I hope TiddlyWiki goes back to the hand-crafted, human banners of yesteryear?
I confess that the wind has gone out of my sails for banner design, after many contests in which I was eager to submit a design or two or five.
Worse, I did not vote this time, and I even dropped the ball on keeping my contest website updated. I feel I owe the community, and @jeremyruston, an apology for that lack of engagement.
To be fair, the past three weeks have also been a very busy time in my worklife, and that’s part of the reason for my silence in this thread (after the initial announcement).
Still, I suspect my shifting relation to the banner-contest may be worth sharing.
My own designs were always entirely worked up through a modest graphic-design toolkit pretty analogous to what I could do with pen and paper and scissors — nothing more computationally sophisticated than bezier curves, gradients, and layering. (One of my designs was chosen once, and I happily cheered for others’ creative graphic work every other time.)
These days, I’m completely uninterested in enlisting a “free” (never really free) big-data-genie, just in order to delegate graphic-design details to its power-hungry circuits. (Some tasks may be high-stakes enough. Let people turn to giant-data-digestion-engines to solve heinous crimes, sift through bio-medical-diagnostic-epidemiological patterns, etc. And there’s plenty of grey area… But submitting a “winning” design — by “me”? — to a beloved small community of TiddlyWiki users? Such a task is not in the grey area for me; it’s not worth setting up a mild-meld with any of the beckoning behemoths.)
But I’ve discovered I’m simply not at all motivated to put hours into graphic tinkering (through trackpad clicks to build up each line and shape and fill gradient), if my work is “competing” with corporate-Windigo eye-candy, and thus would probably call for including a sales-pitch in the “Hey-folks-this-is-old-fashioned-hand-shaped-with-only-one-organic-brain-plus-local-software!” vein.
Yours, deflatedly,
-Springer
5 Likes
Thanks for raising this @Springer. Likewise, I really do not like AI generated imagery, and am not happy with using it to represent TiddlyWiki. I think it conveys to visitors that we don’t care about creative work, which is particularly unfortunate for a product intended to support creative people.
So, I regret not stipulating a “no AI” rule for the banner image competition, or at least raising the question of doing so.
The winning entry was by @Peter, who said in his submission that he would understand if AI entries were not allowed. It’s embarrassing to do it retrospectively, but I don’t think we should use any of the AI generated entries.
That gives us a problem because there were no entries that were not generated by AI.
I don’t think there’s time to rerun the banner competition. It would be fantastic if there was anyone out there who was able to make a manual interpretation of @Peter’s idea. I have some thoughts about alternative approaches if that doesn’t work out.
I’d welcome the thoughts of others.
3 Likes
I’m very glad to see you taking this stance!
We did have some discussion about the suitability of AI-generated banners last year, in conjunction with the 5.3.7 banner contest. At the time, the majority view seemed to be that it harmed nothing to allow AI entries, and that people who disliked them could simply vote against them. And since the winning banner used for 5.3.7 and 5.3.8 was indeed AI-generated, I suspect that the community may have gotten the message that AI imagery is welcome… and, like @Springer, that there’s no reason to do one’s best with Photoshop or Illustrator when a flashier AI image is likely to win anyway.
If this doesn’t align with your current views, I think you may need to formalize (and publicize) your position on the use of AI in TW’s branding. Otherwise, I imagine people will continue to assume that it has at least your tacit endorsement.
I personally did like the submissions. I could identify with all the ideas behind the images. For me that’s probably more important than the image quality itself. But I am fine if they are pleasant to the eye too.
IMO it looked like, we would have gotten no submissions at all. May be this time there was too much time upfront. It seems the community gets more creative or more motivated, it there is time pressure … I don’t know.
I did like the “idea” the “story” behind the initial submission. Sadly, I did mess up the title. Reading the initial description again, I just saw that it said: “New Frontier”
I personally can do some basic stuff with Gimp, which I did for my own submission for v5.3.6. But in no way, would I have been able to make something even similar to the “New Frontier” image.
I could / would have come up with a similar idea as for v5.3.6, but the sheer amount of changes to v5.4.0, did overwhelm me.
That’s why I also liked the submissions from shenzhy16. He did focus on one “key feature”, which I think was smarter than me 
As I wrote, if one user submits several images, they do “canibalise” each other. If there would have been only 2 images, the “race” would have been much closer. …
Just some thoughts
Mario
2 Likes
Yes, I think in the past when I submitted multiple designs, I was encouraged (and/or simply individually motivated) to settle on just one before the final candidate pool was closed. Especially when they’re variations on a theme.
It’s not fair to stipulate a rule after a contest has completed. Especially since the issue of AI has come up before, and there was plenty of time to add the requirement.
Logically speaking, the rule should be that everything has to be drawn with the little editor that comes with TiddlyWiki.
The time to stop AI was 10 years ago when it was still producing 6 fingered cats. I don’t think there’s any putting the genie back in the lamp now.
4 Likes
Perhaps the community can adopt the Australian way and allow preferential voting, ask people to order all entries from favorite to least favourite. so simple maths will resolve the deeper all round preferences.
Wether we allow LLM entries in future, which in many ways is more democratic, or stick with skilled submissions keep in mind we will look back at such submissions as there at the beginning of an era, there is no harm having the history recorded.
Indeed, because today it is already producing a generation of script kiddies proudly calling themselves vibe coders. At least an image is just an image and can’t cause much harm. But code hallucinations making their way into peoples’ computers scare me. I prefer to be the stupid guy who is clueless about a topic and may casually copypaste an oneliner found on SO (old one, from pre- so called AI era). But when it comes to industrial levels of generated code, I feel helpless to fight against.
Hi folks,
Thanks for considering my submission!
I tried making another one, matching the spirit of adventure from the first. It was not generated with any AI tools, and is a collage of images that are in the public domain.
Citations:
Excited about the release of 5.4.0!
5 Likes
Peter,
That’s awesome. I like it!
Just saw this - was AWOL - for a long time - (still will be 
).
Mine - simple - and yes - “some” AI.
I haven’t been following along to the banner vote or the AI discussion - but just comparing this to the original AI version - I much prefer this one.