Eliezer Yudkowsky Tweets (https://twitter.com/ESYudkowsky/status/1715753440024854983): Predicting trajectories is harder than predicting endpoints, but, possible scenario #3: we may be in the last year or two of the era where it's easy for a human to get another human's attention - where you are not a biological legacy media company competing with 1000 AI tweeters.
Michael Vassar: I'd take that bet.
This wasn't easy to operationalize exactly, so comments suggesting a better target are welcome, but I figured I will start out on this. And take suggestions on better options if there is interest.
The question's intention/spirit is: Will 75%+ of views on social media be of posts written primarily by AI?
If the answer is clearly YES or NO that will be good enough - e.g. if it's trading >90% or <10% near the end and I have no reason to think that's obviously wrong, I will simply resolve that way. Or if I see the answer as overwhelmingly clear for whatever reason. Ideally if necessary I will take a random sample and evaluate. If I can't find a good way to gather the data at reasonable time cost, or otherwise there is real uncertainty, I will resolve to the non-manipulated market price.
This will resolve early if it is >95% or <5% and the answer seems very obvious.
1,000
3.00Artificial intelligence is continuously creating possibilities. IBM’s statistics reveal that 35% of companies have already implemented AI in their business while an additional 42% are actively exploring AI in 2022 (BMI, 2022). Furthermore, the survey conducted by Marketers Worldwide highlights that 76% of marketers employ generative AI for basic content creation and copywriting, ranking these as the top two applications for generative AI (Hootsuite, 2023). An interesting study also shows that people tend to trust more on AI-generated tweets (The Verge, 2023) which could potentially result in a surge in AI-generated views by 2026. However, I believe it will not grow quickly enough to reach a 75%+ share of AI-generated social media views even though this is surfacing the web at an alarming rate.
I think you want to say that “posts” only count towards your metric once they’re seasoned for some de minimis period of time. If AI posts are being made cheaply but Twitter is automatically squashing most of them within a minute or two such that almost no users see them, you don’t want to count those, I think.
Separately, I think probably useful to clarify whether or not a LinkedIn-style “you can use AI as a native feature to help write your own posts” would count against this. (I think yes, but YMMV)
@DaveK Depends on the relative contribution of the human vs. the AI and whether or not it is centrally still talking as and to a human? My prediction is I'd say YES to most such features.