Unleashing Creativity with Blockchain Technology

Transcript

hello okay hi hi yes thank you perfect um hi my name is Paula I am the CEO of linim labs uh we are a web 3 software development company so we develop web 3 solutions for our clients so smart contracts full stack end to end um before I get into it I would like to ask everyone to bear with me up until yesterday I was dead certain I was on a panel um so to my surprise yesterday I really read everything and realized it was a presentation but I am excited to share some things with you today and I figured it would be a good spot to talk about some of the things we have been building so we've been around since 2016 and we've obviously that's been a while we

have evolved with the market so we started off with Consulting we actually started off largely here in Switzerland Consulting Banks helping them learn what the heck blockchain was over time we then built internal projects that we spun out different Dows that we built over the past three years we've been focused in on web3 development as an agency so 2021 was a big year of nft marketplaces we built quite a few of those I'm not going to talk about that today because it is not 2021 but we built quite a few of those this year we've built different staking contracts so we've currently built something with ether5 that's currently holding I think over 50 million in the contracts so we built

quite a big range of web 3 projects but uh it would be hard to ignore um that over the past three four months we have seen a huge um growth in the AI department and I think it is I think everyone here can agree it's most likely the future in some way and we've said that about blockchain for a while too and internally we decided like how can we marry those two together how can we have blockchain or how can we have ai support blockchain and this is something we started looking into around September of last year so um started thinking about it and in January we brought on a team of AI developers and machine learning

developers to you know see what we could do there so we brought on a small team and we were just like hey just break things just just break things in the corner what what's possible learn CTO I want you to learn about machine learning too so we put them in there and we were like you know where could this actually be of use and something We Came Upon was was Dow's machine learning is obviously really really great for managing data helping go through data and and suggesting things and one thing we see with Dao's is you know a huge issue in governance a huge issue and information overload where to start so we decided to focus in on there and we were like let's

start by building a prototype internally and see what we can do first off we were looking at all the Dows and you know how many are active and how many have proposals and one thing we saw is that over 60 percent or 55 of all Dows on Snapshot never launched a single proposal period um and only one percent launched more than a hundred and only about 12 percent launched even more than six so I mean most Dows that make the snapshot don't even come out with a proposal and then of the ones of the proposals that do go out only about 30 of them thirty percent of them get even 11 votes so

um yeah we We examined uh 13 000 and 51 000 proposals to come to that so one thing we saw in a place where we were like we can start off here is trying to help Dallas even figure out you know let's get a proposal out the door where to start so that's oh actually let me take it back so then we also analyzed all the keywords that you would see in proposals so um of of the the different proposals which words were more likely you know if included would be accepted verse rejected and in what frequency and we came out with this this cluster map which is um yeah one of the tools that is used to help analyze data and comically AI was

very very often rejected so it sits there in the corner but yeah so some the words that we saw that get accepted the most is governance trading voting liquidity and then a lot of these words that don't get accepted and how and harder to put on a map but like how the proposals are written so well actually so using this data they created an algorithm internally that basically would help write proposals that are more likely to be accepted by users versus proposals that are rejected um and so this is pretty basic but it's you know we created a pipeline that a user could come in come in with an idea for their Dao it would be prompted using the the llm that we created with the

keywords we'd be able to spit you out of proposal that you could launch um and yeah so we used few shop prompting and for those who don't know if you shot prompting uh isn't a specific type of AI model that uses all the data but then also uses a couple specific data pieces that you might feed it a smaller domain amount to come in with something even more specialized so obviously chat gbt is not that because it uses all the data but if you feed it a couple of pieces of data to specifically look at then you start going in the fuse shot area and while it is more contained and maybe isn't as doesn't have quite a breadth of x peace it is more domain knowledge so we so a

model around that um and just create an emergency plan in the event of Maker Now regulation we've added some examples used the algorithm and it basically sped out that um it's funny it says after an extensive discussion with the Dow I'm like okay all right sure I don't know who the AI was discussing with but yeah it proposes an emergency plan and it basically says to trade out 20 000 of ghosts quickly to you know usdc and it comes out with a kind of a specific plan there but it's it's you know it doesn't really say why and and everything about it we thought this was a good start

and then we went a little bit further and we said um so basically come out with a proposal in general we wanted to see you know you're a brand new now we didn't give it too much information besides you're kind of in The Lending and D5 space so pool proposals that are more from The Lending and defy area verse nfts or anything like that but um yeah create a proposal that's useful for the community and we'll gain popularity and oh wait no actually no sorry did not want to do that one um yep this is where I was putting it together yesterday um yeah so create a proposal that uses some examples for Dallas to attract token holders

it came out with just the Dallas white list protocol for V3 portals showing the growing importance of multi-chained Solutions in the D5 ecosystem it was a good proposal it it's like cool do it so we wanted to go a little bit further and have it create you know a couple of ideas because you you know maybe one won't work so let's see if you provide um several ideas so then it's provided more and said perhaps create a dow grants program and this would attract users to it and incentivize people to get involved or alternative launch a dow

ambassador program so both of those made sense but we did find them to be just a bit generic so then we went a bit further and we were like okay can you create specific proposals for these specific Dows please don't come up with generic ideas if you don't have any ideas just let us know you don't have ideas so then it went further and we did the bottle down and generates out and basically it said um the bottle Dow should create a partnership program for more artists to um yeah help increase their popularity and it decided that because the motto Dow had just had a successful collaboration that brought in quite a bit of um yeah engagement and they the or the

AI suggested building upon that um and then additionally for the Gen art Dao it basically said um there was an issue with the liquidity there so it suggested to increase the number of token holders and provide liquidity and suggested a secondary sales liquidity program to be implemented and so it gives a motivation it gives the proposal um and you know we've all probably used Chachi PT at this point chat CPT can do some of this but this is built on top of it and it's it's very specific two Dows it goes through it reads the proposals it doesn't come up with generic ideas um so yeah so that's yeah so this is some of the things we did create a basic um UI implementation

around it and we still have a bit to build on it this is a very early prototype just so our team internally can learn how to use this stuff um but basically we came up and we're like okay you come in with your down name let's say this is your brand new this is your keywords you know liquidity treasury LP lending you don't even know what you want to start out with so then it would suggest some proposals like hey start looking through these this could be relevant to you you can look through those if you don't think that they're good enough you can keep having it render more and then it would based on based on that if you liked the things it proposed to you it would create several ideas over so you could Implement a lending platform you could create a liquidity

provision you can establish a strategic strategic partnership and so it'd give you nine ideas and you could pick whichever ones you thought was more relevant and be like hey actually generate the full proposal for that so if you picked one you could Implement a lending platform and it would give you the description the reasonings why and you know the technical specification what it would require to actually build it you can then expand on it more but this is how you would start off with the first version of The Proposal and you know I think we're at a place where um chat uh is is exciting on the AI front but I think we're at a place where it's like okay what's next um and some of the things that you know

could happen here if if that's something that people want it to happen um is you could you could potentially connect the AI with that Dows in general and so not only is it pitching ideas but it's also perhaps it can connect to you as a person and help you as a person actually read go through the proposals on Dows that you're engaged with and help you decide which way you should maybe vote so it's not as much information coming to you specifically if you wanted further you could connect it to the gnosis multisig and this is where it gets a little wild but you know the AI could potentially make the decisions if you wanted it to do that and a lot of this the Dow mechanics that you know we like in theory we like the

decentralization we like all that but the people aren't really going through the hard work of it you could connect AI to it and you could optimize it in that way I mean you do come into questions like you know will the AI take over I don't know but you know that is a potential next step of where it could go um internally we've just had additional fun just having the the AI team on on our team so they've also you know idea for anybody with the company we've also done things where we've given it all our docs all our employee handbooks all our notion docs access to some of our chats and it's basically like you can ask it like what are the values I mean that's kind of boring but um you know who's this who's this person what's this person's job and it can just

spit it out what's the vacation policy who's working on this project and that's just kind of nice and efficient because I mean those are just the questions that everybody's always asking each other so we um we put that I deleted the slide that now I wish I had and my CTO had put one where it's like is Paula funny and I was like I don't know what data you gave it for that but it's like oh she's nice um internally and it's something that we already and we've started working with some of our clients to see you know what kind of overlap they could have and started working to build some things like that but yeah we actually feel quite confident with the team that has has joined since January and their understanding of both web 3 and

blockchain so if you're interested in developing in that specific Niche um chat with me or just general web3 but yeah thank you so much um yeah I'm Paula