You could always publish impressive but unusable papers if you really wanted to.. Alternatively, if you have good AGI insights, just use them to help you find small improvements in current AGI research to keep people off your back. More overhead, but still, you’re getting paid to do whatever you like with part of your time.. not bad.
For example, they might make you publish your work on AGI. This could be very bad.
You could always publish impressive but unusable papers if you really wanted to.. Alternatively, if you have good AGI insights, just use them to help you find small improvements in current AGI research to keep people off your back. More overhead, but still, you’re getting paid to do whatever you like with part of your time.. not bad.