On 12/1/2023 I will use the best available paid model of midjourney (top paid or best plan available under $100/month) to run this prompt 5 times "a picture of a large overgrown concrete building with a large neon sign that says Pizza on top"
This will produce 20 images (5 mosaics of 4)
If the word Pizza is spelled correctly in at least 10/20 of the resulting images, the claim resolves YES. If not, it's NO. If the claim is tested successfully before this final test, it can immediately resolve YES, but I won't test it more than once per day.
Below is a sample of the images for that prompt today, 3/16/2023
Only the lower right image from the last of the 5 would count, so in total today's score would be 1/20.
@SneakySly Why's that? Other AIs can do this well, this market was not about some capabilities gain, it was about a particular product's version update
@SneakySly I chose the due date based on vague estimates of dev time after listening to midjourney office hours combined with experience waiting for deepfloyd, another text generating AI system.
So this is more of a "software is slow and hard to develop" thing, especially for a company which doesn't value it much and is growing hugely in other ways, so priorities change. David Holz likes text but doesn't view it as a killer app so isn't publicly rushing it in reports.
@Ernie Exactly. Image generation AI that does text well was around 6+ months ago, the most popular interfaces like midjourney's or DALLE's aren't even the best in the game. This is about a software release, not at all about some capabilities leap.
@firstuserhere agree mostly, but what's better than dalle3 rt now? It's amazing. Their rate limits hobble it to death. If it had MJ style availability it'd completely dominate
Interestingly, dalle-3 with this exact prompt is fine, but if you allow it to modify the prompt it sometimes gets it wrong:
picture of a large overgrown concrete building with a large neon sign that says Pizza on top:
A large, overgrown concrete building, depicting signs of age and abandonment. Nature has started to reclaim the structure, with vines and plants growing over its surface. On top of the building, there's a large, brightly lit neon sign that reads 'Pizza'. The sign should stand out, glowing vividly against the building's worn facade. The setting is somewhat eerie, highlighting the contrast between the decaying building and the vibrant neon sign.
@StrayClimb This is the market you're looking for, then: https://manifold.markets/firstuserhere/will-dalle3-create-correct-text-in?r=Tkxlc2V1bA
@StrayClimb seems more like one person bought the market down and the subsequent jump is others buying it back to the original %
Hmm, the tech seems to be coming along and this may be technically possible for at least some people right now. https://twitter.com/DrJimFan/status/1694358069638275463
The technology exists https://www.deepfloyd.ai/deepfloyd-if and could in theory be integrated into midjourney tomorrow, but I have no idea how high this is on the teams list of priorities