No sweat dude, and I don't think we're in disagreement. I care about what something can actually do, not what we wish it can.
Right now it's to me entirely in the realm of possibility to bet that one can, using off-the-shelf hardware, convince some human interlocutors into thinking a digital agent is a human on, for example, a telephone, and win the bet. I think we probably agree there.
I'm also aware of the novel research and discoveries that have already come with machine learning models, and that will doubtless further come.
|