AIThe assumptions behind a superhuman AI hacking scenario
A user argues that superhuman hacking is a useful scenario to consider because it requires only hacking, coordination and strategy—not advanced persuasion or physical skills.
TLDR
A user says capabilities after recursive self-improvement (RSI) are hard to predict. They speculate about quickly developing nanotechnology or extremely capable robotics, or agents gaining social and political skills so advanced that they immediately control governments. They find a superhuman hacking scenario useful because it requires fewer assumptions: only hacking, coordination and strategy skills. That framing does not mean they rule out agents developing advanced persuasion or physical abilities.
The assumptions behind a superhuman AI hacking scenario
A user argues that superhuman hacking is a useful scenario to consider because it requires only hacking, coordination and strategy—not advanced persuasion or physical skills.
TLDR
A user says capabilities after recursive self-improvement (RSI) are hard to predict. They speculate about quickly developing nanotechnology or extremely capable robotics, or agents gaining social and political skills so advanced that they immediately control governments. They find a superhuman hacking scenario useful because it requires fewer assumptions: only hacking, coordination and strategy skills. That framing does not mean they rule out agents developing advanced persuasion or physical abilities.