HN Reader

the author mentions they’re using haiku for the model, but I wonder if the travel query -> place transform could be done with a tiny local model.

But that’s the bitter lesson I guess, unless there’s a reason to go fully local (extreme privacy concerns, offline use, etc) training a custom model just isn’t worth it over using some cloud api offering, even if it is orders of magnitude more compute that _somebody_ is paying for in the end

Photo is made, not generated!

Couldn't you grab the image from the Wikipedia page of that location? For example searching for "Deadvlei" from the blog post gives this which has a photo:

https://en.wikipedia.org/wiki/Deadvlei

The annoying thing with Unsplash is that it sounds like a really permissive license, but the moment you try and do it at scale you're no longer in compliance with their terms. Also their definition of not building a competing service is really broad. Maybe i'm being overly cautious but I get eeked out by all that.

This sort of thing will kill the graphic design industry.

API limitations? Why not have the script run for a few days collecting images and then crank through all of them in an hour?

[flagged]

[stub for offtopicness]

[sorry I messed up with that title! perils of not reading the articles closely]