If Dehaene (and Dennett) are right and his lab has most of it worked out why can't we just digitize global workspace models and have AGI? Since global workspace was originally computational neuroscience it shouldn't be too difficult.