EG "make me a 120 minute Star Trek / Star Wars crossover". It should be more or less comparable to a big-budget studio film, although it doesn't have to pass a full Turing Test as long as it's pretty good. The AI doesn't have to be available to the public, as long as it's confirmed to exist.
1,000
3.00You know it is over, right? Or do we have to wait until Jan 1st for people to realize that another year has passed without much progress? Or was there enough progress since March 2023 to justify the same valuation now? Resolution was 5 years away then and it is 3 years away now.
@skibidist I keep pointing that we are obviously not getting anything close to proportional progress towards that goal and people just keep denying reality.
I will mention it again at the end of the year, and at the beginning of 2026, and 2027. And the year after that, I will collect my winnings.
@skibidist Sora didn't exist in March 2023. We had imagen video. My gut is there isn't enough progress in 2 out of 5 years or so.. but we'll see.
@Usaar33 An individual person can make a short clip using a few dollars worth of time.
A full high quality movie takes millions of dollars. But by the standards of the people in this market, you should be able to do it with a few hundred dollars at most, because it is "just" the short clip multiplied by a few tens or by a hundred at most.
That of course explains why the standards are wrong. Making a short clip is not even 1% of the way to making a high quality movie.
A bit rambly, but Ray Kurzweil predicts AI movies within a year or two (minute 50):
@Joshua having both play money with loans and having sweepstakes makes manifold the life of the prediction markets party
@jim to be fair, I’d argue that AGI is a necessary but insufficient condition to generate full-length Hollywood films from a single prompt.
@benshindel The opposite could be argued as well: Generating full-length Hollywood movies from a single prompt is a necessary but insufficient condition for AGI.
I tend more towards your argument, but I'd guess if this market would resolve yes, there'd still be no consensus (outside tech bubble) for "AGI exists".
@jim We won't have AGI in 2028.
But even if we did, it is true that we would not have full length AI generated movies.
@benshindel Correct. That is what I said. We will not have AGI in 2028, but even if we did, we would still not have the movies.
Same with the Turing Test market; even if we have AGI by 2029, it still won't pass the test.
"pretty good" in the resolution criteria is extremely vague and really should be improved given the popularity of this market.
The criteria I bet on in June is "whether I'd give the movie at least a 6/10 on IMDb (which would put it in the top 85% of the 1377 movies I've seen/rated, i.e. not horrible) if I wasn't rating it especially favorably due to being impressed by the fact that it was created by AI (a novelty factor that would wear off quickly after a few AI-generated movies)."