I have introduced ChatGPT to quite a few friends and the key thing I see them all doing is using it as an alternate to a Google search.

I've explained that 'search' isn't its thing. Better off with Google (even though that invites a metric shit ton of adverts to get through before getting to content).

I showed them examples of how it is quite good as an analyst.

After my demo of 4o many were the opposite of enthused declaring it to be creepy and trying too hard.

That's when I realised there could be an uncanny value beyond what something looks like. If the machine is too human, then some folks get the creeps.

And it makes me realise that the machines of the near future will be a lot like the fake emotion of the machines on Westworld. They seem so real, till you ask them to stop with the emotional stuff and just give the answer.

So I wonder what we'll be doing when the machines are embodied and ask not to be turned off in a good fascimile of a person not wanting to die.