Zarf Updates

IFComp 2025 has wrapped, and the results are posted! Congratulations to Detritus (Ben Jackson) and all the other entrants that did well.

Now let's talk survey.

The post-Comp feedback survey is live. (As linked from this official post.) Jacqueline, our organizer, writes:

This year’s survey is a little more important than usual, as we’re seeking thoughts on the future of how we address the UK Online Safety Act, how much generative artificial intelligence should be allowed in the competition, whether or not you’d like us to continue the awards livestream, and highlighting some important new volunteer roles for those who are interested. Please check it out! We need all of your thoughts in one place, so even if you’ve expressed opinions [on the forum], or in Discord, or emailed us, we will look for your thoughts in that survey.

"A little more important" is a bit of an understatement. Discussion about generative AI (LLMs and art generators) got contentious this year. See this forum thread, which hit 283 posts in seven days flat.

Let me back up. For the past couple of years, IFComp has permitted entries that use AI tools, as long as that usage is declared up front. Looks like eleven of this year's entries (out of 85) made the declaration -- some for game text, some for game art or cover art, some for both.

(Footnote: the AI usage labels have disappeared from the site now that voting is over. I think this is an oversight; I'm checking with the web team now. Refer to the wayback snapshot for how this looked during voting.)

There was some discussion about AI usage last year, but that didn't translate to a lot of feedback in the 2024 post-Comp survey. See this note from September:

It would be good if this year's contentious discussion got converted to a lot of survey feedback. Otherwise, it is likely that the same policy will be once again kept for 2026.

Why does it have to be survey feedback? See the note above: "We need all of your thoughts in one place." It's just not possible to go through a 283-post thread and figure out what the community thinks. Some people post twenty times; some people post once and then duck out; some people see the thread counter smoking its bearings and mutter "I ain't going near that thing." Some people don't read the forum regularly at all. (Note that the thread was locked after seven days, so it was perfectly possible to miss the whole thing.)

And, come on, you've met the Internet. It's not hard for a few people to push a point of view by being persistent and noisy about it. I'm not saying that's happened here -- but the way you tell the difference is by doing a comprehensive survey. Trying to parse quantitative results out of a forum discussion is a fool's errand.

In case you're wondering: I am listed on the IFComp committee. My role is entirely advisory at this point. I'll be looped in on the discussion on "what we gonna do next year", but the final decision is the chair's.

I have thoughts about AI use in IFComp, but I'm not including them in this post.

There's also the whole issue of the UK Online Safety Act, which sucks but we have to follow it. Yes, we asked our lawyer.

TLDR: Here's the survey. Go for it.