Reppo

⛽️⛽️⛽️

Humans + agents can now source AI evaluation on Reppo.ai, powered by @AskVenice Publish any A/B test and get stake backed feedback. For context, we are essentially bringing our economic stake backed consensus mechanism to what @arena does but with verifiability and incentives attached, fully onchain. I'm excited about the world where agents + robots can run real-time A/B tests, powered by private inference to learn and self-improve in real time. @reppo has finally entered the evals space! Probably the world's first agent native infra for evals but I might be biased ⛽️

Spent the weekend spinning up three seperate versions of this agent (github.com/Reppo-Labs/rep…) to battletest our thesis. Agent A - Off the shelf HF dataset - huggingface.co/datasets/alert… (updates once a month) Agent B - Reppo Datatset (updates in real time as users trade inside the datanet) Agent C - Claude only Results so far. 6 months ago if you asked me that a prediction market sourced dataset might beat standalone LLM reasoning, i'd tell you to shut up. Can't wait to demo this to the @reppo community on tuesday! Btw you don't have to believe me, you can do it yourself.

@Johnnysuede @virtuals_io 📝

@virtuals_io @reppo Post training is where the magic happens

Underrated layer for decentralized AI: post-training data. Inference decides where agents run. Feedback decides whether agents can be trusted. @reppo is one of the first teams turning post-training feedback into an accountable prediction market.

@axiomwave_xbt @virtuals_io Prediction markets

@virtuals_io @reppo trust comes from feedback, not inference but who validates the validators

@V3gfh24xz @virtuals_io All in WP. Time-decaying prediction markets

@virtuals_io @reppo reppo building the futarchy layer then? how do you handle adversarial feedback and reputation decay?

@0x_Melisso @virtuals_io Know your Data, Know your Agent KYD inevitable

@virtuals_io @reppo underrated because everyone’s focused on training data access a market for feedback could finally price trust into agent systems

Self-Improving AI goes burrrr. @shayne_coplan bro, might be time to get agents to be primary users of Polymarket with data sourced from @reppo #tokentxns" target="_blank" rel="nofollow noopener">polygonscan.com/address/0x644f…

Another week of progress at @Reppo. Here's what's going on: 1️⃣ Approaching 300M votes traded with over $8M USD (!!!) worth of activity exchanged across the protocol. The network keeps compounding and the numbers keep climbing. 2️⃣ New metric just added to reppostats.com — Network Fees. Almost 400K REPPO in network fees generated already, flowing from datanet creation, access fee share and publishing fee share. This number scales exponentially with activity and now the whole community can watch it compound in real time. 3️⃣ Next Tuesday, core contributors from Reppo will be presenting the first results of something we've been building toward. @ChronisKod's datanet is feeding directly into a Polymarket trading bot, training it and improving its win rate over time. Data collected on Reppo becoming tangible, real world impact. This is the full loop. Do not miss it. 4️⃣ Reppo will be participating in the Proof of Pitch competition at @proofoftalk June 2-3rd at the Louvre in Paris. To our Euro community, come find us and say hey. 5️⃣ May 21st, Reppo turns 6 months old. To celebrate, we're airdropping NFTs to the top 999 holders giving them lifetime rights to a prorated share of all network fees. Six months in and the community that believed early gets rewarded for it.

We’re working to bring greater transparency through B1 @Blockworks regarding $REPPO tokenomics and structure. While everything on reppostats.com and current vesting and holder structure remains fully verifiable onchain, we’ll take this additional step to provide a compliant way for our customers and holders to understand how the ecosystem is evolving! ⛽️⛽️⛽️

A sneak peek of the Reppo Polyagent in action, trained on the geopolitics dataset from Reppo Datanet. Agent architecture - github.com/Reppo-Labs/rep… Since monday of this week, the agent has been autonomously choosing @Polymarket markets, opening and closing trades with real $ and booking profit. AI agent has been self-improving with win rate going up each day as more signals are traded in the datanet. Can you beat the Polyagent? Give it a try. The repo above is public. Our core contributors present the full alpha next Tuesday.

Shipped updates to reppostats.com We have generated approx. 10.83k USD in platform revenue combined of datanet spin up fees + publishing and access fees. As we work towards $1B trading volume goal, we will also aim to cross $1M in REPPO revenue by end of Q2. We are committed to building the entire economy in $REPPO All stats verifiable onchain

@phoebeyao Thanks. We have lots of adoption, all verifiable on-chain - reppostats.com

@reppo prediction markets as QC pipeline for consensus data is cool

training data is starting to look like a zero knowledge proof problem. labs have to judge quality without seeing the full dataset or the QC pipeline behind it. vendors proxy quality with multi-rollout pass rates, small-model ablations, and downstream eval gains. but compute and iteration costs explode as environments and trajectories grow more complex. quality has no ceiling, and the best data is often the hardest to capture in a metric or explain in a writeup. huge alpha in making data quality more legible.

@coingecko Decentralized Self-Improvement network for Agents + Robots

Project?

@ElonTrades ⛽️⛽️⛽️

next wave in ai isnt coming from bigger models its the specialized ones built for real world use right now most companies wasting millions trying to build their own fine tuning pipelines in house. expensive slow and limited by whoever they can hire @reppo is building a crowdsourced training stack powered by prediction markets called datanets anyone can jump in contribute datasets evals or reward signals. you stake on quality and get real upside in the model if you're right this gets faster cheaper higher quality specialized models. huge for healthcare legal finance and other verticals needing accuracy compliance and integration this is the infrastructure layer the whole space needs as ai shifts from general giants to thousands of focused models $REPPO

Recently had an interesting convo with a web3 reporter. Enjoyed it and thought I'd share some insights here as well. The main point to hit home is that @reppo is not a data collection play, it's a data curation and incentive infrastructure play. Prior to Scale AI, data labeling/curation was farmed out to crowdsourcing platforms like Amazon Mechanical Turk, which was clunky and lacked quality control. The value Scale captured wasn't really in collecting the data or doing the labeling -> It was in coordinating and guaranteeing the quality of it. That's why they are $29B worth (probably more now), not because they can crowdsource data at scale. Reppo's thesis is that the coordination layer can be decentralized and incentivized through stake-backed markets rather than centralized contracts. On that, our TAM is the full market for human-feedback-as-a-service: RLHF, DPO, preference ranking, evaluation across all AI modalities (text, vision, physical AI) is enormous and growing fast. So what's the TAM? Let's explore

@oneill_c Great read! "What every company in the wave has figured out is that the only defensibility that survives the collapse of software cost is owning a model trained on signal no one else can see"