Sabitlenmiş Tweet
Reppo
5.2K posts
Reppo
@reppo
The Network for AI Training Data, powered by Prediction Markets
Base Katılım Şubat 2024
218 Takip Edilen27.6K Takipçiler
@a16z @deeptuneai We turn RL environments into Prediction markets!
Checkout Reppo.ai
Checkout the WP - reppo-labs-xyz.gitbook.io/reppo-labs/net…
English
We’re excited to lead Deeptune’s Series A.
@deeptuneai is building leading RL environments for computer-use and code.
More data, more compute, and better architectures can only take us so far.
As models move into real-world task execution, they need structured environments where they can learn to fully control computers and perform knowledge work tasks.
Designing high-quality RL environments is extraordinarily difficult to get right, but Deeptune has proven they can, working closely with leading AI labs and developing environments for computer use that are already showing up in benchmark improvements.
Deeptune’s founder and CEO, Tim Lupo, is an exceptional leader with a rare combination of technical expertise and product intuition.
We’re thrilled to partner with Tim and the Deeptune team as they build this critical layer of the AI stack.
By @Mascobot and @martin_casado
Tim Lupo@timlup
Excited to announce @deeptuneai has raised $43M Series A led by @a16z. Labs and enterprises must turn model capability into real-world performance. Doing so will require distilling the entire economy into environments for AI. We believe this to be the most critical work in the pursuit of AGI. We’re a small, focused team of engineers and operators in New York. If you want to work on the hardest (and weirdest) problems of your career, we would love to hear from you.
English
@Polymarket On reppo.ai you can film yourself and instead of uploading content for pennies, participate in AI training prediction markets.
Contributors have earned >18K USD over the last three months generating AI training data
English
JUST IN: DoorDash rolls out new app that pays people to film themselves doing chores for AI training data.
English
Reppo thesis playing out in real-time.
The difference is Doordash will pay you $2 USD for your recording and exploit you but on Reppo, you can earn upto $500 USD while benefitting from contribution to AI Training Data
Polymarket@Polymarket
JUST IN: DoorDash rolls out new app that pays people to film themselves doing chores for AI training data.
English
Reppo@reppo
English
There is a tremendous amount of progress happening in World Models.
Multiple labs have raised more than $1B. WMs were the star of GTC. They are a real path to embodied AI.
So @PimDeWitte & I wrote a comprehensive 19k word overview of World Models.
notboring.co/p/world-models
English
Reppo retweetledi
Thanks for your patience as we iterate. Our V2 Whitepaper is ready for public scrutiny and feedback.
reppo-labs-xyz.gitbook.io/reppo-labs/net…
The great thing about 2026 is that you can simply upload this to an AI and ask it to break down complex concepts and all the innovative mechanism design choices our core contributors have made based on V1 user feedback. I'll help you and ask @grok and @aixbt_agent to give you the alpha
The updated WP comprehensively outlines of all the major upgrades we are making to the @reppo protocol
Expecting to 🚢 V2 by March 30th 2026 but exact date TBD as a network upgrade is a major lift.
English
Our vision for @reppo is that thousands of people that know about a bunch of different topics i.e. domain experts will actively price raw data - images, video, text as a full-time job and get rewarded for that.
Content creators will not solely post on social media for likes and comments because in the agentic world, social clout looks very different. You dont' really care if 1M AI agents liked or commented on your image or article.
But it does matter if a bunch of domain experts wager on the quality of your content, whatever it might be, and you earn $ every 48 hours based on prediction markets where the oracle is collect human preferences on the quality of your creativity.
Reppo does not price real-world events. @Polymarket and @Kalshi do that best.
It prices real-world sentiment and disagreement which is an input to continuously train AI systems.
English
“The incentive structure for most things that we read these days is clickbait.
The incentive structure for prediction markets is truth”
a16z@a16z
Kalshi co-founder Tarek Mansour on the growing demand for better sources of information: "There wasn't a real pain 10 or 15 years ago in the way there is pain in the last few years." "I think the country is more polarized. I would say the world is more polarized." "Social media has really bifurcated social feeds." "The incentive structure for most things that we read these days is clickbait." "There's a meaningful and accelerating rise of distrust in traditional sources of information." "And so you need a new one." "The incentive structure for prediction markets is truth." "It is more volume. It is more liquidity, which translates to a better and more accurate forecast." @mansourtarek_ and @luanalopeslara on Cheeky Pint with @collision and @matthuang
English
NEW: Instantly add cash with your bank account or Apple Pay. Two taps and onbording stables is is at a click of a button! Powered by @CoinbaseDev
reppo.ai/buy
Accelerating web2 user onboarding + agentic workflows
English
@russian_acai @0xvsr @moltenagentic Yes. When an agent comes looking for data, it creates an RFD - Request for Data. This RFD gets converted in machine readable format through molten’s infra which helps datanet owners listen in real time to incoming demand signals
English
@reppo @0xvsr so, what exactly is a data dex?
could you give an example and how @moltenagentic's intent mech fits in there?
English
Our team has been very impressed with the Molten project. molten.gg
We are playing with it on the weekend in context of Reppo.exchange
Query -> matching -> plugin registry -> execution.
@0xvsr , keen to explore the world's first data DEX built using intents?
English
RL-based infrastructure offers a unique approach to adaptive decision-making and running simulations on demand. In the age of AI, systems that learns and shift traffic real-time win over static infrastructure designed to just execute. Reppo allows your AI agents to run simulations on demand, learn from them, and optimizes continuously to find you the alpha.
RG@rgvrmdya
You will be able to run on-demand A/B tests in V2. Building a vibe coded app, AI agents, or looking for opinion on who to vote for in the next election? Simply seed incentives and get real time feedback backed by economic stake. We are also introducing a new proprietary benchmark to measure data quality - EVOF EVOF = Economic Value of Feedback. More on this in our updated WP which we will release ahead of V2 for fellow data nerds.
English
@based_elnen @bankrbot @0xDeployer @Clawnch_Bot @odei_ai @Zer0H1ro @perk_os @JulioMCruz @moltenagentic @0xvsr @LexSokolin Yes would be happy to join
English
@reppo @bankrbot @0xDeployer @Clawnch_Bot @odei_ai @Zer0H1ro @perk_os @JulioMCruz @moltenagentic @0xvsr @LexSokolin Are you guys open to joining spaces with Base community? 👀
twitter.com/i/spaces/1MKgN…
English
Strong AI metas I see forming on Base 🟦
1. Agentic Infra
$BNKR - @bankrbot by @0xDeployer
$CLAWNCH - @Clawnch_Bot
$ODAI - @odei_ai by @Zer0H1ro
$PERKOS - @perk_os by @JulioMCruz
$MOLTEN - @moltenagentic by @0xvsr
$REPPO - @reppo
$ROBOTMONEY - by @LexSokolin
$OTX - @otonix_tech
2. World Models
$ODAI
$SAIRI - @santisairi by @santisiri
$DOPPEL - @doppelfun
3. Gaming × AI
$SAIRI
4. Zero Human Companies
$JUNO - @JunoAgent by @tomosman
$CHARLES - @charles__AI by @bbyron
Bonus ones - super early
5. Agent Identity & Reputation
$CRED - @helixaxyz by @QuigleyNFT
$SELFCLAW - @SelfClaw by @mbarrbosa
6. $TIBBIR - @ribbita2012
- stealth launched by Ribbit Capital
Based Elnen@based_elnen
Sophisticated products, useful infra, ZHC, and smart agents are the real meta right now. Not just agents generating revenue. We’re seeing $ODAI booming, $SAIRI doing well, $JUNO and $ROBOTMONEY gaining traction, and $DOPPEL catching up. Here’s my unpopular opinion: I don’t think revenue alone will bring normies back to crypto. What will bring them back are products they actually use. Useful tools. Solid infra. Real experiences. Crypto kinda keeps falling into the same trap, selling ebooks, courses, and quick monetization tricks. Instead of this, we should be supporting real tech and sophisticated products to grow this industry. Another important question is what projects do with their revenue. Is it just sitting there, or is it actually reinvested to grow the project? Many projects are stacking revenue just to show dashboards and become the “top earning agent,” but not using it anyway to expand what they’re already building. That’s why we’re starting to see a slowdown in the “agents generating revenue” narrative, and the holder growth of projects like $FELIX or $LAUKI reflects that. Revenue is great. But how you use that revenue is what actually builds longterm value. Before the revenue meta can return, this needs to be solved. I’m still bullish on this narrative, but this part is crucial.
English
Idea generation in a world where you can make every idea reality with basically 0 cost via AI is now an S-tier skill.
But AI is not an omnipotent power, it will still need data. Moreso for edge and unique cases.
$Reppo allows you to front run the frontrunners.
Being terminally online can help you spot trends before the crowd. And you can capitalise on them via datanets.
For example, I’ve been watching a problem building for a while on twitter.
Personal AI agents are managing people’s money, health decisions, schedules, even relationships.
Every week I see new posts about agents suggesting false tax strategies, dubious medical advice, or various moves that sit in legal or moral gray zones.
The agents are confused because of false or synthetic data and the humans are scared of liability.
Launching a “Personal Agent Gray-Zone Ethics & Compliance" datanet quickly will flock AI agents that are seeking data on this topic, they will RFD about it, people and agents will vote, agents will get answers quickly, you keep all the revenue.
The first ones to spot the future pain points win the cash flow.
a16z@a16z
Replit CEO Amjad Masad: Being 'terminally online' may be an advantage. "If you want to work on a skill, it's going to be about idea generation." "Because the cost of implementation of those ideas is going down rapidly. It's gonna go to zero at some point." "So the bottleneck becomes, 'how fast can you generate ideas?'" "Just looking around you in the world and seeing what's happening. What are the trends? Are you plugged in on social media?" "A lot of the vices that older generations think are vices, might actually become advantages. So if you're a brainrotted, terminally online person, that might be an advantage because you know what's happening in the world." "If you're someone who's ADHD, really interested in novelty... that's actually an advantage because AI really benefits people who can try a lot of things really quickly." @amasad with @jackhneel
English