This market is fuzzy and ambiguous. On 2022-Dec-31, Greg Brockman (Co-Founder of OpenAI) tweets:
Prediction: 2023 will make 2022 look like a sleepy year for AI advancement & adoption.
Around 2024-Jan-01, I will reply to Greg Brockman's tweet thread (or DM them), and ask if they feel that 2023 "made 2022 look like a sleepy year for AI advancement & adoption". If we agree about whether this "happened", then I'll resolve the market that way.
But if we disagree about whether this "happened", then I'll convert this market into a Keynesian Beauty Contest. I would wait for the price to stabilize, either spending 5+ days at <=10% (resolves "No"), OR spending 5+ days at >=90% (resolves "Yes"). This is only meant to be used if me and Greg Brockman don't agree afterward.
If Greg Brockman doesn't respond by 2024-Jan-7th, or gives an unclear answer, then we'll use the KBC. The close date is deliberately set excessively far in the future, so replies and reflection have whatever time they need.
It's a statement designed to confuse rather than clarify - it's a statement that can't even be wrong because it's so metaphorical - but, it's also a statement that is more approachable / viral to the vast majority of people in the world than a lot of the more technical stuff that is out there trying to figure out more clearly where things will go. I'm trying to put together markets that are much more 3-rd party verifiable but likely a much smaller segment of folks are interested in these:
@PatrickDelaney crap, here is the other one:
So far it feels like it's happening. I would say the start of 2023 is comparable with the end of 2022, but the start of 2022 clearly feels sleepy in comparison.
@Gigacasting I think my previous comment still holds: "I hope we agree that so far nothing like that is happening, and you're expecting growth in adoption on average speeding up, and advancement at least not slowing down?" [by you I meant people holding YES] Making the previous year look "sleepy" is a big ask, changes would have to be very easily visible in many ways.
I hope we agree that so far nothing like that is happening, and you're expecting growth in adoption on average speeding up, and advancement at least not slowing down?
For the record, this "event" seems plausible to me. It would partly depend on how much better GPT-4 is. Other developments could add weight to the claim.
But that said, people get into exaggerated frenzies. All the time. So in this market, the wording says it shouldn't be based absolute excitement levels in 2023 -- we're making a comparison. Remember the crazy AI blast of 2022, such as Stable Diffusion and the ChatGPT demo. My Twitter feed has been frequently plastered with memes about how out-of-control people feel AI progress has been this year. I wouldn't want to unfairly discount a sufficiently-non-sleepy 2023. But I also expect many easily-excited people will have a kind of history amnesia.
Seems pretty hard to beat StableDiffusion in terms of adoption, but maybe something in that space will supplant SD itself. Audio generation could plausibly get to near-SD adoption levels this year.
Presumably Greg's insider knowledge means he has some good reason to believe this is true, wonder what in particular it is.