Cory Doctorow makes some great points about how even the consent argument is a distraction and there are more useful ways to think about criticising AI.

pluralistic.net/2025/12/05/p...

Interesting. I’ve usually found new sessions helpful, apart from re-explaining context. I wonder if you might have more luck getting it to use something like pydot rather than modifying the raw .dot file directly, so that you can better seperate things you don’t want changed into eg a function?

When you feed the file back in are you starting a new conversation?

I don't know if it's a good model of the math, but I've found a helpful analogy for dealing with LLM behavior is to think of the whole thing like a map of language use, and it interprets the prompt as a map coordinate. That part of the map can have a lot of detail or not, or weird splotchy bits.

I mean, at this point I’d settle for just bringing back the swamp 🫠

Make America Swampy Again?

This article by @jrothman.com is the earliest example I could find:
www.jrothman.com/newsletter/2...

I also feel like there must be a hidden body of literature about this somewhere disguised under a different name.

As a Melburnian AFL is something you kind of have to deal with growing up but at some point it stops and the last time I spoke about it was years ago with a colleague from interstate who said that you had to have a team here and also what else would you talk about in a taxi?
1 - No.
2 - Who cares?

Oh, do you get people who kind of can’t grasp the concept and they are all “but you have to care because reasons” and you’re like.. “I hear what you’re saying, but I just don’t care about that” and their brain breaks a little?

Can you give an example?

true, I’m not suggesting that a viable 🅱️❎ always exists

and even if it does, the choice to pursue 🅰️❎ may still be valid and worthy

The line between optional and necessary can become blurry. If there are two plans 🅰️ and 🅱️ that lead to ❎ the devout yak-shaver will tend to define their goal as 🅰️❎ rather than ❎ and will do anything in their power to solve for 🅰️❎ no matter how much simpler or easier it would be to pursue 🅱️❎.

I’m sure there would be folks in govt who share your concerns and are evaluating risks and considering contingencies in case US doesn’t deliver. But if this is happening they wouldn’t (and shouldn’t) make any deliberations public unless it’s time to pull the trigger on an alternative.

The question you just asked is an example of the former and not the latter. If I were to ask you a question in response, would that be an example of the latter?

They certainly did an awesome job freeing Assange!

www.theguardian.com/australia-ne...