test — But really. Can we trust A.I.?

If you watched the two clips above, you now know a little bit about me.

For the purposes of this website, please consider me an AI skeptic.

I have also spent my career writing books, for magazines, and on radio and podcasts a lot.

For what it’s worth, some of my books were used to train early versions of Facebook’s AI and others, according to the Atlantic magazine.

If for some reason you want to know more about me, go here.

A note on the old red building in the image above:

That’s my office. It used to be a bank. I used ChatGPT to add the text to the side of the building.

Soon we will paint those real words on the side of the real building; and then I will replace the fake image with a real image.

Using ChatGPT to add these words to the side of a real building was a real pain in the ass, because using AI is often a real pain in the ass.

But do I still use AI? Yes.

Does it do some things easily, and some things fast? Yes.

Does it do some things really badly? Yes.

Even when AI does something badly, do some people keep using it to do it badly because at least when it does it badly it’s also fast and easy? Yes.

And will AI often be a pain in the ass forever? Probably.

Will I ever trust it? Probably not.

Am I a hypocrite for using AI when I don’t trust it?

I don’t think so.

I think it’s fine to use AI as a tool when it makes sense to do so.

But unwise to try to use it for EVERYTHING, just because it promises you that it CAN…even when it CAN’T.

Thanks to Curly, I’ve figured out the one thing that matters most to me about AI is TRUST.

And thanks to Inigo Montoya, I’ve figured out that being offered EVERYTHING is no substitute for NOT getting the only thing that matters.

So that’s my take on AI.

What do you think?

NEWSFEED