4 Comments

not surprised, but still astonishing.

Expand full comment

Of course, it could all be a parlour trick and it’s using the photo metadata ;-)

Expand full comment

This is good stuff.

It does seem to draw the wrong conclusions from some observations though - the building style and electricity pylons. It’s more that it’s cheap and easier to build in a remote location rather than robustness being the primary reason for the style. I wonder if such false reasoning could cause pollution.

Are the (only) candidates of Canada/Alaska really “educated” proposals- with other cues being evaluated or more of a guess resulting from a heavy North American bias?

Also, it would be interesting to see what it does with conflicting signals-in this case, all the evidence aligns. Not a criticism, this is incredibly impressive, just my curiosity.

What if it were just abandoned and the satellite dish is simply misaligned?

I don’t know whether the picture actually merits a response that offers multiple options with assumptions and reasoning given for it’s proposals - have you seen anything like that in your ChatGPT interrogation travels?

Expand full comment

Excellent observations on bias - thank you.

Expand full comment