Based on this question:
However, Scott's question description says: "The AI doesn't have to be available to the public, as long as it's confirmed to exist."
This market is about whether the AI will be publicly available. I intend to interpret "publicly available" quite strictly. It has to be possible for a normal person to obtain. This is codified as follows:
Either it's some kind of remote service that:
Costs less than $200 for a 90 minute (or greater) film.
Or, it's something you can get set up on your own computer where:
Your computer is a "normal" computer that can be bought for less than $1,500,
The software itself costs less than $1,000.
Either way, the whole process must:
Be legal (in at least one OECD country),
Take less than 24 hours per film (90 minutes or longer),
Not require excessive technical expertise. (This is probably more relevant for the "run it on your own machine" scenario. I don't have a good idea for how to quantify "excessive technical expertise" so I will use my judgement.)
Importantly, this market will resolve based on my opinion of "high-quality", not Scott Alexander's. I expect his opinion will be reasonable, and I think I'll probably agree with him, but I reserve the right to disagree if I think he's too strict or lenient.
Since this market has several subjective elements, I will not bet. However, you may consider it relevant that I hold a very large (by my standards) NO position in Scott's market. I don't think this is a direct problem because I can't influence Scott's resolution with this market, but maybe it makes me biased somehow? Use your judgement.
Edited to remove section about suggesting modifications to the resolution criteria. The criteria as they are now will stand (unless some extremely surprising thing comes to light that makes the letter and the spirit of the market diverge catastrophically).
1,000
3.00@ErickBall I'm coming at "high quality" from the point of view of "does this count as a proper film" and not "does this hold up as a film for the ages". So most existing films would qualify. Examples of existing films that wouldn't count as high quality:
The Room (plot and dialogue doesn't make much sense, one character changes actor halfway through)
Epic Movie (plot doesn't make much sense, jokes don't land)
Catwoman (can't really remember it; I think it was more coherent than the above two but it was just so awful)
I'm expecting that if this happens I'll probably get a chance to see several AI films, so if there are a few I think are bad, that won't be enough for a NO resolution, as long as there are some that are good (or if the ones I don't like are highly regarded by people I respect). (Question is about whether AI is able to generate high quality films after all.)