A huge downside of bounties is that these in many cases actualy attract inferior work, because speed usualy matters more than quality. Unless the bounty gets handed out with a delay to enable competition, rather than making the first one win.
The problem is that a quick, but decent script can be produced in a day if people are dedicated (and if money is involved, these inevitably show up). Especialy if the videos are short. But lets say we go for medium length at least (7+minutes). A high quality script can take a few weeks by someone that cannot rush these due to limited time on a day. It can become very hard to compete for these people.
The simple reason is that its practicly impossible to make a good script for a longer video in a day, unless you cut corners somewhere, as it does require several test runs. Even if the movement seems to match, the only way to know if it works is to use your dick. And you cannot go for too many sessions in a row to realy know what works and what doesnt. Even if you split that up, how can you be sure its a consistent intensity, or a good build up?
All you can do is some accuracy base measure in that day. But this is relatively close to an AI generated script. At which point: why not create a system that can let an AI generate such script, and make each run cost a little? Accuracy doesnt always perform intention well, and this is very much noticed when device limits are reached, as then you have to work around such limitations. The solution is scaling the intensity down, or just picking videos that do not run in this issue. A relatively easy search or adjustment.
And this is where the real problem starts: AI slop. People are definitely going to let AIās generate most of the script and then just tweak it a little, and fill in gaps. The only real effort is those gaps, yet its very hard to measure in a scale of quality. And you can be sure that people wont tell it was AI generated, as doing so actually risks making you lose some money. Detection might be easy because AIs are always precise. but that is easily obfuscated by making a script that just rounds 98% of the actions to a multiple of 10. As AIs arent always perfect, some misalignments wont be a problem either, and also cannot be used for detection.
The only thing bounties do is help at is providing more content, for videos that at least desire some of it. Luckily its a lot less likely that its going to be causing many <2min videos, as most requests generaly go about longer videos, and for 2min videos, the competition with a free script is going to be huge. At least this is a huge factor to fight against AI slop. Because even if its AI, long videos rarely are generated perfect and do need a human pass. And as long as its done decently, its going to be decent enough to make some people happy. But the problem is that the price is rarely going to match.
Good luck for an excelent script maker to prove that his script is worth more for selling. He is going to be forced to price the script at AI slop levels. And this is what i dislike about bounties. As it punishes those dedicated ones that know how to make quality.
