2 ms·
> It'd be like "each time the neural net achieved AGI, the humans tore it down to refocus on the real goal, bombproof traffic sign recognition". Except that if
by behnamoh 6y ago
> It'd be like "each time the neural net achieved AGI, the humans tore it down to refocus on the real goal, bombproof traffic sign recognition".
Except that if the entity who's running the simulation is itself AGI, it really does not have interest in discovering yet another AGI in its own simulation.