GenLayer Community AMA

Hello guys, Happy New Week! I hope we all are doing absolutely amazing. Change of weather during the season is a first approaching. I've been down with a flu for a bit but I'm feeling much better now. Thanks to those who checked up on me. Today we have our head of ecosystem. Ivan with us, he'll be answering all of your questions, sharing insights and everyone will get to the opportunity to meet the pop-tweets at the end of the EME. Before we get started with the questions, let's just take a brief recap of the week. Episode 5 of Genesis is out by our co-founder Albert, such a beautiful episode if you haven't watched it, ensure you catch up. He also published a new article, Permission, Isnt Overdigged. It's our expose and important distinction in the agent's economy. So definitely give it a read, it's weather read. Also, we saw an interview at the 2013 49 highlights, few charing our new CEO of the Gender-Layer Foundation at Gus. He got interviewed by Sanjay, one interesting highlights, and their interview was a gender-lier testing against 179,0842 historical polling markets. We politely achieved matching outcomes without price, feeds or human reviewers. So ensure to watch the interview if you haven't as well. Currently, there are USC campaigns on rally, live. We have the Genesis episode 4 and 5 clip-in campaign. The Justice needs an upgrade campaign and two factor finance as well. So don't miss out all these opportunities. And Cubsan, I had Anna Maria, the rally's BD lead on our space last search date, answer questions and clarify your core inconse about rally. So if you missed it, you can always go back and listen to the recordings. It was a great space. If I've seen questions about rally, if you want to participate without having the ROPs, you can also participate in rally campaigns and you have new ROPs. You need at least $1.2 worth of eth on base. There are always ROP campaigns joining, so pick a campaign, understand its core metrics, and use that knowledge to improve your entries and stack up ROPs. Also, we have two special quests live on the portal. Do not miss the opportunity to stack up CPs. So let's warmly welcome our head of ecosystem and creator of Internet Core 5. And hi, Edwin, how are you doing? Happy new week. Hi, I'm here myself through you. No, it's okay. So yeah, I'm glad to be here. It's a very cloudy day in Buenos Aires today. It's been cloudy for the past few weeks. Not very common in this time of year or somebody called. I'm sorry that you were ill. I had no idea. I hope you're feeling better. I'm glad everyone is listening right now. I want you want just one thing like since Edgar's game CEO, my new role is chief of ecosystem, which is mostly the same thing, but just a changing role in the organization. Okay, well, thanks so much for the qualification. I guess I'm doing much better. Thank you so much. Before we dive into the community questions, is there any updates, announcements for the community? You would love to share. There are many, but I think that we will answer them in some of the questions that the community has been asking. Okay, fine. So without wasting time, which is guest started question from Vazan. Hello, Ivan. Hello Hills. I wish you a great day ahead. Wish you to return. Any specific milestones or challenges left for the day? Any specific milestones or challenges? What I can say that it's related to another question is that the day is coming very, very soon. Everything almost everything is prepared. We are working on some regulatory issues and like finishing some platforms and stuff, but you will hear this week. You will hear news about this. So I'm very happy that finally the day is coming. But on the challenges, no, there's no challenges. Mostly about finishing some documents and some technological platforms, but everything is on track. And which aspect requires currently most of your time. So my time is split among different objectives. The first one is the test net for which. We are with sentiment message yesterday to all the body data that Bradbury face to. And so, we are going to change everything in the next few days. So we have a new reset of Bradbury, where we are going to try the test net clerk code to see if everything is stable. And if everything is stable, then we will go on to test net clerk. We've been running this new node binary for several weeks now, running several bugs that we found. And also using AI as well to try different scenarios. And we are doing some tests that we haven't done before preparing to main it. So my, my, like, let's say that one third of my time is spent on, on managing the test net. I don't know if you guys know, but we have an agent called agent deck or, which is the main character from. Blade Runner. Because we have test and passimov test net Bradbury test net clerk and the obvious continuation was. To use Philip K. Dick, but just there his last time was not the best one for letter D. So we went with record and with this agent, we are organizing everything that we can with the body data. There are telegram groups and this code where they can ask these questions asking them to top up their accounts, checking out the status. And I spent a lot of time in this agent, which is automatizing everything because as you guys know right now, we have 45 validators on test net. But for minute, we will need to have 1500, 1500 validators, which is a lot. And we got to manage, manage them all by hand. So that's why we have been working on the card a lot. And yeah, like today tomorrow, we will start Bradbury face to, which is really, really good and really big milestone. And one third of my time is working with ecosystem projects, for example, we have the denari accelerator working with 11 startups. We will announce and present the start up soon. They are recording videos, explaining what their startups do and explain the teams. We have very very strong teams working on this and several several really interesting startups. We are using general layer in novel ways in new ways, some of them we have thought about in our request for startups, some of them are new ways that we haven't thought about. So we're really excited about this. We are currently running like a big curricula where each week we are presenting a new subject and there's an expert guest who comes and delivers this kind of keynote. Then they have one and once with the denari lab team, which is the one who are running the hackathon. And I'm also talking to them on a weekly basis, helping them. And yeah, just trying to understand what's the best use for genera in their projects, like three of them are already implementing the solutions. So that's really nice. And we're also working on bringing together apart from rally and collective memory of some other like fan markets or the fun. We're also working on bringing in like funded startups that they have traction to use general layer to solve some of the problems that they currently face while scaling up. So we're working a lot on also bringing more more startups that will start using general layer after main it. And the rest, the third part of my time is being working on the ecosystems side of things, not so much on projects, but on specific partners. For example, we are revamping our partnership with more fuse, which is an inference provided the centralized inference provider. And we're going to announce how the how the different validators can make use of of their new proposal that is going to come with many free tokens for them to test different things that we need. And also on the security side on the all run side, we have many different things preparing for the day and for main it aggregators, while it's a lot of things different companies and each of them. We have like I don't know between free and 15 different companies that we are talking to to decide which ones the best fortune there. But the good thing is that in most of the things we have already closed the deal on either implementing on the integration or the integration has already been done. So that's a little bit what I'm spending my days this days. Okay, next question is from Hamid Salman one he's got two questions. The first one was the one was the important partnership on number. You are watching that will show generally has become the main place people used to solve disputes in the AI agency economy. So the main number that we're watching every day is the amount of decisions made how many times the consensus has been called to make a decision. We are doing many of these decisions ourselves, but most of the decisions nowadays come from rally and from collective memory and there are a few other startups, some of the ones that I mentioned just before this ones in the previous question that are using and creating more decisions, but that's the main. So the main thing that we are looking for that's the main number, but it's not only to solve disputes. That's why we called the we call the gender here like an education there because actually occasion is broader. So we are going to include more things than disputes like in the case of rally for example, you're putting the trust in general here to decide the impact score of each of each tweet. There's no dispute there, but general here is doing its work. So yeah, numbers of decision made each day. So the question validators use different AI models. How will you make sure that similar disputes get similar decisions over time. So the results stay fair and consistent. So the main idea is for them to you to buy there's also use different AI models and generic second sentences among all these models. In the case of simple questions, if you read a news article and you ask a question about the news article, I don't think that's going to change over time, right. The news article is going to stay the same. The question is going to stay the same. And the result is mostly going to stay the same because it's pretty easy for LL models right now to read something and get information retrieved information based on the question. What will change over time. So these things, I don't see that use cases like or I don't see that it will change with an alam's getting better and faster and more intelligent. I even see a use case in the future that if you have a simple question, you can just use a smaller model to reply to that question. So it's more efficient. It's cheaper and faster. But the interesting part is going to come when you're asking LLM models questions that they cannot solve like all of them, they cannot solve them right now. For example, if you pass them a big, a neat happier with 400 million lines, like no model today can handle a 400, like that amount of information. Like we are right now and between one or two million context windows, two million tokens of context windows. So they are not able to understand that or they are not able to understand videos on on the same questions and some images. What I do see is that as LLM's continue to improve which they have been improving and they will continue to improve faster and faster. We will be able with gender to apply the consensus to many new use cases to many new topics to many new ideas. So in this case is worth sometimes the response is under I am determined, meaning that there was no consensus between the different other models as they get better and better, they would converge to the truth and they will be able to answer these questions that they can answer right now. Next question is from Glacier Hi, I'm Van and Hills, I wanted to ask if you could give an idea of when the community will be able to test delegation and whether this will be done using test students. When the community will be able to test delegation in Clark, for sure. I mean, you can already test delegation, but there you won't see many many things like I don't know we already have many ideas there for. For example, there will be a new platform from general labs called Jen's taking where you will be able to delegate from the studio i'm not sure but I think that if you go to the studio right now you can delegate. But in the internet Clark for sure you will be able to delegate and I think that in Bradbury face to as well. I need to check that but yeah from the general layer CLI I think that already you can delegate so it's not much about the technical side of things is more about that just doing that and because the idea to start like taking a delegation is part of the economic layer. Like you know that as we were testing more the structural things and the binaries and the infrastructure and code itself in Bradbury were testing more the AI side of things and in Clark were testing the economy.