flows.wtf
4 months ago, we proposed a protocol for streaming funds to Nouns builders. The prop passed, and we built flows.wtf
- Open source streaming grants protocol and platform
- Recursive set of money streaming pools
- Nouns holders vote on streams
- Grantees gated by TCR
- Farcaster and onchain data pipeline, enabling:
- Apply for a grant via chatbot
- Tailored builder support for all nounsfolk
- An agent that writes gonzo style journalism about Nouns builders
In just over three months, we delivered more than we promised. Check a video of new features here
So far
Flows has been well received. We've onboarded 50+ talented builders globally, each earning part of the $10k monthly stream.
A few examples of projects we’ve funded:
- Digital literacy sessions for the elderly in Chile
- A community garden in Colombia
- Esports streamers around the world
- A skateshop owner in Brazil
- Athletes in Uganda
- Food for the needy in Nigeria
- A food forest project in Northern California
- Denim art from Latin America
Flows is funding some absolutely epic stuff. These sort of non-consensus small grants are exactly the types of decentralized opportunities we want to champion in Nouns. Check out the /flows channel to dive in.
To recap the problem
Building in Nouns has been tough for smaller projects. Before flows, small grants were effectively non-existent, and certainly not scaleable.
As a Nouner, everyone cares about funding different things, but For/Against voting requires majority agreement.
The best capital allocation system minimizes voter effort and maximizes funding for impactful outcomes. We're trying to solve effective capital allocation onchain.
We believe that effective grant-making is key to the future success of Nouns. Our incentive system for Flows must ensure that scamming is more costly than the rewards.
cost to scam > potential rewards
Working towards a solution
If it's more costly to fake impact than to work on a genuine idea, we can solve the problem of free-riders.
We partially solved this using a Token Curated Registry (TCR); you can read more about its incentives here. But even with incentives in place to curate a solid list of grantees, "ain't nobody got time for that", and it's unfortunately still open to collusion as well.
The game has changed. AI is now incredibly cheap, and we have onchain data plus Farcaster’s social graph. So, we can build AI systems to spot, question, and ultimately challenge underperforming grantees.
Directionally correct incentives are already in place with the TCRs, and we have all the data we need, we just need intelligent agents willing to do the work.
Technical details
Our long-term goal is to fight scammers and champion rockstar builders. We aim to scale Nounish grant-making to support one million builders per month. The tools we put together are:
- Streaming money
- ✓ Funds paid out every second
- ✓ Builders can be removed at any time if they stop performing
- ✓ Streams adjust in realtime based on votes and new recipients
- Incentivized Curation (TCRs)
- ✓ Slashable application fee disincentivizes fraudulent applications
- ✓ Curators incentivized to maintain quality list of grantees
- AI
- ✓ Farcaster data pipeline to check all builder updates
- ✓ Assess whether casts qualify as grant updates based on builder history, requirements, and grant details.
- [WIP] Agents to tag builders on Farcaster, ask for follow up impact proof, and ultimately challenge under-performing grantees if necessary
- [WIP] Agents to champion builders and help them accomplish their wildest dreams
While we believe the above mechanisms will help us scale the current program to $100k+ per month in builder salaries, we're not sure it will be sufficient at numbers above that. We plan to explore other mechanisms so that effective grant making can emerge from the incentive structure of Flows.
We're concerned that there's no clear incentive to find and reward those who spot underperforming or scamming grantees. We need an onchain mechanism to inform us whether or not someone is legitimate, continually, over time, in the context of their grant work.
We’re most excited about exploring small prediction markets for each project or grantee. We are open to more ideas on this front.
Our ask
We have been streaming $10k a month to builders in the grants program (weighted by Nouns votes) for the past 4 weeks. We want to 4x this over the next 4 months.
We believe we have the tools and data to effectively catch scammers and free-riders.
Even with our limited infra today, 94% of grants recipients have posted legitimate updates on the work they're doing (4/67 have not, but 2 are newly approved, and 2 are having trouble with Warpcast).
Plus, we have a huge number of incredible builders waiting to apply, including an Olympic bronze medalist. Great builders are ready to scale their impact.
Our Plans:
- Build Small-Scale Prediction Markets: We'll create prediction markets for each project to rank quality builders and incentivize spotting scammers.
- AI Agents on Farcaster: These agents will help builders scale their projects, write stories, and verify their impact.
- Personalized Grant Pages: Each builder will have a customizable grant page.
- Data Pipelines and Automation: We'll use automated systems and incentives to maintain a 90%+ quality rate among funded builders.
- Expansion: We aim to scale up to 250 builders per month.
Our Commitment:
We'll keep iterating based on feedback, adding new features for growth, and simplifying the platform. We plan to deliver beyond what's outlined here and hope this gives a clear signal of our direction for those interested.
The team
Rocketman and Wojciech. We've been building in and around the Nouns space for nearly 3 years. In that time, we have built an alternative protocol for launching Nounish DAOs, House of Nouns, and Vrbs DAO to name a few.
The funding request
We are requesting $20k USDC a month for 4 months to continue work, split between Wojciech and rocketman ($10k a month per person).
We'd like to scale the grants program to 12ETH a month, streamed to builders every month, for 4 months (up from 10k USDC a month).
Total ask: $80k USDC + 48 ETH
Questions
Have any questions? Want to discuss with us? You can reach me on Warpcast here or Twitter here.
flows.wtf
4 months ago, we proposed a protocol for streaming funds to Nouns builders. The prop passed, and we built flows.wtf
- Open source streaming grants protocol and platform
- Recursive set of money streaming pools
- Nouns holders vote on streams
- Grantees gated by TCR
- Farcaster and onchain data pipeline, enabling:
- Apply for a grant via chatbot
- Tailored builder support for all nounsfolk
- An agent that writes gonzo style journalism about Nouns builders
In just over three months, we delivered more than we promised. Check a video of new features here
So far
Flows has been well received. We've onboarded 50+ talented builders globally, each earning part of the $10k monthly stream.
A few examples of projects we’ve funded:
- Digital literacy sessions for the elderly in Chile
- A community garden in Colombia
- Esports streamers around the world
- A skateshop owner in Brazil
- Athletes in Uganda
- Food for the needy in Nigeria
- A food forest project in Northern California
- Denim art from Latin America
Flows is funding some absolutely epic stuff. These sort of non-consensus small grants are exactly the types of decentralized opportunities we want to champion in Nouns. Check out the /flows channel to dive in.
To recap the problem
Building in Nouns has been tough for smaller projects. Before flows, small grants were effectively non-existent, and certainly not scaleable.
As a Nouner, everyone cares about funding different things, but For/Against voting requires majority agreement.
The best capital allocation system minimizes voter effort and maximizes funding for impactful outcomes. We're trying to solve effective capital allocation onchain.
We believe that effective grant-making is key to the future success of Nouns. Our incentive system for Flows must ensure that scamming is more costly than the rewards.
cost to scam > potential rewards
Working towards a solution
If it's more costly to fake impact than to work on a genuine idea, we can solve the problem of free-riders.
We partially solved this using a Token Curated Registry (TCR); you can read more about its incentives here. But even with incentives in place to curate a solid list of grantees, "ain't nobody got time for that", and it's unfortunately still open to collusion as well.
The game has changed. AI is now incredibly cheap, and we have onchain data plus Farcaster’s social graph. So, we can build AI systems to spot, question, and ultimately challenge underperforming grantees.
Directionally correct incentives are already in place with the TCRs, and we have all the data we need, we just need intelligent agents willing to do the work.
Technical details
Our long-term goal is to fight scammers and champion rockstar builders. We aim to scale Nounish grant-making to support one million builders per month. The tools we put together are:
- Streaming money
- ✓ Funds paid out every second
- ✓ Builders can be removed at any time if they stop performing
- ✓ Streams adjust in realtime based on votes and new recipients
- Incentivized Curation (TCRs)
- ✓ Slashable application fee disincentivizes fraudulent applications
- ✓ Curators incentivized to maintain quality list of grantees
- AI
- ✓ Farcaster data pipeline to check all builder updates
- ✓ Assess whether casts qualify as grant updates based on builder history, requirements, and grant details.
- [WIP] Agents to tag builders on Farcaster, ask for follow up impact proof, and ultimately challenge under-performing grantees if necessary
- [WIP] Agents to champion builders and help them accomplish their wildest dreams
While we believe the above mechanisms will help us scale the current program to $100k+ per month in builder salaries, we're not sure it will be sufficient at numbers above that. We plan to explore other mechanisms so that effective grant making can emerge from the incentive structure of Flows.
We're concerned that there's no clear incentive to find and reward those who spot underperforming or scamming grantees. We need an onchain mechanism to inform us whether or not someone is legitimate, continually, over time, in the context of their grant work.
We’re most excited about exploring small prediction markets for each project or grantee. We are open to more ideas on this front.
Our ask
We have been streaming $10k a month to builders in the grants program (weighted by Nouns votes) for the past 4 weeks. We want to 4x this over the next 4 months.
We believe we have the tools and data to effectively catch scammers and free-riders.
Even with our limited infra today, 94% of grants recipients have posted legitimate updates on the work they're doing (4/67 have not, but 2 are newly approved, and 2 are having trouble with Warpcast).
Plus, we have a huge number of incredible builders waiting to apply, including an Olympic bronze medalist. Great builders are ready to scale their impact.
Our Plans:
- Build Small-Scale Prediction Markets: We'll create prediction markets for each project to rank quality builders and incentivize spotting scammers.
- AI Agents on Farcaster: These agents will help builders scale their projects, write stories, and verify their impact.
- Personalized Grant Pages: Each builder will have a customizable grant page.
- Data Pipelines and Automation: We'll use automated systems and incentives to maintain a 90%+ quality rate among funded builders.
- Expansion: We aim to scale up to 250 builders per month.
Our Commitment:
We'll keep iterating based on feedback, adding new features for growth, and simplifying the platform. We plan to deliver beyond what's outlined here and hope this gives a clear signal of our direction for those interested.
The team
Rocketman and Wojciech. We've been building in and around the Nouns space for nearly 3 years. In that time, we have built an alternative protocol for launching Nounish DAOs, House of Nouns, and Vrbs DAO to name a few.
The funding request
We are requesting $20k USDC a month for 4 months to continue work, split between Wojciech and rocketman ($10k a month per person).
We'd like to scale the grants program to 12ETH a month, streamed to builders every month, for 4 months (up from 10k USDC a month).
Total ask: $80k USDC + 48 ETH
Questions
Have any questions? Want to discuss with us? You can reach me on Warpcast here or Twitter here.
