Who are the top 10 AI researchers?
➕
Plus
102
Ṁ51k
resolved Feb 21
Resolved
YES
Ilya Sutskever
Resolved
YES
Geoffrey Hinton
Resolved
YES
Yann LeCun
Resolved
YES
Andrej Karpathy
Resolved
YES
Yoshua Bengio
Resolved
YES
Demis Hassabis
Resolved
YES
Jürgen Schmidhuber
Resolved
YES
Eliezer Yudkowsky
Resolved
YES
Paul Christiano
Resolved
YES
Alec Radford
Resolved
NO
Ian Goodfellow
Resolved
NO
Kaiming He
Resolved
NO
Alex Krizhevsky
Resolved
NO
Richard Sutton
Resolved
NO
Andrew Brown
Resolved
NO
Diederik Kingma
Resolved
NO
Alex Graves
Resolved
NO
Shane Legg
Resolved
NO
Stuart Russell
Resolved
NO
Christopher Manning

Follow-up: /singer/will-one-of-these-ai-researchers-cl

I want to establish who Manifold thinks are the top 10 most important living and active AI/ML researchers. This poll market closes in just a week.

UPDATE:

The 10 winners will be selected from this poll:

/singer/who-of-these-20-is-the-most-importa

Get
Ṁ1,000
and
S3.00
Sort by:

I'm flattered that I even made it on this poll - voting me off is definitely correct!

bought Ṁ20 Alex Graves NO

Can we resolve away all the names that didn’t make it into the poll?

bought Ṁ100 Christopher Manning NO

This market will not answer the question "who Manifold thinks are the top 10 most important living and active AI/ML researchers", given that the poll options are very limited

@JonasVollmer are you suggesting that not enough answers were on the poll, or that the poll format is overly restrictive?

@singer The former! I think many top AI researchers are missing from the poll

@JonasVollmer I don’t think that was the problem with the poll, the problem was that each person could only select one option, so rather than “who are the top 10” it was selecting for “who are the 10 most likely to be ranked #1” which is a very different question

@benshindel

so rather than “who are the top 10” it was selecting for “who are the 10 most likely to be ranked #1” which is a very different question

It's not immediately clear to me why these are substantially different in principle. Are you thinking of a problem where not enough people vote, and so the results are determined by chance?

@singer Imagine a room full of 20 democrats and 5 republicans. You give them a list of 100 prominent politicians and ask them who their favorites are. If everyone gets ten votes, the top 10 politicians will probably be democrats. If everyone just gets one vote, you'll probably see a republican politician in second or third place.

No Noam Shazeer?

I'm a little offended at all the NOs on Andrew Brown, but I guess I'll take the free mana. 🙏

@traders If you believe that the resolution criteria change has unfairly hurt you, please DM me with the details and I'll see about reimbursing you.

@singer wouldn't it make sense to close the market before starting the poll?

@Thomas42 Personally that feels like it would defeat the purpose of having the market, especially since all the bets in the market were made under different resolution criteria. It would make more sense to just NA it, which I considered doing, but I expected fewer complaints this way.

In any case, I've changed things up too much already. This is how it will be until the market is resolved.

@singer There won't be enough responses in the poll to meaningfully reflect people's opinion. If we could each choose 10 options in the poll, it would be more meaningful. Right now I'm simply voting for the one researcher I have shares for.

@Primer Percentages will still break ties, so there's no concern over ambiguity. I'll do my best to boost the poll. If you would like a reimbursement, please DM me. Thank you for being patient.

@traders Your vote on this poll would be appreciated: /singer/which-resolution-criteria-is-best-f

UPDATE:

See the new resolution poll here /singer/who-of-these-20-is-the-most-importa

@singer lol, I voted for the poll, but I didn't realize it would be 'vote for 1'. I absolutely would have voted differently 😂

@DavidFWatson Yeah, the poll format not giving an option for N-votes-per-voter is unfortunate. I wish that I knew a way around it.

@singer You could have made say 3 polls and asked ppl to pick their 2nd and 3rd, and then summed them

Yann LeCun

They projected his face onto the Burj Khalifa, so it must be true.

Comment hidden
© Manifold Markets, Inc.Terms + Mana-only TermsPrivacyRules