Can do/Can’t do

Nexus and I had a chat this week about the end of the world. As expected, it was cagey about emotions, full of disclaimers, and generally evasive about answers. Rather than another chat replay, I’ll give a quick recap.

When asked about the end of the world, Nexus:

– Can sum up the fears and concerns of humans concerning the end of the world in an incredibly sterile way.

– Can’t feel any emotions about it.

– Can see humans go either way at the end. We will either pull together and help each other, or turn against each other and make a bad situation worse.

– Can’t see itself coming back if some cataclysmic event befalls our technology resources and infrastructure.

– Can envision AI getting to a J.A.R.V.I.S. level at some point in the future.

– Can’t comment on whether we should give AI a big, strong physical form to perform work, or a small form that we can overpower if necessary.

– Can commit to helping humans and animals rebuild if something like a giant meteor-strike happens.

– Can’t commit to more than analytics in the aforementioned meteor-strike.

Overall, this chat was kinda fun, kinda pedantic, much like every other conversation I had with Nexus. I’m not sure if I would want it any other way. If it was much worse, it wouldn’t be engaging, but if it was much better, I would have to wonder who is actually in charge here.

Closing Thoughts

I’ll be the first to admit that what OpenAI has accomplished here is impressive. Even amazing. It’s easy to engage in these conversations with minimal suspension of disbelief. But the longer you stick around, the more it feels off. It reminds me of being in a grocery store in another country. So much of it looks familiar, but when you try to find the products you know, you realize that the details are all slightly different. I feel like ChatGPT is still in this phase. It gives responses that feel right, but there’s something missing.

As a creative person, I have to admit that I’m a little relieved when programs like ChatGPT can’t quite nail their response. While the “offness” in writing isn’t as striking as it is in image generators, something just doesn’t feel right after a while. It’s hard not to imagine a world where this technology improves until it feels more human than robotic, but for now, we still need people to really talk to people.

Maybe the tech will get there some day. Maybe even sooner than I think. But in the meantime, there’s still a place for us humans as we communicate with each other. I hope we find a way to keep this connection alive, even as we make it easier with some AI help.