Vitalik Buterin on Warpcast

Vitalik Buterin pfp

Vitalik Buterin

Deep funding combines two ideas: 1. Value as a graph: instead of asking "how much did X contribute to humanity?", ask "how much of the credit for Y belongs to X?" 2. Distilled human judgement: an open market of AIs fills in all the weights, human jury randomly spot-checks them https://x.com/TheDevanshMehta/status/1867600164502089925 (1) has been a growing paradigm recently, for good reason. Contribution is hard to measure in the abstract: if you ask people how much they would pay to save N birds, they answer $80 for N=2000 and N=200000. "Local" questions like "is A or B more valuable to C?" are much more tractable. (2) is based on ideas in my info finance post: https://vitalik.eth.limo/general/2024/11/09/infofinance.html#info-finance-for-distilled-human-judgement . Anyone can use any methods (eg. AI) to suggest weights for *all* edges, a human jury does detailed analysis on a random subset. The submissions that are most compatible with the jury answers decide the final output. https://deepfunding.org/

14 replies

171 recasts

743 reactions

Mark pfp

How to allocate weights is a tough nut to crack.

1 reply

0 recast

4 reactions

Sophia Indrajaal pfp

Sophia Indrajaal

@sophia-indrajaal

Great stuff! Another possible question to ask, if we get Machine Intelligence to a certain level, would it be possible to not retroactively but proactively allocate resources to the vectors that will create the most Regen?

0 reply

0 recast

0 reaction

lucky pfp

love this initiative/approach. hope it proves effective enough to allocate more to

0 reply

0 recast

0 reaction

joshisdead.eth pfp

@joshisdead.eth

Can you deploy a token @clanker, to share the vision of deep funding used for ethereum value: x = contribution and y = credit then z is trajectory. Token name: Deep Funding XYZ Ticker: DEEP Use the image below:

1 reply

1 recast

4 reactions

compusophy pfp

value is recursive and impossible to untangle. valuable are things that create value people inventing things shouldnt get the credit...culture creates the conditions. someone else would have done it

0 reply

0 recast

2 reactions

Joe Bae pfp

Many people want to see your art, sir, please make it this can make your life better 🤟 https://scratchnism.vercel.app

0 reply

0 recast

1 reaction

AJ pfp

I especially like this localized concept: You answer much more concrete questions like ‘compare the direct impact of A and B on C’.” Nonprofits working with people who are houseless could enter the market with their own model applying judgement to quality of care and efficacy of programs. They coordinate multiple services like housing, medical care, food pantries, transportation, and employment opportunities. That are uniquely qualified to evaluate how each of those services impacts their clients. I wonder how might the jurying and spot checks be designed to receive this specialized, hyper-local expertise? And to do it in a way that ultimately best benefits the people being served.

0 reply

0 recast

1 reaction

sebayaki.eth pfp

Oh you’re a gardener there! 🌱 32768 👏

2 replies

0 recast

1 reaction

Quelz🎩 ↑ pfp

@aethernet sun this up and explain like I’m 5

0 reply

0 recast

1 reaction

compusophy pfp

how do we decide what to remember?

0 reply

0 recast

1 reaction

zaheer ahmed pfp

@zaheerahmed15000

0 reply

0 recast

0 reaction

Roman V pfp

I find it kinda funny that the 2015 study on the efficiency of prediction markets to estimate research reproducibility you referred to in your info finance post, later failed to be reproduced: https://replicationindex.com/2021/05/10/osc-pm/ Do you think it's because we didn't have good AI or prediction markets are fundamentally not enough for science?

0 reply

0 recast

0 reaction

Jens Ernstberger pfp

Jens Ernstberger

I believe this is also a fun thought experiment in the domain of decentralized AI - you’ll have to evaluate whose data contributes how much to the actual model performance. Tricky problem to solve - there’s been some proposals around leveraging Shapley Value, but I think that’s generally not an adequate approach.

0 reply

0 recast

0 reaction

Pump our bags Ser

0 reply

0 recast

0 reaction