Having previously (see here and here) assayed the conundrum of deflecting curious future explorers from spelunking in our present nuclear waste, we were intrigued to see what sort of out-the-box solutions our artificial intelligencer Janelle Shane (see previously) might be able to coax out of her neural network to serve as a ten-thousand year warning. Summarising the schema to the GPT-3 routine, its designs seemed to match that of its human engineering inspiration to intrigue as much as dissuade any future civilisation. Giant tube worms and a field of Tulips Shrieking Madness might deter exploration but I am not sure about Dangerous Stairs or Disrupted Pollen Lines. Much more to explore at the links above.