[REDACTED] 2024 | The Path Towards Funding AGI AI Powered Grants

Captions as YouTube publishes them, usually machine-generated, so product and people names can be misheard.

1. 0:00

   well today I'm here to talk about AI GF which is not only a program for Builders but a movement uh at the intersection of exploring Ai and public goods funding um just for reference who here needs funding who needs funding bad oh wow I'm I'm okay all right I I'll catch you guys uh later on that um a little bit background about me um I'm plug AKA shot I'm focused on supporting teams and Builders across the near ecosystem really specialized in kind of Grants

2. 0:31

   funding mechanisms and also open source contributions um I'm at also the near Foundation low key focus on public goods and founder of potluck and also a steward for aipf and you could follow be on foraster lens near social at plug Rail and hopefully make an agent of Me based on my open social data and so here is ultimately the vision we're trying to build for um uh imagine a world where you could get funding automatically as you build

3. 0:59

   something users SW um a lot of you here are founders looking for funding and often we see ourselves not even just spending time focused on funding but even at times having a whole co-founder spending one to three months or 3 to six months raising Venture Capital raising grants and it's honestly a huge distraction and uh different incentive alignment to actually building a product that users want um so ultimately uh

4. 1:24

   through AJF through building out this open funding stag we're looking to not only solve the problem of making it easy for Founders to find funding but people with treasuries to actually find the teams uh that are uh qualified as well um and so at the same time um building a lot of these funding mechanism and these allocation tools I found that governance is really hard to scale and there's too

5. 1:49

   much processes there um so this is why we kind of t pivoted or not even really focused on the AI Asian part of the stack um as of now because we can build all the funding mechanisms and contracts and toolings to allow anyone to allocate funding uh but ultimately if it doesn't reduce overhead and provide that business value um what's what's the point we're just essentially a lot of

6. 2:12

   times in blockchains we're building a lot of uh barriers to entry rather than talking about the problems ahead um and we noticed a few Trends happening in the Eos system in web 3 in general Tech um well firstly we see the growth of AI um now especially the past 2 years llms are actually disrupting administrative work they may have not necessarily taken jobs but you could you just saw uh Louise actually use AI for the development of

7. 2:41

   projects people from every part of the professional life are actually using AI right now to reduce overhead and save money and then already automation is prevailing in web 3 we have me Bots we have uh so many different tools that uh uh really make blockchain a a prim gr for programmable incentives in money and at the same time now more so with web 3 that did web 2 uh we have more data

8. 3:10

   about funding and open data blockchain provides a source to provide information about traction pass grants data and there's a bunch of people building this space uh so that we have visibility not only on Venture funding but grants and uh subsequent traction at the same time time ecosystem funds are huge and people ask why are you focusing on public goods funding or funding in general in the web 3 space as opposed to web 2 everyone needs funding at web 2 as well um and um in the

9. 3:41

   reality is we're at a very early uh start of AI uh and it's really hard to get traditional foundations and traditional uh VCS to actually experiment with uh AI in the context of funding at the same time if you look at all these different ecosystems and this is just not evm blockchains um when a blockchain is started they usually have a token and that token Supply is allocated towards an ecosystem Treasury and that ecosystem treasury is usually distributed through funding and grants and if you look at just the nod evm top

10. 4:14

    20 blockchain market caps and then look at their ecosystem treasuries we're talking almost a hundred billion dollars that we can experiment with to give out to builders and a lot of times the projects that go to these ecosystems are the same projects so not only having C on how they're progressing in there but how they're progressing um in other ecosystems is very very relevant um as

11. 4:37

    well um so this has a huge potential in terms of cost savings if we save 10 to 20% on uh the operation budget for B's grant program which is honestly the usual overhead of what it takes to run a grant program um and at the same time I think a lot of these foundations and these crypto Cubbies are very open to experimenting with kind of building out um at the intersection of AI and GRS at the same time here's some things we also notice that um agents yield

12. 5:08

    better results when working together you saw right now in bit's example uh to generate you know a swap you had to uh you know get a price Discovery tool and conat it with other agent and if you want to make it uh a dow proposal for swap you have to plug into the Dow tool Le so um agents are just like hbid where each individual agent is kind of a subject matter expert and they work together to produce uh

13. 5:34

    these more experienced teams at the same time we see agents progen out composedly they can tackle a lot of the problems outside of public goods funding so in the context of building a due diligence agent to check whether grantees are doing their Milestones or are even uh doing the things that they said they did Um this can be used in Venture Capital this could also be used in banking and and other use cases but as an open indiv visual agent can be tapped into this public goods funding use case um and

14. 6:04

    then at the same time agents are only as relevant as the data that they're privy to um right now we've reached the deflection point in AI where most of these foundational models have already indexed public information um so it's important that not only we set up these new data pipelines that are open but we start leveraging private information so I'm really excited about Nyan and the and the integration that are coming sued

15. 6:28

    at the same time open source limits vendor lockets and it also speeds up collaboration and itation um and I'm a big proponent of Open Source and one of the key pillars about user owned AI is actually Building open systems that people can interact with so here are also some some challenges and some limitations um of AIP GF like handling funds is really scary um also the metrics G be gain and we have a lot of external dependencies on apis for like Twitter data LinkedIn data crunch based data so many external sources that

16. 7:03

    we have to put in context for funding a funding AGI it's uh it makes it really difficult to build this system um also there's a really hard timelines to understand a specific Roi even without AI um so how do we know if a project is successful and um that knowing that is a multi-year maybe even a decade time Horizon um and so although we can you know see whether a project that is

17. 7:33

    receiving funding would have gotten funding from subject matter experts and tubid evaluators we don't necessarily know whether uh that was affected of a long-term time Horizon and there's a couple of ways we can limit this uh but I'll kind of go into like where we're at in AIP pgf as a move bid uh in here and this is kind of how I look at the evolution of AI pgf so right now we're

18. 7:59

    at deware like we have things that look cool and that demonstrate the power of AI but they don't necessarily plug into larger systems um that's where we're at right now we're very early on this um and then eventually we'll have many agents and this is kind of the phase that we're in so different parts of the grant workflow and the funding workflow essentially building agents around this

19. 8:23

    and then kind of pipeline them into uh data pipelines where they can continuously learn and then have Conta with aade information and off trade information to uh to kind of optimize that workflow and then eventually plugging into a multi-agent workflow where now you have an RFP agent that develops RFP you have an application submission agent that submits the application you have an eval agent and

20. 8:47

    you kind of have these workflows where eventually you can actually build a whole end to- end funding Pipeline and then experiment with the micro grants program where you actually have an agent giving out grants autonomously with a small amount eventually at a funding AGI that can give out grants at allocated funding better than a human CAD um and so that's the vision and we are just at

21. 9:11

    demo Weare and my main point of this top isn't to tell you what we build and what we're building it's honestly to inspire you how you can start tackling a lot of these cases and then what actually it looks like um in here and so this is kind of a little bit breakdown of different stakeholders and uh different parts of the grant workflows that can be built uh and also even Within These workflows you could build multi-agent systems and then apply them to different

22. 9:39

    use cases I mentioned earlier um even having diligence on tees you could build a whole company around that individually you can build a whole company around uh getting qualified leads in the ecosystem U so I'm primarily looking for projects that have a business use case with this particular workflow that's willing to build a open Agent and then pipeline to the overall funding AGI so we have a few case studies if you want actually see what's being built all

23. 10:09

    the open source repos if you go to AIP gf.com exlore a lot of people have been building a lot of cool stuff um right now primarily we're working also on agent framework to more easily deploy funding agents uh but uh like we've built agents on potl potluck agent on B there have been people who built uh agents for determining budgets for proposals for evaluating proposals for evaluating elig eligibility the grant program to checking Twitter and using tools like Masa and other partners

24. 10:40

    you've seen here to uh actually do diligence on projects so many people are building different parts of the pipeline um within our ecosystem but also uh externally as well and we've got pretty good reception from the overall like funding ecosystem of kind of using AI to ultimately reduce this overhead so yeah what to build I'm always very for giving guidance for Builders on what to build if you go to AIP gf.com

25. 11:08

    ideas um all of this is a public board with requirements of exactly what you could tackle and also any relevant projects and information there um so yeah if you want to find more information on that uh if you go to AIP gf.com um you can you can actually we have a kind of small Builder grants program um for this but we also have these agents here and then pretty soon we'll be deploying our agent playground so this is a little bit about

26. 11:41

    what we're building at potlock uh right now I talked a lot about the funding agents layer uh but we're actually building a whole stack for anyone to deploy Ed toed Funding Solutions um so this is you know funding distribution of the contract layer the identity layer the incentive mechanisms impact tracking and building an all up all on social and then building different agents and applications um on top of that to

27. 12:04

    interface with this stack so yeah we're also coming out with potluck V2 actually Beyond just quadratic funding so anyone could actually deploy their own mechanisms on there but stay tuned for that so yeah here's the call to action like if you are a funding program and you want to experiment with the intersection of AI to reduce overhead to uh fund your Builders if you're a builder build a ipgf with us and if

28. 12:28

    you're a dream and want to replace your job with AI which is the primary reason of what I'm try to achieve uh because I think work is overrated and I'm trying to not have a job um yeah hit me up