Allow me to rephrase.

Shouldn’t the algorithms be trained to assess whether the answer they are providing is within the
capabilities they have been trained to have, before the algorithm gets to be labeled AI

That’s some fine looking granite, as I like to say, and a fine landscape photograph as well.

Once I had the thought to go see granite in every state that has it, but it’s actually present in more eastern states than one might guess

So does that make the days between Black Friday and cyber Monday as scandium and yttrium

Asking for a friend

There's solid evidence that a FFP2/N95 masks keep filtering for at least 40 hours. By then, they're probably starting to smell even if the straps haven't broken but as you say, the reuse makes them a lot more affordable.

www.sciencedirect.com/science/arti...

Congrats!

I’ve enjoyed your posts and pictures

Yes. If the tool were intelligent it would realize not to try to do things it’s not made to do.
Like count the number of A letters in artificial intelligence and come up with four, or six!

An AI guy told me the token involved, and so of course it gets it wrong. Shouldn’t the AI know this itself?

Strongly agree. Here's my review from last May: thetyee.ca/Culture/2025...