 All right, welcome everybody to the October 19th Hyperledger Technical Oversight Committee call. If you've been on the call before, you already know this, but for those of you who are new, we do have two things that we have to abide by. The first one is our antitrust policy. Basically, there are a number of different people on the call, and we need to make sure that we are aware of and not participating in any activities that are prohibited under any of the antitrust and competition laws across the world. And then the second thing that we have to abide by is our code of conduct, which is linked in the agenda. Respect, be respectful of everybody on the call, their ideas and thoughts. For announcements today, we have the standard deaf weekly developer newsletter that goes out each Friday. If you do have anything that you would like to include in that newsletter, please do leave a comment for consideration on the link that's in the agenda. The second announcement that we have is that the TOC call for nominations period has begun, and it will end on October 31st at the end of day Pacific time. So if you are interested in running for the Hyperledger Technical Oversight Committee, please do add your nomination. There are instructions at the link that's in the agenda for you to submit your nomination. The last announcement that we have is that on Thursday, November 16th, there will be a workshop that is the Atomic Cross Ledger Transactions between Hyperledger Basu and Corda Ledgers using Hyperledger TechDive V2. And if you are interested in joining that workshop, please do click on the link to register for that particular workshop. Any other announcements that anybody has or would like to make? No, okay. So we do have three quarterly reports that have come in Calipur, Cacti and Fabric. As of yesterday, when I took a look, it looked like each of them had about seven reviews from the TOC members. I didn't see anything as far as questions or comments that were in those reports. I did, however, see that there was a question from Peter in the Cacti report about Hacktoberfest. As far as I know, Peter, we do not advertise that to anyone within the Hyperledger community. I think it's up to each of the maintainer groups as to whether or not they want to participate in that. But yeah, anybody know of anything that we specifically do for Hacktoberfest? I don't think Hyperledger has ever formerly participated in it, but if projects want to, that's great. I think you just use the tag on your bug for your issue and go forth and prosper. And Peter, Cacti is participating in that this year, is that correct? Yes, we've been a recurring user participant of Hacktoberfest. And it's pretty good for getting new people exposed to the project. Of course there's, the nature of it comes with a lot of drive-by contributions too, but we'll just take what we can get. And in that philosophy, it's great because there's always one or two people who submit more than exactly one contribution as well. This has just triggered something in my brain. I want to show you guys this. This is a website that's generated automatically for CNCF that, you know, you can, it picks up the issues that are good first issues and, you know, it kind of pulls everything in together. This might be something that Hyperledger could look into later when I have a bunch of free time. That looks interesting, right? And I'm sure it's open source code, so we could probably utilize it for Hyperledger and other projects there, so, Arno? Yeah, just, this is interesting. I was exposed to another tool apparently they have called Cielo Monitor, you might try. And it only covers CNCF and LFAI, but it does a bunch of like, gives you a bunch of dashboards and information about the state of a repository. It's also interesting. I only found out about that in the last couple of days and I have been working to try to get the whole. They have Clotrimuter, Clomonitor, Clowwarden and ClouUI. I've been working to get that set up so that we have better tools. So I'm working on it. Very cool, thanks. Okay, sorry, I had to go off and do my searches to find those things, so I think that's distracted. You guys, thank you for distracting me. Interesting things to take a look at for us to consider moving forward. Okay, so any questions or comments then on the quarterly reports that I may have missed that we need to bring up in this discussion? No, okay. So we'll obviously wait until everybody approves it or we reach our two week timeframe on those reports before we merge them, but it looks all like they'll probably be merged within the two week timeframe. So for upcoming reports, we do have the Q4 Hyperledger Sawtooth Report that is due next week. So we'll look to see that one coming in next Thursday. And then for discussion items today, the only thing I had on the agenda was the documentation task force update. Is there anything else that we do need to discuss today that I may have missed? No, okay. So then, Bobby, off to you. Documentation task force update. And Bobby, you're on mute. Yes, I know, I'm off mute now. Thank you very much. I'm trying to do that presentation. Get it in slide mode and then share my screen. Yeah, that didn't work. Hold on just a moment. I thought we're gonna go over some quarterly reports for a second. You thought there was gonna be some discussion there, huh? Yeah, it usually is. So I was like, okay, I got a minute. There we go. Should be good to go. Let me share my screen. Let me know when you see my screen. Yep, we can see the screen. Okay, excellent. So welcome everyone to the October 19th TOC call in which the recommendations from the documentation task force are being presented to the TOC for their feedback. The documentation task force started from the TOC's need to organize the documentation in the community and get it organized so that it's consistent, reliable and what everybody expects from the excellence of the Hyperloser Foundation. So we were working as a task force up until the mentorship program. And then we had put it into the mentorship program and got a mentee, the manager, who I call Arunama, the manager and then some other volunteer mentees that stayed with the project because they believed in what we were doing and wanted the experience and they'll present today. And I think you'll be impressed with them as much as I am because they are a fabulous group of people. So let's get started. Today's agenda, we're just gonna go over the objectives of the group and the goal of the group. We're gonna talk about the areas we worked on, those six areas. And then we found during those discussions of those six areas, we do weave in a little AI tools in the documentation because I don't think you can really have an effective enterprise documentation effort without AI now. So some of the demos we do are during the six topics and then some are overall for the Hyperloser Foundation in general and then some final recommendations for documentation for Hyperloser. Like I said, the agenda and now I'm going to turn over the rest of the presentation to Arunama. Arunama, it's all you. Yeah, thank you, Bapi for handing over the presentation. Hello, everyone. My name is Arunama Choudhury and currently I'm a master student at NIT Warangal. I have been working as a task force leader under the documentation task force for the past four months. I am also LFX90 under the documentation standards project and the committee chair of the documentation task force advance. Okay, so there are some problems that we identified with the current documentation that we are working on to solve it. There is lack of clear guidelines, lack of consistency, also some lack of standard templates and graphic library. So the goal of the Hyperloser documentation task force is to create and maintain high quality, accessible, innovative and community driven documentation that supports both users and contributors of Hyperloser technologies while upholding the highest standard of compliance and inclusiveness. Yeah, so these are some of the objectives that the documentation task force is working on starting with interactive and community driven content, accessible templates and style guides, modular documentation. The need of modular documentation arises from the fact that if there is an update in any feature, so for that we don't have to update the entire documentation. We can just update that particular module. Then we have user guides for different audiences, accessibility and education at once. So identifying the needs of the document, of the various documentation needs, we divided ourselves into six subcommittees to cater to each of the needs. Starting with GitHub templates, the chair of GitHub templates is Gianluca and I would like to hand over the stage to him so that he can share updates on the GitHub templates subcommittee, that is doing some wonderful work in the GitHub templates area. Gianluca, the stage is yours. Okay, thank you Arunma. Hello everyone. I'm Gianluca Capuzzi and on this project, I'm focusing on a standard documentation bucket. The name is GitHub template and it is related to documentation in order to support the different status of the project and also maintainers and the new maintainers in the life cycle. And next slide. Thank you. Okay, this is a list of the different status of the project. For example, one is the proposal, which is important for people who wants to propose a new project to the upper ledger ecosystem. And on the right, we have a list of example of support. For example, the organization and badges, user guides and dashboard. And next slide. And we have also an example about the proposal status. We have an example of support. For example, the checklist, which add in a sort of way, consent to people and support the people who wants to propose the project in the process. And also it's important, for example, to add other improvement, could be AI tools that supports documentation creation, updating management and so on. For us, okay, thank you. Okay, so the next subcommittee to cater to another section of needs of the documentation is the GitHub make the docs. And the committee chair of GitHub make the docs is Tripur. So I will again hand over the stage to Tripur so that she can share the updates of the subcommittee and what we have been doing there. Thank you, Arunima. Hi, everyone. Hope everyone is having a great time here in the presentation. And my name is Tripur Joshi and I'm the committee chair of make the docs. I am helping in turning repos into amazing guides. Moving forward. I will start with spreading some light on how Hyperledger is providing a very premium tool to every member in the committee, which is make the docs, which is not relatively known among the members and due to which it creates challenges in the project and they try to find other alternative solution which are like open source. And in fact, make the docs itself is a open source documentation platform. And even though it is a very valuable resource for us, I would like to provide some alternatives to it, which I will be talking about in the coming slides. Now this is a quick setup guide, how we install the make the docs. First of all, we need to install Python and PIP. For then we will install with the help of PIP, we will install make the docs and other essential dependencies. And through that, we will install custom themes, what type of themes do we want in our site. Next, we will go to the configuration of the site, how we want our navigation description to look like, what name do we want for our site. So this is a quick setup guide of how we will set up make the docs in our own device. Moving forward on the next slide, we, I am coming back to the alternative that I started with that even though make the docs is a very good resource, we have an alternative for that and we just get book. Why I am comparing both the documentation tool is that I am going forward with five points. First is setup. The first point, as we have already seen, while we were setting up make the docs, we have to download Python, then we have to download PIP, then we have to download other stuff with that and other dependencies. But with get book, we just have to do a simple login with our own GitHub ID and we can start working on our documentation. Second point that is ease of use. As you all know that as a technical writer, we should know markdown. If we are working on make the docs, we should know markdown. But with get book, there is no necessity for that. Even if you are a non-technical user, you can simply press on the plus button and you can start writing whatever you want, paragraphs, code, do you want to insert in it. Other is collaboration. The collaboration while working on documentation is very important. We need multiple people to collaborate with us, coders, developers, maintainers, everyone. But in make the docs, we require external tool for it. But in GitHub, it is already a built-in collaboration tool. So there is already a built-in collaboration tool. We don't need any external tool for that. Same goes with version control here. We need to create a separate GitHub repository when we are working with make the docs. As get book is part of GitHub, it has a seamless integration with it. So we don't need any external things for that. Same, the last point, which I think is very important is analytics. Make the docs does not provide any analytics whatsoever once this website is published, which I think that should be there and it is a built-in feature and GitHub book. Like how many people are visiting the website or SEO search is there working as and all the important things. So that's why I think that GitHub should be there. And while I like to present a demo about it, we all know how make the docs work, but let's look at how get book work. So I'll be sharing my screen. Hope everyone can see it. We just have to type get book and pick on the first link that comes up. Simple process. We will sign up with our GitHub. So, and this is my team library. I'm working, I created one hyper ledger standard documentation, user documentation for all the coders who are new to technical writing and don't know how to move forward with their documentation, how to create documentation and I have created few steps for it. Like how can they do and why it is easy for me to work on GitHub rather than we make the docs is if I want to add a page, I will simply do new documentation. Now, once it opens, even if I don't know markdown, I can just simply click on the add button. I need heading one and I can start working on it like documentation. Simple as that. And if I want to add a code, I can simply insert a code block here. I don't need any external things with it. If I want to share with it to anyone, with it anyone, I've already shared it with Bobby. I can do it as a reader, as a creator administrator, editor whatsoever. And if I upgrade it, I can also share it as a commenter. So it has multiple benefits of using the book. You can publish a space or website space for that, export as a PDF, share to an audience who can access it, all these kinds of things which I think is missing and make the docs and one of the accessibility things which is missing. And this is how we will use, get book. Moving forward with that presentation, I would like Bobby to share the screen again. So yeah, now moving forward is, I will introduce one of my favorite AI tools that I came to know when I was working with the team is Vemsical. And it is basically an AI generating, diagram generating tool, which helps you with different kind of stuff like flowchart, diagram, UML, wirefames, and you don't have to do anything with that. And it is easily compatible with Git, Google Docs and Figma and all the other stuff. Now I would like to show a demo of it again. Hope everyone can see my screen. If not, please let me know. You just have to write Vemsical, click on the first link that comes up. It's a very seamless website. And you know, once you start working on it, you get a hint, like a hang of it. If I create a new one, a board, and I want to create a mind map, suppose. And I want to create about, let's start with documentation. Like I need some mind map on document, documentation and I don't have any ideas for it. However, I'll expand it or anything. I will just click on generate ideas and it will do the work for me. Like what type of documentations are there, user, manual, technical documentation, API documentation, okay. So yeah, this, I can also zoom it out. And if I want to go deeper into it, I want to know about how can I start working with API documentation. I will again click on generate ideas and it will start doing it for me. Like quick start guide and all that stuff. So I think it's a very useful AI tool that we should like know about it. And it will also help us in being more effective in our work. Again, going back to our presentation, there's one another tool that I like to discuss. Bobby, could you please share the presentation again? It looks like maybe you're on mute. Can you see my screen? Can you see the screen in here? Oh, I was sorry, I was on mute and I was talking. So sorry for that. So this is one of my favorite AI tools and it was introduced by Bobby when we started working on our subcommittee presentation. And it was so helpful because we didn't have to do any work. We just have to like say what the presentation is about and it will start generating it for us. I would again like to show a demo for it. You cannot start sharing while others. So here we are. Again, it is as simple as memcicle. We don't have to do anything. We just have to say gamma and click on the first link that comes up on the Google search. If we want to create a new one, we just have to do this generate and I've lost the credit for it because I started building on it. And this was like the presentation that I created. I just have to write what the presentation is. Like even if I just had a hyper ledger as a task, AI task force, it will create the whole theme based presentation for me. I can change the themes or theme for it. Like the colors, if I don't like it, I can edit them. If the image I want to edit, everything is so seamless that I don't have to do anything. The conclusion, even these texts, I can also change them. Like it added with AI. I need to expend the text. I can do that. It will do it for me. And I can chat here and I can give my inputs, like turn it into a timeline or something. And it will do it for me. So this is one of the easiest app. If you have to go and have to create a presentation in a second, you can do that and you can start presenting. You don't even need your notes with it. So I think this is one of the most amazing AI tools that is out there. And that's all from my side. Thank you so much for listening. Bobby, you can share this with me again. Hello, everyone. Hope I'm audible. So I'm the committee chair of community template. So like the main goal of the template team is to like simplify the use of templates with the hyperledger community and sharing like the documentation created is streamlined as unified. So like we have discussed what is the role of a community template because like we know in everyday life or like while presenting our PPT or while like in any important presentation the like role of template is like a very great role. So like what the community template team did was like we reviewed the old templates of hyperledger and updated them with the new guidelines and logo redesigned the existing templates for the better onboarding experience. And like there are some awesome AI tools which we used to modify and like review the templates. These are like a chart GPT, Dal E2, Fit Jam and Jamboot, my personal favorite, luka.com which is like for the logo generation and template generation clip drop which can like generate images on AI prompt. So like the various templates we have in the hyperledger community are like white paper standard user cases, block basic, GitHub presentation, graphic set standard. So like we like did some research on these. Like review the and modify the template. So like I would like to show the Jamboot which is like a Figma AI tool. Figma is a design software which is used for the UI UX designer to enhance the user experience of the website and the app. So I would like to give a demo for this. Hope my screen is visible. So like this is the Jamboot. I have a, I'm sorry. Yes, and this is a Jamboot. So like from here, I can use the multiple plugins and all with Fit Jam gives us. So through this Jamboot, what I can do is I can ideate, ask a quick question. I can tell, teach me about this or anything. So like if I want to ask something like tell me about the project maintainer of documentation. So like it will help me ideate. Yes. So like this is the ideation phase like where I want to like create a dashboard for a project maintainer for the documentation. So like it gave me some points like create a virtual assistant that update and maintain project. And these I can like further ask to ideate it or quick question or teach me about or anything. So like I use this AI tool to like develop many templates and like the do the ideation and all stuff. It can like also code this up. Like if I give up a prompt to like, let's take a very easy example to like write a code to add two numbers. So yes. So it wrote the code in Python. So this is the code. So like it can like help us a lot because like generally like as a document documentation writer, we don't generally like to code these stuff. So we can use these tools to like write the code or generate the code, which can like guide the user or the newcomers to approach the project we have on the GitHub. So this is about the Jamboard. Bobby, you can proceed further. Okay. So next we have the best practices subgroup advance. So in the best practices subgroup, we are planning to build a batch builder. So the batch builder will integrate batching and tokenization seamlessly into the broader hyperledger ecosystem. Next. Okay. So for documentation badging, these are the steps we have planned to do for to gain as a part of the documentation badging system, we can gain badges for documentation creation and once the documentation badges earned for that stage in the life cycle, advancement is possible. That is, we can move to the next stage. Okay. So then we have the community dashboards, the purpose is to describe dashboards to aid community by persona. As a persona, we have four parts, maintainers dashboard, contributor dashboard, community member and community curious. Bobby, can you please explain each of these? From the next slide, we have each of the dashboards. Yeah, sure. I can explain them real quick. I'm sorry. I was on mute and couldn't find it for a second. So what we see the dashboards are kind of tied to your username so that when you log in, you can click on your dashboard and see, like kind of like the TOC has a list of the tasks that we have to accomplish for the week, something like that, but for maintainers and community people, they can assign themselves tasks where people can assign tasks that would show up on the dashboard. If you're a maintainer responsible for a project on your dashboard, you would see all of the five sections that you need to accomplish for best practices you would see, again, how far you are in each one of those sections to achieve your badge or token and tied into again, like a gamified, once you get to the whole thing filled out, you've done 99.9% of the work you get to move on to the next, whether it's a stage of your life cycle if you're a maintainer, whether your community curious, it's just another course in our identity series or whatever it is that you're doing, that dashboard would be tied to your username and your efforts in the community. Thank you, Arunama, it's back to you. So Bobby, in the next slide, I think we have some discussion. After this, we have the description of each of the dashboards so maybe you can go through them as well. Can you refresh once? So while I'm finding the right slide to also just incorporate something that we have come to discuss is with the learning token mentorship program, we're trying to work with Alfonso and his team to try to incorporate that lab into the idea of tokenizing the efforts. So here is the slideshow back, sorry for the delay. Can you share my screen? Let me know when you can see my screen. Yeah, we can see it now. So here we have each of the personas. I would like to explain these personas and dashboard. Okay, so for each of the personas which Bobby has mentioned, we have like four sections, user guide and checklist, batching requirements and token status bar and a what page and profile card. So yes, so for the first persona, which is the maintainers dashboard, like we introduced these four checklists where like in the user guide and checklist, we want that the maintainer explain the purpose of the user guide to the mentee at the screenshot for the user understanding and like for FAQs mentioned FAQs instruction and checklist. In the batching system, there must be a batch for the code contribution and also the clear criteria and information on how the mentee can achieve the batch and token. For the status bar, there should be like the status of the project where it is graduated or like in progress or like upcoming project. Also the visual representations of badges, tokens and awards, the timeline or history of badges when the maintainer or the project has on that batch. For the profile card, this is like a normal profile of the maintainer mentioning his or her current project, organization, mentee and et cetera. Coming to the contributor dashboard, we have like in the user guide, the contributor will develop the user guide reading the instructions again by the maintainer dashboard. Also a hand in the available checklist for the batching requirements, there will be batch for code contribution, meetups and like community meetups and events. Status bar is like the status of the mentee project, evaluation and review by the maintainer and also the award for successfully completing the mentorship period. And the profile card there would be the mentee name, his mentor project name and the previous projects and mentors for the like community members who are like volunteers. So there would be like, we will like introduce the open source to them. Then there would be like guidance and information on channels that how the community members can get involved, contribute in hyper ledger organization. There would be detail on how to like contribute bug reports, feature, et cetera. And checklist for a to track their progress within the organization. For the batching requirement, there would be like a batching for them. And yeah, so if the like a community member is like first newbie, new beginner, so there would be like specific badges for him like first pull request or like that. He can also like review the badge and history. The status bar and award page, there would be like, yes, the contribution metrics and like the what commit he has made, what documentation contribution or any special feature he has like done or the bug reports and also a ranking system we can introduce like where the top contributors, the volunteer contributors can be like showed in the present organization and also the profile card where the details of the each member would be there. Yeah, so this is the last, like the community curious people who are like would just want to like explore the organization. So this is for them, the user guide and checklist that we will like introduce open source project to them like without contributing, how can they help us? So we would be providing resources on and information on like how to navigate project repositories and all a checklist to get started and explore the project. There would be the batching requirement would be like there would be specific badges for the explorers and like demonstrate them like how to meet the batch criteria without contributing much and exploring more and helping the community. So the status bar and award page, this is same like everyone metrics related to number of projects they explore, information like how exploring projects support the open source community, how that is helping the organization. On that basis, there would be the ranking system to highlight active community curious members and in the profile card, there would be like the name description of interest in organization and like which project they are interested in and they like to deep down more and like contribute. So like this is about the four personas. So the last is like the documentation, documentation checklist where we will be having like these five points, the user guide, white paper, meetup slides, community presentation, available templates we have, available checklist we have. This is like for the documentation builder we have in our organization. Okay, thank you, Kajal. Next, we have the user guide subcommittee. So the user guide subcommittee is focused more on creation and accessibility. Currently, we are also working on metaphors, content creation hubs and dashboards. And it aims to provide tools to craft comprehensive and easy to understand community user guides. Bobby, could you please explain the slide of how we are planning to create dashboards for user guides? Of course. So we envision a community content creation center so that if you have to create content, you go to that section of the Wiki page, I personally see it in the metaverse where you can start building your content. We'll give you the AI tools, the style guides, whatever you need. So you're more worried about the content of your information rather than the formatting and the style. And there's great, this is just learn worlds, they drop that into your community content builder section wherever that is in your website or your metaverse and you just start building the content yourself. And again, we'll talk about that a little bit more. I do wanna run past onboarding quickly because basically onboarding, we just wanna have all of these great things accessible in two clicks so that any one of those personas kind of know exactly where to go seamlessly. And we're working with the onboarding task force for that. But when they get to onboarding, that the documentation they need and the user guides or whatever information they need is clear and accessible. So I'm gonna turn it back over to Arunama to introduce the very exciting next section of our presentation. Okay, so the next section, we have the AI trends in the enterprise documentation. Next slide. Okay, so we are planning to create AI-powered community user guides. These are the few areas that we are planning to cater to like a codes planner, AI-powered tools to assist in generating codes, outlines and learning activities. We are also planning to create assessment designer to swiftly create diverse besides assignments with AI with efficiency and accuracy. Our email creator, like engage learners with AI enhanced impactful emails fostering motivation. We also plan to use learning tokens like reward learners with learning tokens like some educational bonus bonuses to encourage them to learn more. We also plan to incorporate a feedback generator to deliver constructive and personalized feedback promptly. We don't know pun intended. We would like really focus to deliver some constructive feedback using AI. Another can be the eBook writer. And eBook writer, the AI crafted comprehensive eBook writer to seamlessly to systematically develop learning tracks to convey knowledge effectively to designated audiences and the content editor to elevate content quality with AI polished text, ensuring clarity and consistency. So moving ahead, we have a really interesting section but I would like to hand over the stage to Chianduka and he will be discussing his work on open source AI tool that he has been working on. Okay, thank you, Arunma. Yes, this is a very exciting topic for me, open source because I'm working with open source model. And the goal of the project is, actually is a proof of concept to have to provide a system like charge PT for people that don't want to read, for example, documentation and prefer to ask question to the system. Okay, in the next slide. Next slide. Yeah, this is a picture and it's an example. Okay, the technology background technology is the transformer. We also know a large language model like GPT engine and there is a list of main application in natural language processing, text classification, also called sentiment analysis, text translation, text generation, for example, from a sentence to a paragraph and also on the contrary, the text summarization from a paragraph to a sentence and question answering that is interesting for us because we want to use this technology to do that. And the last is means also we can use different application as combination to have best performance. Okay, this is the open source model downloaded from Agingface, the MIS-XLMR based on BERT and it has 560 million parameters. I test in my laptop and consists of two main components, the body pre-trained using the standard standard for the question answering data set and the head for the fine tuning, which means the domain application using hyperledger data. In the next slide, we see a snippet of the standard question answering data set and there are 100,000 question answering in English from English articles and this is a sentence which explain this kind of trick, the trust and learning to give you a sense of difference between fine tuning and training a model. Fine tuning is like taking your car to the mechanic and getting new spark plugs, whereas training is like getting a whole new engine. So it's a way to save much time for the training phase. Yes, this is the domain adaptation. Now I'm using a few documents from hyperledger existing the validation. And in the next slide, this is an example and this is the architecture, two main component, the retriever and the reader. Retriever works with unstructured tons of unstructured documents and retrieved the most relevant documents and retards to the reader. The reader is the AI model that extract the answer for the user. Next slide. Okay, thank you. Yeah, this is the current techniques use a pipeline with a stack library and the elastic search is the retriever and the reader, the open source model, XLM that already see. And next slide, I think. Yeah, I'd like to share my screen for a short demo. Okay, thank you. Okay, now the system is running and it tells me, ask me something about hyperledger documentation. An example could be, what is hyperledger fabric, for example, and also is the most active of hyperledger project. And it also give us the text snippet where the system extract the answer. Just one other, what is the mission of hyperledger? We can also expect wrong response, nutriting open source project under open governments that grows strong, sustaining computing and thriving ecosystem. Anyone want to ask a question to the system? Okay, no. How about what's the first step in installing fabric? We try. Okay, so step eight, I don't know, I think it's the wrong response. I think it depends also to the fact that I'm using a few documents and one improvement could be to add more documentation to the system to improve the answer. And okay, just to finish, if you can share again, thank you. Thank you. Okay, for future improvement on the left, we have the current model and technology using and on the right, we can improve the system. For example, using GPT for all, which uses billions of parameters, add other documents to the dataset using existing hyperledger documentation. And also using, for example, R-A-G technology, which is generative, it means that extracting the answer, it will be able to change words, in the response. Okay, in the next slide. Yeah. I hate to interrupt, but I do know that it is eight o'clock. I personally have to drop, but feel free to finish up if you want, but I will make a drop. Okay, let me just round up for the TOC. So our two recommendations would be to again, have a discussion over the points that we mentioned and how to incorporate them or to incorporate them. And then we also are recommending, which if you can stay on the call, you can see two additional task force, one for the Enterprise AI FAQ, which we're going to introduce as a lab. And then the other one is just to get this stuff in a library and we'll show you next how we think that should look. So thank you if you have to drop at the hour for listening to us. I have to drop, but I'd love to hear more and maybe a short section next week on the recommendations because I'd love to be able to take some of this. So how can I get at my hands on, the learnings you've come across? You know what, we can pause right here and come back next week and do the recommendations. If everyone- I think that would be best. Okay, let's do that. Then thank you very much for your time and for my team, the mentees, you guys are awesome, you did a great job. So thank you. Awesome, thanks. Thanks everybody. Thanks, yep, bye.