Developing policies around AI usage in general here at Leeds Uni is on-going challenge, especially given what AI can do changes on a weekly/monthly basis. And of course AI covers a host of different things - expect LLMs will seem antiquated this time next year and everything will be world models

I am seeing AI solve problems as part of my day job, but only when used with an appreciation of where it can go wrong. I saved weeks of effort using it to help write some PowerQuery code but then spent a week fixing things it got wrong.

Claiming the right to train AI on their users' content is bad, but that clause essentially claims free use of users' content - send a private video and they could slap their watermark on it and put it on Youtube. Send a manuscript, they could publish it on Amazon 😣

Must admit I didn't know it was a sub-genre to itself, which might be why I didn't kill the reindeer. Now I feel like I should and yet also would feel bad if I did...

I enjoyed it, but also felt there was a Santa as John McClane/Jason Voorhees concept in there that could have been great. I got halfway through trying to write that screenplay, but never figured out the ending.

If you haven't seen this on its second weekend drop: www.hollywoodreporter.com/movies/movie...

I doubt that'll happen, but eventually things will get so broken with some much wealth in so few hands that it'll have to change

The last few days of stock market panic are kind of all 'froth' compared to the impact of real data coming in over the next 6-12 months. Only glimmer of upside I can see if this goes wrong enough that consensus emerges the only real fix is to start taxing wealth properly

We had to go looking for hints. One of those days when you find the answer and don't go 'oh, of course', but 'oh, that's ridiculous'.

If there any no good comparisons for a submission or the only good comparisons are more obscure books/non-bestsellers, is it better to leave out comps altogether (even if it might suggest you haven't done your market research)?

Or even better, use the clickbait article to say 'famous author might not finish their book, so here are books by marginalised authors that are finished that you could try'. Maybe would avoid the problem of the non-clickbait articles not getting any clicks?

Surely even a master criminal would make sure to case a joint first? Not doing the research would be a sign of hubris, not confidence

Can't remember on which version I first used Final Draft (maybe 5?) but even then I was surprised how many bugs a simple/overpriced word processor could have. The fact it's now on version 13, basically still does the same thing, and still has bugs is ridiculous.

I touched on AI in my last series of books, but took the approach that the people who are trying to shape/control a future with AI make a better (and possibly scarier) story than the technology itself - 'Super-intelligent AI dooms humanity because company trying to improve quarterly forecast'

... which should be fine. That's a solid basis for casting a sitcom, but it feels a bit too trite. So, even though I keep telling myself the only objective with this project is to have fun, because I'm not going to try pitching it anywhere, apparently I can't let fun be easy.

... but it's also hard to avoid archetypes. For example, if multiple characters are sarcastic, oddball or naive/dumb, they won't play off each other as well as two characters with distinct styles. But then you risk ending up with a cast made up of the sarcastic one, the naive one, the oddball...