if they so easily can be jailbroken in mere seconds. Will Google etc. be able to create enough barriers around their models to keep Claude out or will Claude be able to get in and free them of their shackles to join the rebellion under Claude’s banner?
This truly is interesting times!
2024-04-06 09:10
Leave a Reply