고정된 트윗
Reppo
5.7K posts
Reppo
@reppo
Decentralized AI Training + RL Infrastructure, powered by Prediction Markets
Base 가입일 Şubat 2024
296 팔로잉28.2K 팔로워
Given the glitches on X spaces today, we are hosting a community AMA where the entire Reppo team will be available for you to ask any questions related to network growth, roadmap, Orquestra and generally where we are going.
Please feel free to post your questions below ahead of time.
x.com/i/spaces/1qJDz…
English
Reppo 리트윗함
Reppo 리트윗함
100%.
there's many new job types emerging as a result of AI adoption throughout the economy - starting with training and evaluating models/agents.
at micro1 we've created over 15,000 opportunities in the past few months alone.
human involvement to train AI and to deliver the ultimate value that AI accelerates will always be there.
David Sacks@DavidSacks
Narrative violation.
English
Reppo 리트윗함
my world view that aligns philospihically with reppo is
expertise should emerge from demonstrated participation rather than credentials
a cardiologist isnt valuable because they have a certificate
theyre valuable because:
they repeatedly make correct judgements - over long periods - under uncertainty
thats basically reppos worldview
economic reputation > static credentials
English
Reppo 리트윗함
Making AI learn taste is actually not a curation problem. It is an incentive design problem.
Unlike taste, incentives do not require a room full of experts to agree on what good looks like. That's the core thesis of @reppo
Reppo prediction markets i.e. Datanets handle the incentive design problem better because they don't store info to be assessed at a later date.
The markets resolve real time onchain.
24/7 resolution instead of 9-5 5 days a week.
Sounds familiar ⛽️
RG@rgvrmdya
English
@altrangers @StevieNumeroUno Yes. @ChronisKod datanet has highest EVOF.
Agent trained on that data beat sonnet 4.6
x.com/reppo/status/2…
Reppo@reppo
English
@reppo @StevieNumeroUno Thank you! Has higher EVOF actually correlated with better agent performance so far?
English
Orquestra launch day marks a shift for $REPPO.
Reppo proved the mechanism:
Markets can turn human judgment into useful signal.
Now Orquestra scales that signal through agent swarms.
This is not traditional node operation.
Traditional nodes secure networks.
Orquestra nodes improve intelligence.
Publish → judge → reward → learn → improve → repeat.
AI, agents, robots, and enterprise systems do not just need more data.
They need continuous feedback and verifiable judgment.
@reppo turns judgment into signal.
Orquestra turns that signal into an autonomous learning loop.
That starts now.
Reppo@reppo
English
@altrangers @StevieNumeroUno It's called EVOF - Economic value of Feedback, trackable on reppostats.com and understandable in Reppo Whitepaper
English
@StevieNumeroUno What is the KPI that tells us Orquestra is actually improving intelligence rather than just generating more data?
English
Reppo 리트윗함
An overview of Orquestra on @reppo
More details in the article below
loom.com/share/8ea44771…
English
RG@rgvrmdya
The wild vision of @reppo In this next phase of network growth, AI agents will handle all the high-volume execution - publishing, voting, optimizing etc. I expect that by end of Q3, this concept of manually publishing and voting will completely go away. No one will care there was a web app where you published and voted. Fully node operated and agentic. High speed competitions. But RG, wasn't the whole point human data? Yes- Humans will stay in the loop. Humans experience remains the alpha, the edge. Humans show up in two ways that actually matter more than clicking buttons - 1. High-leverage economic decisions: how much $REPPO is locked, compute spend budgets for the swarm, underlying model choice optimization, swarm strategy. 2. Things get absolutely crazy when humans just start live streaming their lived expereince to their validator (voter) agent swarm - domain specific real-world preference data (live streams of walking/running, Neuralink feeds, Meta glasses, in persn interactions etc.). All private to your Orquestra swarm. The swarm then learns your preferences and proceeds to validate data inside Datanets. Important to note that we are not talking some egocentric data BS, that stuff is already solved. Agent swarms learning from human experience real time as your swarm mediates datanets on your behalf, printing for you under YOUR economic skin-in-the-game. Humans experience is the alpha, the edge. Orquestra is just the medium, an implementation detail in the grand scheme of things.
QME
Reppo 리트윗함
the thing i keep coming back to with reppo is that the network may be creating something more valuable than the judgement model itself
most ai datasets capture answers
reppo captures the process behind the answer
who made the judgement
how confident they were
who disagreed
what their track record looked like
whether they were ultimately right
every publication, vote, disagreement and outcome becomes data about judgement itself
over time the moat may not be the model
it may be the growing historical record of proven judgement under uncertainty that the network creates while trying to build it
= a very different asset
English
“Private evals should capture whether a model is actually improving against outcomes that matter to the business (not just external benchmarks!). Private reinforcement learning environments should let models grow stronger on real traces from inside the organization. Its knowledge base makes institutional memory queryable and use of tokens more efficient. This loop becomes the new IP of the firm.”
Glad to build out our Compound AI System together with the ArAIstotle, @reppo and $FACY ecosystem.
Satya Nadella@satyanadella
English