Which Institution Will Have Shown To Have Been the Most Prescient In Terms of the Direction of A.I. by the end of 2023?
resolved Jan 12
Center for AI Safety https://www.safe.ai/
Eval.AI / CloudCV / Georgia Tech https://eval.ai/web/challenges/list

Preface / Inspiration:

There are lots of interesting A.I. leaderboards out there and I think we could all benefit from knowing what's out there and what they are working on, sort of a meta-leaderboard overview.

Market Description

Several Universities and Organizations lay claim to being highly authoritative with regard to A.I.

Of course hindsight is 2023, so while all sorts of holy prognostications and WAG's (wild ass guesses) may presently be touted as dogma, we can all come back together and review how things went at the end of the year.

Market Resolution Threshold

Which organization will have proven to be the most prescient?

  • This market will be resolved solely by the market author or his desginee via PREPONDERANCE OF EVIDENCE, meaning that the burden of proof is on the betters to support their case with convincing arguments, and that it's more about weighted liklihood that something happened, rather than more of a, "reasonable doubt," thing, which would have a much higher standard. This means that likely percentages will be dolled out among institutions.

  • I really don't like people just stating things and not backing up their arguments. Therefore, for the purposes of this market, comments with links and evidence to back up arguments will be read and reviewed. This includes minutia and lawyering about the definition of the word, "prescient."

  • Comments with no citations will be disregarded.

  • This doesn't mean don't attempt to make convincing argument, it just means something like, "C'mon, it's organization X, we all know they are the best," means nothing. A-priori social beliefs do not count, there must be evidence given.

Get Ṁ600 play money
Sort by:

At the risk of pissing off @NikolasKuhn I think from what I have seen HuggingFace ended up being the most prescient based upon based upon the direction of AI in 2023 simply based upon the fact that OpenLLM's advanced in benchmarks, while showing the OpenLLM Leaderboard within Spaces, while nothing else on here showed any evidence of AI moving in any particular direction, but rather were speculating.

This doesn't say anything about the organizations capability to show what will occur over another decade or more time period, but rather just within 2023.

Please let me know if there is any objection to this.

@PatrickDelaney I think I wanted to mostly boost this market for some interesting discussion. As it is, I think the resolution criteria are vastly under-specified and resolution will be a matter of personal taste.
As public scrutiny of AI developments has increased and several national AI safety initiatives have been launched this year, I would personally indeed say that CAIS has been the most prescient - but it really comes down to personal taste.

Considering the engagement of the market and the lack of clear way to resolve, the most reasonable thing to me seems to resolve to N/A. But I'll leave it up to the author.

Possibly relevant market... ?

Hmm I haven't even heard about one of them, I am a bit surprised there is no Hutter Prize and nothing by Hendrycks etc al.

Still, interesting question!

@ValeryCherepanov one -> some*. In fact, I haven't heard about 3 of them, possibly even 4.

@ValeryCherepanov Added https://www.safe.ai/ and http://prize.hutter1.net/ ... had never heard of Hendrycks but he appears to be the director of Safe.AI

Other relevant markets:

More related questions