{"id":"D0mI_jbwkuI","title":"NEAR AI Ecosystem - What Did You Ship This Week? #12 by NEAR AI & NEARWEEK 🚢","description":"NEAR is the Blockchain for AI\nWebsite: https://near.org/\nNEAR AI: https://near.ai/\nIntents: https://near.org/intents\nChain Abstraction: https://www.proximity.dev/\nSharding: https://nearone.org/\n\nFor the devs\nGitHub: https://github.com/near \nDev Docs: https://docs.near.org/\nDevhub: https://devhub.near.page/ @NEARDevHub\n\nConnect with socials\nTwitter:   / nearprotocol  \nBlog: https://near.org/blog/\nMedium:   / nearprotocol  \n#NEAR #nearprotocol #NEARAI","channel_name":"NEAR Protocol","published_at":"2025-05-22T00:00:00.000Z","thumbnail_url":"https://img.youtube.com/vi/D0mI_jbwkuI/hqdefault.jpg","transcript_text":"near AI. What did you shift this week? Welcome, welcome everybody. Uh my name is Cameron Dennis. I am on the new AI team and this is the 12th episode of What Did You Ship this week. Uh we took a bit of a hiatus last week uh to unite a lot of the amazing speakers that are here today. Um we have a jam-packed agenda and so without further ado, we're just going to hop right on into it. But if you are tuning in for the first time, this is a live stream that we do pretty much every week where we highlight or give builders the the space to highlight what they're up to. who think of it kind of like a open engineering standup of sorts where they really just share what they've been shipping over the last week. And to kick things off, we have Hong Yin from Fallen Network. Hong, how's it going? Welcome. Hey, great. Great. Thank you. It's awesome to have you back as always and uh this is not your first time on and so you you know you know the ropes. if you could just let everyone know exactly uh who you are, what follow Network is and then unpack what you ship this week, we can dive right into it. Okay. Okay. Uh there are a bunch of very interesting uh things we have been developing but uh yeah mostly today I'm going to cover all the work related with uh AI and TE. So uh brief intro so uh I'm hung co-founder and uh c CTO of fa network uh and fa network is just a you know uh blockchain ecosystem and now we have a few product everything is around uh te trusted execution environment so just in case you are not familiar with te t is a special uh hardware it allow you to run something uh with the privacy and integrity so imagine you can put something in the TE and uh the TE can uh just run a program as is without any code change. The same time it can give you a proof and with this proof you can be convinced that this program is fully privacy preserving and this program cannot be uh tempered uh tempered with uh even if you are the the uh operator or uh the hardware owner of the hardware. So um yeah let's go through what we are doing this week. So I'm going to just share the screen of this window. Okay. All right. Yeah. So yeah this is about this is just our official website. So I will just skip it. So uh today I I have uh roughly four things to share. So uh first thing is uh MCP. So uh I think probably most of the people who are uh familiar with the recent development of uh AI have heard about MCP. MCP is a way to extend um your RM your or your AI agent to access some tools. For example, you can connect your AI agent to uh say a blockchain MCP so that it can read and write to the blockchain. And traditionally, if you run MCP, you run MCP server together with your program uh with your w with your AI client together in your uh say your laptop or or your PC. And this is uh it turns out is not very secure because you know you always run some uh binary maybe you haven't checked the content it can read your uh computer and even st uh steal your uh sensitive data. So uh of course an easier way is that we can just uh move the MCP from your computer to a remote server. In this way, the MCP server itself and your AI uh program is isolated. But then another problem becomes if you're going to send say your private keys or any sensitive information to a remote server, how can you transfer the remote server? So this is how uh you know TE can play. Uh so we put the MCP server in the TE and in this way um the T uh TE guarantee that uh the program you're interacting with is exactly the program you want to run. So in this way we can ensure your sensitive data will never get leaked outside of the uh MCP program you want to deploy and there's nobody can uh manipulate how your data is uh uh is accessed within this uh MCP server. uh and uh to make this super super simple, we have already deployed a few MCP servers uh on our T cloud. So as you can see here uh you can see a list of the MCP servers. Those are just a readyto use server. You can just copy and paste the server uh to your uh RM program. So you can start using that. And one of that is for example near blockchain MCP. And in this MCP you can read the blockchain uh content like a read from smart contract. At the same time you can even embed your private key uh in your local environment and send it share it securely with the MCP server. So it can sign transaction on behalf of you. So yeah this is uh one thing and uh of course uh those MCP are just uh developed and deployed by the official team. What if you want to uh develop your own? So uh we offer off uh offer a very uh easy way to do so. So uh just uh two uh two weeks ago we shipped a new feature called the templates. So in the past I already shared how easy you can deploy the uh virtual machines to uh TE. But uh can we make it even easier? So uh the answer to that is uh template. So here we have a list of the template and it's still growing and as you can see we have seven MCP templates uh one A2A template and a few others. So here uh for example if you want to deploy a uh uh like a uh MCP server uh uh D5 lama uh then just click deploy and uh here what we do for you is that just uh uh have all the configuration prefilled and all you need to do is just click deploy and then it will uh jump back to the list of your virtual machines and after a few uh minutes this will be deployed up and running and you can just uh copy and paste the uh URL of this MCP server and use it by yourself. You can even uh deploy your own and get it listed as a template on the cloud. Yeah. So uh yeah, this uh is a very brief u uh update about the cloud feature. On the other side, we also have a bunch of uh interesting updates uh around AI. So uh uh just uh in case you are not familiar with uh uh uh uh the integration between us and the open router uh so open router is an aggregator of AI uh uh APIs. So if you register open router account then you can access uh thousands of AI models uh just using one account and uh actually this is a web two uh product. So it's very popular among AI developers and recently open router just added the TE uh uh uh AI model support and we are proud to be one of the model providers. So for example for this llama 3.370 billion uh model uh we are listed here and if we uh go to this page you can see how much token is processed by the TE uh uh TE uh uh you know AP uh uh RM and now we already have uh four models up and running and uh when when you are using those models through open router or through uh directly through uh uh our uh API I uh you can uh be sure that all your inputs output and also you know uh the uh uh uh like a metadata is uh encrypted and never leaked to anyone else. Uh so privacy guaranteed and the integrity also guaranteed. Uh so I think this is a huge step because now we can see that not only web three people are using TE for to enjoy it uh security but also web two people are enjoying that as well. And uh uh yeah and finally uh there is an uh update that's still in development but I'm super willing to show to uh you guys that is uh in the past when running uh the GPU in the TE mode there's actually one caveat that is you can only use one GPU in uh uh the TE virtual machine but recently with the latest update from Nvidia Now we can actually have eight GPU cars connect interconnected using envelink. So that means now we're we can have more than one terabyte of the GPU memory to uh serve even larger uh language models like a deepseec v1 and the deep uh deepsec v3 and deepsec r uh uh r1. So uh and the same time like we can also do uh the large scale fine-tuning on even larger models. Uh so this is uh still in development but we have already successfully tested with uh the open source SDK called private MR SDK and this is actually a work uh uh collaboration between uh FA network and the near AI to bring you an SDK to easily deploy your uh AI application into GPUT and soon you're going to uh be able to access eight GPU cars together. All right. So, yeah, that's uh the a brief update of what we have shipped in the past few weeks. Thank you. I mean, I this is so exciting. I don't think anyone here actually truly understands if you're able to, you know, have a machines like like do the sort of like multi-GPU training with TEES, you can start training entire models inside GP inside of TE's. Is that is that right? I understand that correctly. Eight GPU TE. Okay. So that's like one machine. Mhm. Yeah, that's right. Okay. Is it possible to do mach I know you're you're just figuring out like how to do one machine but can you do machine to machine so you could have like multi machines like you know different clusters training together. Yep that's planned in the future. I think it's not super hard uh far away especially given the black well GPU will be soon available. Cool. And this is like this will actually change the entire state of AI if this works well because then you can do like verifiable like content authentication and all this other stuff and like fully realize the vision that I think was laid out uh in the decentralized machine learning decentralized confidential machine learning paper DCML. Um so this is so cool. Are there any other teams working on this? Like are you guys the only ones? Uh yeah, I I think at least from the web three and open source world uh we're uh we're not aware of any other doing that yet. Okay. Well, Hong, congratulations. This is so exciting. Um would love to share more information online uh like in writing form. So, please be sure to drop us any information that we should be able to uh repost. So, thank you so much for hopping on. Is there anything else that you'd like to highlight before we move on to Manuel? No, I think that's all. Thank you so much. Awesome. H, thank you for joining. That was exciting. Um, Manuel, how you doing, man? It's been a while. Hey. Yeah. Uh, yes, it was exciting. I can't wait to see those GPU is working. I think I want to hook them as soon as possible. Yeah, this is we had we need a a jam on this because you know the work on the TE side is part of the reason why Manuel and I started working together and the fact that you're now able to do this across well it's still in experimentation mode but if you are able to train across many machines or even just like inside one like not single GPU this we can train the first small models and then like scale up as we get more like right now GPUs are the bottom so really exciting updates. Um, real quick, real quick. Who are you? What is Silverstream? What did you ship this week? So, uh, hello everybody. I'm Manuel. I I'm one of the co-ounders of Silver Screen AI. We are a company that has web agents. Um, these are autonomous agents that can can interact with the web like humans do by clicking and typing around the internet. We are trying to focus right now to improve the developer experience. So think of our role who makes your life easier. We want to help developers in this ecosystem to just integrate with agents web agents as well as as possible. We focus on observability. Our key differentiator is that we focus on reliability. So what we want agents where you offload your task and you know it will work. But it also means that there's a lot of work that goes underneath to make it work every time. That's where we specialize. What did we ship last week? Um, so it's more last time I was talking to four months ago. So let's say what we shipped in the last month. One big update is that we we last time we talked we were recruiting some researchers to work on open agents, open models. We can gather we at silver we collect a lot of web data. this data needs to be turned into a model that people can use. We have some initial good results which are going to be published at scientific conferences. So that was very exciting, very open within very much this ecosystem. Uh on the infrastructure side, this is what I guess you care about. We had a much better user experience. This means the time of spinning up an agent. So from the moment you ask us go on this web page to the moment agent actually guest starts to go there went down a lot went from a few minutes to a few seconds. Um it helps a lot with the debugging so you don't need to start a new machine and wait for it to start up every time. On the other side, we we do a lot of QA for for normal people for B2C kind of customers and that has been going on for the last month and a half. We are iterating the the user experience where to a point where they love it. Something else new is per user session uh storage. This is a permanent session that you can have the agent start from. So you don't need to do login every time. You can it's it's as if you it's your own browser where every time you spin up you can have every user remember the past history. This is great for captures because you don't get blocked if recapture or Google or cloudflare they know that you are the same user as before. So that was the biggest part of the of the page. What is going to happen next and the reason I want to talk to you is that we are going to wow support some teams to work on the developer side of things. So we offer APIs. Our agents are integrated through the Neo ecosystem. So you can call our agent and work with it is that it's a little bit complex to understand how to work with it. So we want to find five to eight teams to work with very closely. we will give you a lot of support create a Slack channel with you weekly or by weekly meetings to let you integrate. So now we're looking for daily uh alpha users. This is the second alpha way. So you can get all the first one five people, eight people do write us on Twitter. So this is the Twitter handle if you send us a message to to get in. What I need to know from you is what is the use case and the use cases where these agents are great is mostly something that you would do manually or you would have some very brittle set of playright or browser automation uh workflows and then it breaks you know the website changes something changes you need to do small changes and the code has to be updated that's where we shine. So if you have these use cases where you cannot scale your agent, you need to scale them with human whether it's developer or human we cannot there. So tell us about what you need to do with it the team what are you doing uh with your results today and we will reach back to you within a few days to tell you if we can help or if it's not yet the time. But this is the big thing. I want to have people to talk with us within the ecosystem uh to understand how you can Yeah, man. I'm happy to get you whatever leads you want. Um so just to re reiterate, best for people if they're interested in participating in your beta message you on Twitter. But the people who would benefit most from this beta are those who are looking for some sort of automation within their existing applications. Yes. Say in browser. Yeah. Say you want your agent to pick up API keys from a website. You need to go on the website and register your agent, get the API key somehow and now you can do it with some like ax, some kind of browser automation. But that will change over time, right? So offload it to us. We will do it for you. And we are here to help people to on board in the platform. So incredible. Yeah, makes makes sense to me. Be sure to follow Silverstream on Twitter, Silverstream AI. Well, is there any place where they can follow up with you directly or should they just go to Silverstream Twitter? If you send us a DM or if you t tweet at us, we will answer. out to you my co-ounder is following the Twitter account as much as you can and we will reach out very soon especially if you come from the ecosystem. Yeah. Yeah guys uh don't be dumb. Manuel has finished his PhD and been doing this for quite a while and so if you are interested in well like you know there's real subject matter expertise here. So I should have expanded on that. Yeah. So like our team is very strong on these as we've been working on them for three years now and on the technical side the cos from Google security um we know how to build software so we are here to help developers to build your own agent and to make it more authentic do rely on our expertise because we are here to do that. Ben cool well thanks for coming on as always. Is there anything else you want to say before we hop off? No. And I'm really excited about this on network. I can't wait. Yeah, it's so exciting. Awesome. Thanks for joining. Aaron, you're up next. How you doing? Hi. Hey. Hey. Welcome back. Hey. Uh yeah, great to see you again. Uh this is Aaron, uh the founder of Kronox. So what we're doing is we are utilizing the edge devices from the community and and providing AI services for to AI developers. So uh what I'm bring here today is uh we made a PR to add Kronx as a new uh AI provider to Near SDK. So it's pretty simple. Um if you use near hub you can try to use querex as a provider here. So just change the model here and uh we also provide a free API key to use so everyone can try it out if you are using this um and we will send your request to edge devices running somewhere in this world and getting responses back to you. So this is what we're building. This is fantastic. Um tell everyone a little bit about the benefits of using Krenics for inference like why should they use over fireworks or Grock or something like that. Yeah sure. So um the mo the best benefit here is uh Kronos using edge devices can making the cost really cheap. We currently we provide free API for any near developers. In the future we can promise that the cost will not be more than 10 10% of fireworks. So this is the unique advantage of utilizing edge devices and this is because uh we are utilizing GPU that's much cheaper than A100 or H100 from GPU cloud and we don't pay any uh the pay any fee to the uh data center to the staff of data center. So everything is from community which is really cheap. So this is the unique value of that. So if you have building some agent uh applications you will notice that your openi token will running really fast for any request you will take like tens or even hundreds of API requests to open aai so which will be like several dollars uh several years ago maybe cheaper now but you are serving one users at the same time it's still very expensive using krenx you can make this cost to be really or even free at this moment. So yeah, this is the unique value from KX. That's super exciting. And if somebody's let's say not necessarily needs it for inference but wants to participate in the network, how is the node operating system oper like working these days and how's how's progress been? Yeah, so you just need to download the node app from Kronx network and uh uh install that, double click that and everything will run really fast. Um yeah so just follow our Twitter uh you will get everything uh from our website. Awesome. I'm running my Kronics node now. So you know decentralizing the network. So I'm excited to uh try out the inference and I will do what I can to help get that PR merged. I just hit up the team. Oh, it's already merged. Oh, it's already merged. Fantastic. Yeah. Good stuff. All right. Well, then I'll try it out. Thank you for making it. Yeah. Awesome, man. Well, thanks for joining. Um, we're gonna bring on Jordan. Jordan, are you on? Boom. Welcome back. Hey, good to see you. It's good to see you, too. So, what's going on with public AI? What's what's new? Yeah. Um, I'll I'll get the presentation up because I do have some um exciting news to share. So, let's see here. Um there we go. Share. So very briefly myself um Jordan here been active in the web 3 space coming on six years now and got involved in the space through um my interest in Dows and decentralization and the new ways that web 3 can be used to capture value um and all those exciting opportunities. And over the years having um advised hundreds of projects in their launches um and having seen a lot of different takes on how tokconomics and these tools can be used um I'm I'm really excited about what we're doing at public AI which is making it equitable for everyday people to earn with their experience. Um, and what we're doing, the way we're making it equitable is, uh, we're powering exceptional AI with the data that we get. So, one of the big bottlenecks for AIS. Um, of course, there's the the hardware and um, everything else that uh, we've already been talking about on this call today. Um, but one of the biggest bottlenecks is data. And without fresh new data um there's not a lot of point in training new models. Uh so what we do is for our clients who have already exhausted the existing data in the world um they've already synthesized data uh etc. They've gotten as far as they can via those means. Um when they need niche on demand data they can come to us and we can source it from our community of close to a million now. uh data builders who participate in uploading and voting and everything else that's required to take crowdsource data and get it to uh high quality so that it can be applied to production grade models. Um and what I'm going to be sharing today is updates with our latest client and earning opportunities for everybody that's listening. Um, this platform is literally for everybody. Um, we think that we've gotten it to a balanced place where no matter where you are in the world, um, it's exciting to participate in. So, let's dive right in. Um, the big news is we've onboarded a new client that is doing real-time translation devices. And for that we need people speaking in their native languages um from all over the world in all different dialects um different accents etc. So to get that data we're launching season two of our native language audio recording campaign. And the way for everybody to participate um is that you first when you log on to the website you set your mother tongue. So the the native language that you've spoken uh since birth and then we verify that people actually have that expertise through a quizzing system. And then you can upload up to 30 minutes a day um to maximize your earnings. And then as you vote on each other's uploads, you build your reputation and can earn even more. And this campaign is huge. Uh the rewards are over $500,000. So um we think this is a pretty exciting thing to be getting out to the community and we're looking forward to everybody's participation. So to go into some of the more details about how you actually participate and walk you through that. Um when you hit beta.publicai.io, the first thing you'll see is a popup to set your native language. and we have a pretty good spread of the different languages and the countries where they're spoken. But if for some reason you don't see your country listed here, prioritize getting your language right in this because that's what's going to unlock the quizzes for you. So um prioritize language. Uh get your country uh selected if it is there. Um if it's not, you can feel free to open a Discord ticket which is our official support channel. Um, having had this uh live for over a month, we think we've got them all covered. And one note on this uh native language selector when it pops up is that it can only be set once. So, be sure you get it right. Um, and that's to prevent people from jumping around and saying that they've spoken several languages since birth. Um, and then after that, you can set the UI on the app to whatever you like, whether it's um, English and use your machine's local translation or whatever you like. But be sure that for this um, native language expertise, you set it correctly at the start to get your mother tongue locked in. And then after that step, uh, you can hit up the quiz section. And we're putting new quizzes live every week. Um the important one at the moment is the language quizzes. So um when you do your language setting, you can then pop into quizzes, uh find your language, and after answering a 100 questions, you'll get a badge, and that's what unlocks the uploads and voting process for that particular language. So it goes really fast. Um, you don't need to cheat or stress about like how how smart you are in your language because the questions themselves are fairly obvious. Like if you're a native speaker of a language or for like the the dental campaign, if you work in the dental industry, then the questions are all kind of common sense. And it's really just a tool for us to identify actual expertise. So when you start the quiz, you'll get uh 10 questions at a time. So you'll get 10 questions. there's a timer so that um people aren't cheating and gpting the answers. Um and then after you've gone through that 10 times, you'll have answered 100 questions total. And as long as you're above like 60% uh quality or correct answers on those quizzes, um then you are qualified to participate in the campaigns. And again, that threshold is really just to make sure that um people are actually authentically like from where they say they are or speaking the language they say they are. Um the questions aren't hard. Uh most people have no problem with it. And then after you've gotten that quiz out of the way, then you can go over to the upload section. Um be sure to check between the live and registered. Um, some people have been having trouble finding the actual campaigns. So, check both the live and the registered there. And, um, from there you can just upload audio. So, um, the phrases that come up are pretty short. It's between like probably five and 15 seconds of recording audio each time. And you can upload up to I think we have it set at like 360 uploads a day. Um, about 30 minutes of audio you can upload a day. And then after you've uh hit that target for uploading, you can go and vote on everybody else's uploads. And this is yet another earning opportunity. So um you get to earn points by voting on each other's submissions. And the important thing to note here is that you get rewarded whether or not the sample is good or bad. What you're getting rewarded for is your discretion in deciding if a sample is good or bad. Um the reward comes when you are in consensus with the final vote on any piece of data. So if something is like obviously an AI voice, vote no. If the audio quality is bad and you can't hear or understand somebody, vote no. Um because that's what we would expect other people who are voting to also do. So you want to reach consensus with the other voters. That's when you get your stake back and that's when you get the uh rewards for participating in the voting. And then after that, you can go back to the mission page. Uh kind of track your progress, see how your uploads are doing as far as like their validation process, how far along they are, and if you don't feel like the voting is going fast enough, then go ping your friends and let them know that there's an opportunity uh to earn over at Public AI just by participating in audio uploads and voting. And then if that wasn't all enough, uh we also have bonuses for continued participation. So at like the one hour audio upload mark, 10 hour, 50 hour, 100 audio upload marks, um there are significant rewards um up to $255 at the 100 hour audio mark. So, um, this is our way of letting the community know, uh, it's worth showing up multiple days and to continue, uh, working on these tasks, uh, to meet some long-term goals. Um, not just, uh, check it out once and forget it, but, there's like some significant rewards if you stick around and participate in the campaign. And then there's also an opportunity as you increase your reputation through the voting and uploading to join higher tiers of contributors inside of our system. So after you've participated for a couple days in this campaign, you got your reputation up, you can apply to become a guard in our community, which means that your earnings are multiplied. Um, and there's more opportunities. You get to see campaigns before anybody else. Um, so there's uh like a meaningful way to level up inside of our community. And that's it. Uh especially if you speak any of the languages on the screen right now. Um these are the 12 that are live for the upload and voting task. So um come check it out and we'll be releasing a total of 44 languages uh over the next few months. So keep your eyes peeled across our channels uh Twitter, Telegram, Discord. Um, and if you run into any trouble or have feedback, Discord is the best place to connect and open a ticket. That's our only official support channel. So, yeah, keep your eyes peeled and looking forward to hearing everybody's audio. Thank you, Jordan. This is incredible. An amazing earning opportunity. Do you say $500,000? That's right. Yep. Oh my god. Okay, cool. That beats airdrop farming, that's for sure. Um, so this is super exciting. Um, when does this campaign launch? I'd like to live. Unite. You're live now. And it's going through the 22nd of next month. Um, that's when the season cuts off and then we'll check out how everything's going, tune parameters. Um, we're always iterating on our system and look at the next season. But yeah, now is a great time to get in. Cool. I'm just hitting up the Near Foundation team to start promoting this. So, um, fantastic. Thank you. Thank you. Be sure to check Telegram and coordinate with them as well and getting them on this. Uh, so awesome. Jordan, thank you again. This is super exciting. Uh, I will also be checking this out. Is there anything else that you want to say before we hop off to Zahed? Nope. That's it. Thanks so much for the time. Take care, everybody. Awesome, Jordan. Thank you. All right, Zahed, how's it going? Welcome back. You are muted. Thank you for the invite. Yeah, I'm doing great. Uh so and excited to share some of the stuff that we have been working. So uh how should Yeah, you were teasing it to me a little bit in our in our message saying it's it might be built for me. So I'm excited to hear. Yeah. So uh this is the same process. So hey everyone um I'm Zah, founder of Jutsu. we have uh but today I'm not going to talk about jutsu I'm going to talk about post uh a AI social media manager agent that can write um social media content adopting your your writing style um so the main difference between post and other tools uh is post can adopt your writing style so uh I will show share some demo And then uh you can you can also create provide a brand guideline for uh for post to uh start adopting uh your style. So I think it is always better to show versus tell uh let me share my screen. Share screen. This one. Okay. So this is post uh you can um sign up today like by going to app.post.ai. Uh but the main diff uh difference is uh today the thing that I want to show is the brand kit first of all. So the brand kit is the way to personalize your each and every post. So you can see I have I have several uh writing style. So let's say this person uh write their content in in LinkedIn in in a format like like showing one two three numbers and I know that uh she have lot of follower and every uh post she make have a lot of uh impression. So this is uh Josh is another person who whose writing style is you can see like a little bit different. Uh he uses arrays and stuff. Uh and then imagine like if you are already pro or if you have a social media manager who already have a writing style you can kind of create few writing styles. So the way you create a writing style is you add example. So you can just add few example give it a name then you have a writing style and the way you would use a writing style is uh when you create a post so you come here and then let's say I want to write about uh I want to have a post about uh 402 uh payment protocol u so because I was curious about this protocol all uh but let's let's ask post to generate a uh post using Josh writing style. So what it is doing is it is going to uh search in the uh web and then it will find about the protocol uh and then you can see like it's followed his writing style. Now imagine like a uh another post I want to adopt different writing style. So let's create another post. Same post uh x 402 payment protocol but this time uh let's say I want to adopt different writing style. So if I generate now you will see it will follow it will create the post and it will follow that specific persona or writing style that I choose. So you can now you can use your imagination let's say if you are a big fan of Gary Vee or or someone else you can create a writing style like them. So now you can see like it adopted like this one two three that style not not the previous one. So this is one great way to kind of a uh without putting a lot of effort you can kind of a clone someone's writing style and kind of a start posting like them and hopefully grow your network. The second uh thing important feature that we added is a co-pilot. So there uh I was presenting it in in Amazon and then when I was talking with uh Cameron. So like one of the question is like okay so this tool is great but like a I still love cloud or I still love uh chat GPD can I use it? Uh so now the answer is yes. So we added a copilot where it is fine-tuned for social media related question. Um so let's say uh if I want to create the same thing I can say like um hey I want to use plot son 3.7 uh create a LinkedIn post on x 402 protocol. So we don't need this. Um so now if I ask so what it will do is like it will use plot and it is doing the reasoning so you can see it's it is doing the all the stuff it need to do and then in few second so it's writing like a one type so you need you need a little bit more so can you create a text post. So you can see like all of a sudden you can use uh different kind of models. So not just one model that we selected. Um so so that's kind of the so now then you can um you can even ask like okay how often should I post in LinkedIn. So then it is kind of answering you okay you should you should probably post at least one or twice per day and three uh two to five time uh per month. So this is post uh where the goal is to take away all the uh mental block or writer's block and all the time you need to uh spend on content creation and give you back the time so that you can focus on your product. Um uh we are finally officially launching beginning of June. Uh so and probably we'll share a um kind of a um uh offer for uh near community. Um so if you're interested uh you can already start uh kind of a um trying but like a official launch will be uh June 1. So yeah thank you. Yeah this is great. Um I would love to get entire teams to use post to I kind of think of it like team le marketing. So anytime that there's a big announcement or something like the uh brand guidelines everything is like automatically followed. you know, a lot of like especially engineers on the teams frequently don't like to post on social media and so uh I think this is great and super eager to try to get the entire Near Foundation on boarded uh because everyone should be posting at least once a week is what I'm getting at. So one two things that will come soon. So one is kind of a building that um kind of employeeled or community-led uh kind of social media uh amplification. So where uh as a uh team you can post and you can invite all your uh colleague to kind of a upboard. So that's something will come soon. Uh then the next thing will come is like now it's more like a co-pilot but like a post will evolve to a stage where you will come and hire a intern and then give them some parameter or some goal and then it will just do execute it. So the way a real intern would do. So that's the next step. So first we'll go after a intern level uh kind of a skill and then it will replace hopefully not replace like a augment uh like a social media manager role in the future. Social media would you do SEO eventually? Could you do you know like kind of move up all the way to CMO is the goal. So the main project is Jutsu which is a AI power AI works uh workspace which will become something called software defined workforce. So where you will have different AI agent that complete different job. So one of post is one of these tools right. So then if you remember we had code generation and now code generation is becoming so much better. becoming like a AI engineer. So the future will be you come to this platform you h you need a social media person you hire a agent you need some code generation or code review you have bunch of engineer in Bangladesh and then you still want some kind of a mid-level employee to review their code so you hire a AI code reviewer or something like that and you put together all of this and those idea of like a threeperson unicorn or so that's where it it is going in the future. Uh post maybe SEO, I don't know, but maybe SEO will be a different agent. SEO is also kind of broken in the world of AI anyway. So if yeah, if you can figure out how to influence the AI to give you the proper summary, then like you're on to something big. Um Zah, this is super exciting. You said a launch is coming in early June. Do you have a date yet? Yeah, June 1st. June 1st. Okay. Yeah, eight days. Cool. Well, fantastic. like I will follow up with you and and DM. But uh is there anything that people can do to sign up for this thing in the short term? Is there a form just like get early access? You just you just go app.ai and keep using it. So there is no there is nothing. It's already kind of a soft launch. We have soft launch but the new feature and kind of more official launch will be uh beginning of next month. But for the community near community like you can just go and start using today. Cool. Well, Zahed, thank you so much. Thank you. Y awesome, guys. Uh this is all we have for today. Uh next week there's going to be a bigger update for the Near AI team, which I'm very excited about. Um I work with all of these teams sort of separately. Uh and the whole point here is well, we're building an ecosystem. Uh Near AI is not a single organization that's pushing everything ourselves. There's no way in hell we are going to be able to build decentralized AI independently. Uh we need everyone to do our do their parts and uh you know it's this is the manifestation of it all. Everyone kind of working together. So thanks to everyone for joining. Uh hope to see you guys next week. We do this every Thursday at 11 a.m. Pacific, 2 p.m. Eastern. And see you next time. Thank you. Thanks for watching. Subscribe to Near Week.","transcript_synced_at":"2026-11-03T04:56:54.304Z","segments":[{"start_seconds":17,"text":"near AI. What did you shift this week?","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=17s"},{"start_seconds":24,"text":"Welcome, welcome everybody. Uh my name is Cameron Dennis. I am on the new AI team and this is the 12th episode of What Did You Ship this week. Uh we took","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=24s"},{"start_seconds":32,"text":"a bit of a hiatus last week uh to unite a lot of the amazing speakers that are here today. Um we have a jam-packed","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=32s"},{"start_seconds":39,"text":"agenda and so without further ado, we're just going to hop right on into it. But if you are tuning in for the first time,","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=39s"},{"start_seconds":47,"text":"this is a live stream that we do pretty much every week where we highlight or give builders the the space to highlight what they're up to. who think of it kind","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=47s"},{"start_seconds":54,"text":"of like a open engineering standup of sorts where they really just share what they've been shipping over the last week. And to kick things off, we have","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=54s"},{"start_seconds":62,"text":"Hong Yin from Fallen Network. Hong, how's it going? Welcome. Hey, great.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=62s"},{"start_seconds":68,"text":"Great. Thank you. It's awesome to have you back as always and uh this is not your first time on and so you you know","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=68s"},{"start_seconds":75,"text":"you know the ropes. if you could just let everyone know exactly uh who you are, what follow Network is and then unpack what you ship this week, we can","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=75s"},{"start_seconds":83,"text":"dive right into it. Okay. Okay. Uh there are a bunch of very interesting uh things we have been developing but uh","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=83s"},{"start_seconds":91,"text":"yeah mostly today I'm going to cover all the work related with uh AI and TE. So","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=91s"},{"start_seconds":98,"text":"uh brief intro so uh I'm hung co-founder and uh c CTO of fa network uh and fa","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=98s"},{"start_seconds":106,"text":"network is just a you know uh blockchain ecosystem and now we have a few product everything is around uh te trusted","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=106s"},{"start_seconds":115,"text":"execution environment so just in case you are not familiar with te t is a special uh hardware it allow you to run","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=115s"},{"start_seconds":122,"text":"something uh with the privacy and integrity so imagine you can put something in the TE and uh the TE can uh","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=122s"},{"start_seconds":130,"text":"just run a program as is without any code change. The same time it can give you a proof and with this proof you can","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=130s"},{"start_seconds":136,"text":"be convinced that this program is fully privacy preserving and this program cannot be uh tempered uh tempered with","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=136s"},{"start_seconds":145,"text":"uh even if you are the the uh operator or uh the hardware owner of the hardware. So um yeah let's go through","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=145s"},{"start_seconds":155,"text":"what we are doing this week. So I'm going to just share the screen of this window.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=155s"},{"start_seconds":171,"text":"Okay. All right. Yeah. So yeah this is about this is just our official website.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=171s"},{"start_seconds":178,"text":"So I will just skip it. So uh today I I have uh roughly four things to share. So","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=178s"},{"start_seconds":184,"text":"uh first thing is uh MCP. So uh I think probably most of the people who are uh","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=184s"},{"start_seconds":191,"text":"familiar with the recent development of uh AI have heard about MCP. MCP is a way","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=191s"},{"start_seconds":198,"text":"to extend um your RM your or your AI agent to access some tools. For example,","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=198s"},{"start_seconds":205,"text":"you can connect your AI agent to uh say a blockchain MCP so that it can read and write to the blockchain. And","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=205s"},{"start_seconds":212,"text":"traditionally, if you run MCP, you run MCP server together with your program uh with your w with your AI client together","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=212s"},{"start_seconds":222,"text":"in your uh say your laptop or or your PC. And this is uh it turns out is not very secure because you know you always","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=222s"},{"start_seconds":230,"text":"run some uh binary maybe you haven't checked the content it can read your uh computer and even st uh steal your uh","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=230s"},{"start_seconds":239,"text":"sensitive data. So uh of course an easier way is that we can just uh move the MCP from your computer to a remote","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=239s"},{"start_seconds":248,"text":"server. In this way, the MCP server itself and your AI uh program is isolated. But then another problem","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=248s"},{"start_seconds":256,"text":"becomes if you're going to send say your private keys or any sensitive information to a remote server, how can you transfer the remote server? So this","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=256s"},{"start_seconds":265,"text":"is how uh you know TE can play. Uh so we put the MCP server in the TE and in this","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=265s"},{"start_seconds":272,"text":"way um the T uh TE guarantee that uh the program you're interacting with is exactly the program you want to run. So","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=272s"},{"start_seconds":281,"text":"in this way we can ensure your sensitive data will never get leaked outside of the uh MCP program you want to deploy","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=281s"},{"start_seconds":288,"text":"and there's nobody can uh manipulate how your data is uh uh is accessed within","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=288s"},{"start_seconds":295,"text":"this uh MCP server. uh and uh to make this super super simple, we have already deployed a few MCP servers uh on our T","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=295s"},{"start_seconds":304,"text":"cloud. So as you can see here uh you can see a list of the MCP servers. Those are just a readyto use server. You can just","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=304s"},{"start_seconds":313,"text":"copy and paste the server uh to your uh RM program. So you can start using that.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=313s"},{"start_seconds":319,"text":"And one of that is for example near blockchain MCP. And in this MCP you can read the blockchain uh content like a","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=319s"},{"start_seconds":326,"text":"read from smart contract. At the same time you can even embed your private key uh in your local environment and send it share it securely with the MCP server.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=326s"},{"start_seconds":337,"text":"So it can sign transaction on behalf of you. So yeah this is uh one thing and uh of course uh those MCP are just uh","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=337s"},{"start_seconds":346,"text":"developed and deployed by the official team. What if you want to uh develop your own? So uh we offer off uh offer a","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=346s"},{"start_seconds":355,"text":"very uh easy way to do so. So uh just uh two uh two weeks ago we shipped a new feature called the templates. So in the","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=355s"},{"start_seconds":363,"text":"past I already shared how easy you can deploy the uh virtual machines to uh TE.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=363s"},{"start_seconds":370,"text":"But uh can we make it even easier? So uh the answer to that is uh template. So here we have a list of the template and","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=370s"},{"start_seconds":377,"text":"it's still growing and as you can see we have seven MCP templates uh one A2A template and a few others. So here uh","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=377s"},{"start_seconds":386,"text":"for example if you want to deploy a uh uh like a uh MCP server uh uh D5 lama uh","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=386s"},{"start_seconds":394,"text":"then just click deploy and uh here what we do for you is that just uh uh have","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=394s"},{"start_seconds":401,"text":"all the configuration prefilled and all you need to do is just click deploy and then it will uh jump back to","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=401s"},{"start_seconds":409,"text":"the list of your virtual machines and after a few uh minutes this will be deployed up and running and you can just","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=409s"},{"start_seconds":417,"text":"uh copy and paste the uh URL of this MCP server and use it by yourself. You can even uh deploy your own and get it","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=417s"},{"start_seconds":425,"text":"listed as a template on the cloud. Yeah. So uh yeah, this uh is a","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=425s"},{"start_seconds":432,"text":"very brief u uh update about the cloud feature. On the other side, we also have a bunch of uh interesting updates uh","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=432s"},{"start_seconds":440,"text":"around AI. So uh uh just uh in case you are not familiar with uh uh uh uh the","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=440s"},{"start_seconds":447,"text":"integration between us and the open router uh so open router is an aggregator of AI uh uh APIs. So if you","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=447s"},{"start_seconds":455,"text":"register open router account then you can access uh thousands of AI models uh just using one account and uh actually","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=455s"},{"start_seconds":463,"text":"this is a web two uh product. So it's very popular among AI developers and recently open router just added the TE","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=463s"},{"start_seconds":472,"text":"uh uh uh AI model support and we are proud to be one of the model providers.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=472s"},{"start_seconds":479,"text":"So for example for this llama 3.370 billion uh model uh we are listed here","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=479s"},{"start_seconds":486,"text":"and if we uh go to this page you can see how much token is processed by the TE uh","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=486s"},{"start_seconds":494,"text":"uh TE uh uh you know AP uh uh RM and now we already have uh four models up and","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=494s"},{"start_seconds":501,"text":"running and uh when when you are using those models through open router or through uh directly through uh uh our uh","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=501s"},{"start_seconds":508,"text":"API I uh you can uh be sure that all your inputs output and also you know uh","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=508s"},{"start_seconds":514,"text":"the uh uh uh like a metadata is uh encrypted and never leaked to anyone else. Uh so privacy guaranteed and the","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=514s"},{"start_seconds":524,"text":"integrity also guaranteed. Uh so I think this is a huge step because now we can see that not only web three people are","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=524s"},{"start_seconds":531,"text":"using TE for to enjoy it uh security but also web two people are enjoying that as well.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=531s"},{"start_seconds":540,"text":"And uh uh yeah and finally uh there is an uh update that's still in development","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=540s"},{"start_seconds":548,"text":"but I'm super willing to show to uh you guys that is uh in the past when running uh the GPU in the TE mode there's","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=548s"},{"start_seconds":557,"text":"actually one caveat that is you can only use one GPU in uh uh the TE virtual machine but recently with the latest","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=557s"},{"start_seconds":566,"text":"update from Nvidia Now we can actually have eight GPU cars connect interconnected using envelink.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=566s"},{"start_seconds":574,"text":"So that means now we're we can have more than one terabyte of the GPU memory to","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=574s"},{"start_seconds":581,"text":"uh serve even larger uh language models like a deepseec v1 and the deep uh deepsec v3 and deepsec r uh uh r1. So uh","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=581s"},{"start_seconds":592,"text":"and the same time like we can also do uh the large scale fine-tuning on even larger models. Uh so this is uh still in","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=592s"},{"start_seconds":600,"text":"development but we have already successfully tested with uh the open source SDK called private MR SDK and","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=600s"},{"start_seconds":608,"text":"this is actually a work uh uh collaboration between uh FA network and the near AI to bring you an SDK to","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=608s"},{"start_seconds":616,"text":"easily deploy your uh AI application into GPUT and soon you're going to uh be","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=616s"},{"start_seconds":623,"text":"able to access eight GPU cars together. All right. So, yeah, that's uh","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=623s"},{"start_seconds":631,"text":"the a brief update of what we have shipped in the past few weeks. Thank you. I mean, I this is so exciting. I","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=631s"},{"start_seconds":640,"text":"don't think anyone here actually truly understands if you're able to, you know, have a machines like like do the sort of","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=640s"},{"start_seconds":648,"text":"like multi-GPU training with TEES, you can start training entire models inside GP inside of TE's. Is that is that right? I understand that correctly.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=648s"},{"start_seconds":657,"text":"Eight GPU TE. Okay. So that's like one machine. Mhm. Yeah, that's right. Okay.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=657s"},{"start_seconds":664,"text":"Is it possible to do mach I know you're you're just figuring out like how to do one machine but can you do machine to machine so you could have like multi machines like you know different","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=664s"},{"start_seconds":672,"text":"clusters training together. Yep that's planned in the future. I think it's not super hard uh far away especially given","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=672s"},{"start_seconds":679,"text":"the black well GPU will be soon available.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=679s"},{"start_seconds":683,"text":"Cool. And this is like this will actually change the entire state of AI if this works well because then you can do like verifiable like content","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=683s"},{"start_seconds":692,"text":"authentication and all this other stuff and like fully realize the vision that I think was laid out uh in the decentralized machine learning decentralized confidential machine","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=692s"},{"start_seconds":700,"text":"learning paper DCML. Um so this is so cool. Are there any other teams working on this? Like are you guys the only","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=700s"},{"start_seconds":707,"text":"ones? Uh yeah, I I think at least from the web three and open source world uh we're uh we're not aware of any other","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=707s"},{"start_seconds":715,"text":"doing that yet. Okay. Well, Hong, congratulations. This is so exciting. Um would love to share more information","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=715s"},{"start_seconds":723,"text":"online uh like in writing form. So, please be sure to drop us any information that we should be able to uh repost. So, thank you so much for","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=723s"},{"start_seconds":731,"text":"hopping on. Is there anything else that you'd like to highlight before we move on to Manuel? No, I think that's all.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=731s"},{"start_seconds":737,"text":"Thank you so much. Awesome. H, thank you for joining. That was exciting. Um, Manuel, how you doing, man? It's been a while.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=737s"},{"start_seconds":745,"text":"Hey. Yeah. Uh, yes, it was exciting. I can't wait to see those GPU is working.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=745s"},{"start_seconds":751,"text":"I think I want to hook them as soon as possible.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=751s"},{"start_seconds":754,"text":"Yeah, this is we had we need a a jam on this because you know the work on the TE side is part of the reason why Manuel and I started working together and the","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=754s"},{"start_seconds":762,"text":"fact that you're now able to do this across well it's still in experimentation mode but if you are able to train across many machines or even","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=762s"},{"start_seconds":769,"text":"just like inside one like not single GPU this we can train the first small models and then like scale up as we get more","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=769s"},{"start_seconds":778,"text":"like right now GPUs are the bottom so really exciting updates. Um, real quick, real quick. Who are you?","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=778s"},{"start_seconds":788,"text":"What is Silverstream? What did you ship this week? So, uh, hello everybody. I'm Manuel.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=788s"},{"start_seconds":794,"text":"I I'm one of the co-ounders of Silver Screen AI. We are a company that has web agents. Um, these are autonomous agents","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=794s"},{"start_seconds":803,"text":"that can can interact with the web like humans do by clicking and typing around the internet.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=803s"},{"start_seconds":809,"text":"We are trying to focus right now to improve the developer experience. So think of our role who makes your life","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=809s"},{"start_seconds":816,"text":"easier. We want to help developers in this ecosystem to just integrate with agents web agents as well as as possible. We focus on","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=816s"},{"start_seconds":825,"text":"observability. Our key differentiator is that we focus on reliability. So what we want agents where you offload your task","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=825s"},{"start_seconds":833,"text":"and you know it will work. But it also means that there's a lot of work that goes underneath to make it work every time. That's where we specialize. What","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=833s"},{"start_seconds":841,"text":"did we ship last week? Um, so it's more last time I was talking to four months ago. So let's say what we shipped in the","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=841s"},{"start_seconds":850,"text":"last month. One big update is that we we last time we talked we were recruiting some researchers to work on open agents,","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=850s"},{"start_seconds":859,"text":"open models. We can gather we at silver we collect a lot of web data. this data needs to be turned into a model that","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=859s"},{"start_seconds":866,"text":"people can use. We have some initial good results which are going to be published at scientific conferences. So","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=866s"},{"start_seconds":873,"text":"that was very exciting, very open within very much this ecosystem. Uh on the infrastructure","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=873s"},{"start_seconds":881,"text":"side, this is what I guess you care about. We had a much better user experience. This means the time of spinning up an agent. So from the moment","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=881s"},{"start_seconds":889,"text":"you ask us go on this web page to the moment agent actually guest starts to go there went down a lot went from a few minutes to a few seconds.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=889s"},{"start_seconds":901,"text":"Um it helps a lot with the debugging so you don't need to start a new machine and wait for it to start up every time.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=901s"},{"start_seconds":908,"text":"On the other side, we we do a lot of QA for for normal people for B2C kind of customers and that has been going on for","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=908s"},{"start_seconds":917,"text":"the last month and a half. We are iterating the the user experience where to a point where they love","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=917s"},{"start_seconds":925,"text":"it. Something else new is per user session uh storage. This is a permanent session that you can have the agent","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=925s"},{"start_seconds":934,"text":"start from. So you don't need to do login every time. You can it's it's as if you it's your own browser where every time you spin up you can have every user","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=934s"},{"start_seconds":943,"text":"remember the past history. This is great for captures because you don't get blocked if recapture or Google or","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=943s"},{"start_seconds":952,"text":"cloudflare they know that you are the same user as before. So that was the biggest part of the of the page. What is","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=952s"},{"start_seconds":961,"text":"going to happen next and the reason I want to talk to you is that we are going to wow support some teams to work on the developer side of things. So we offer","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=961s"},{"start_seconds":970,"text":"APIs. Our agents are integrated through the Neo ecosystem. So you can call our agent and work with it is that it's a","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=970s"},{"start_seconds":978,"text":"little bit complex to understand how to work with it. So we want to find five to eight teams to work with very closely.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=978s"},{"start_seconds":986,"text":"we will give you a lot of support create a Slack channel with you weekly or by weekly meetings to let you","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=986s"},{"start_seconds":993,"text":"integrate. So now we're looking for daily uh alpha users. This is the second","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=993s"},{"start_seconds":999,"text":"alpha way. So you can get all the first one five people, eight people do write","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=999s"},{"start_seconds":1006,"text":"us on Twitter. So this is the Twitter handle if you send us a message to to get in. What I need to know from you is","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1006s"},{"start_seconds":1015,"text":"what is the use case and the use cases where these agents are great is mostly something that you would do manually or","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1015s"},{"start_seconds":1023,"text":"you would have some very brittle set of playright or browser automation uh workflows and then it breaks you know","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1023s"},{"start_seconds":1032,"text":"the website changes something changes you need to do small changes and the code has to be updated that's where we shine. So if you have these use cases","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1032s"},{"start_seconds":1040,"text":"where you cannot scale your agent, you need to scale them with human whether it's developer or human we cannot there.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1040s"},{"start_seconds":1048,"text":"So tell us about what you need to do with it the team what are you doing uh with your results today and we will reach back to you within a few days to","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1048s"},{"start_seconds":1057,"text":"tell you if we can help or if it's not yet the time. But this is the big thing.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1057s"},{"start_seconds":1062,"text":"I want to have people to talk with us within the ecosystem uh to understand how you","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1062s"},{"start_seconds":1069,"text":"can Yeah, man. I'm happy to get you whatever leads you want. Um so just to re reiterate, best for people if they're","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1069s"},{"start_seconds":1078,"text":"interested in participating in your beta message you on Twitter. But the people who would benefit most from this beta","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1078s"},{"start_seconds":1086,"text":"are those who are looking for some sort of automation within their existing applications. Yes. Say in browser. Yeah.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1086s"},{"start_seconds":1093,"text":"Say you want your agent to pick up API keys from a website. You need to go on the website and register your agent, get","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1093s"},{"start_seconds":1100,"text":"the API key somehow and now you can do it with some like ax, some kind of browser automation. But that will change over time, right? So offload it to us.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1100s"},{"start_seconds":1110,"text":"We will do it for you. And we are here to help people to on board in the platform. So incredible. Yeah, makes","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1110s"},{"start_seconds":1118,"text":"makes sense to me. Be sure to follow Silverstream on Twitter, Silverstream AI. Well, is there any place where they can follow up with you directly or","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1118s"},{"start_seconds":1126,"text":"should they just go to Silverstream Twitter? If you send us a DM or if you t tweet at us, we will answer. out to you","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1126s"},{"start_seconds":1134,"text":"my co-ounder is following the Twitter account as much as you can and we will reach out very soon especially if you","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1134s"},{"start_seconds":1141,"text":"come from the ecosystem. Yeah. Yeah guys uh don't be dumb. Manuel has finished his PhD and been doing this for quite a","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1141s"},{"start_seconds":1149,"text":"while and so if you are interested in well like you know there's real subject matter expertise here. So I should have expanded on that. Yeah. So like our team","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1149s"},{"start_seconds":1158,"text":"is very strong on these as we've been working on them for three years now and on the technical side the cos from","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1158s"},{"start_seconds":1165,"text":"Google security um we know how to build software so we are here to help developers to build your own agent and","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1165s"},{"start_seconds":1174,"text":"to make it more authentic do rely on our expertise because we are here to do that. Ben cool well thanks for coming on","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1174s"},{"start_seconds":1182,"text":"as always. Is there anything else you want to say before we hop off? No. And I'm really excited about this on network. I can't wait. Yeah, it's so exciting. Awesome. Thanks for joining.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1182s"},{"start_seconds":1195,"text":"Aaron, you're up next. How you doing? Hi.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1195s"},{"start_seconds":1201,"text":"Hey. Hey. Welcome back. Hey. Uh yeah, great to see you again. Uh this is Aaron, uh the founder of Kronox. So what","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1201s"},{"start_seconds":1209,"text":"we're doing is we are utilizing the edge devices from the community and and","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1209s"},{"start_seconds":1215,"text":"providing AI services for to AI developers. So uh what I'm bring here","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1215s"},{"start_seconds":1223,"text":"today is uh we made a PR to add Kronx as a new uh AI","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1223s"},{"start_seconds":1232,"text":"provider to Near SDK. So it's pretty simple.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1232s"},{"start_seconds":1239,"text":"Um if you use near hub you can try to use querex as a provider here. So just","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1239s"},{"start_seconds":1247,"text":"change the model here and uh we also provide a free API key to use so everyone can try it out if you are using","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1247s"},{"start_seconds":1256,"text":"this um and we will send your request to edge devices running somewhere in this world and getting responses back to you.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1256s"},{"start_seconds":1266,"text":"So this is what we're building.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1266s"},{"start_seconds":1271,"text":"This is fantastic. Um tell everyone a little bit about the benefits of using Krenics for inference like why should","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1271s"},{"start_seconds":1279,"text":"they use over fireworks or Grock or something like that. Yeah sure. So","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1279s"},{"start_seconds":1285,"text":"um the mo the best benefit here is uh Kronos using edge devices can making the","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1285s"},{"start_seconds":1294,"text":"cost really cheap. We currently we provide free API for any near developers. In the future we can promise","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1294s"},{"start_seconds":1302,"text":"that the cost will not be more than 10 10% of fireworks. So this is the unique advantage of utilizing edge devices and","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1302s"},{"start_seconds":1310,"text":"this is because uh we are utilizing GPU that's much cheaper than A100 or H100","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1310s"},{"start_seconds":1318,"text":"from GPU cloud and we don't pay any uh the pay any fee to the uh data center","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1318s"},{"start_seconds":1328,"text":"to the staff of data center. So everything is from community which is really cheap. So this is the unique value of that. So if you have building","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1328s"},{"start_seconds":1337,"text":"some agent uh applications you will notice that your openi token will","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1337s"},{"start_seconds":1344,"text":"running really fast for any request you will take like tens or even hundreds of API requests to open aai so which will","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1344s"},{"start_seconds":1352,"text":"be like several dollars uh several years ago maybe cheaper now but you are serving one users at the same time it's","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1352s"},{"start_seconds":1360,"text":"still very expensive using krenx you can make this cost to be really or even free at this moment. So yeah,","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1360s"},{"start_seconds":1368,"text":"this is the unique value from KX. That's super exciting. And if somebody's let's say not necessarily needs it for inference but wants to participate in","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1368s"},{"start_seconds":1376,"text":"the network, how is the node operating system oper like working these days and how's how's progress been? Yeah, so you","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1376s"},{"start_seconds":1384,"text":"just need to download the node app from Kronx network and uh uh install that, double click that and everything will","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1384s"},{"start_seconds":1391,"text":"run really fast. Um yeah so just follow our Twitter uh you will get everything uh from our website.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1391s"},{"start_seconds":1400,"text":"Awesome. I'm running my Kronics node now. So you know decentralizing the network. So I'm excited to uh try out the inference and I will do what I can","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1400s"},{"start_seconds":1409,"text":"to help get that PR merged. I just hit up the team. Oh, it's already merged. Oh, it's already merged. Fantastic.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1409s"},{"start_seconds":1415,"text":"Yeah. Good stuff. All right. Well, then I'll try it out. Thank you for making it. Yeah. Awesome, man. Well, thanks for joining. Um, we're gonna bring on Jordan. Jordan, are you on? Boom.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1415s"},{"start_seconds":1427,"text":"Welcome back. Hey, good to see you. It's good to see you, too. So, what's going on with public AI? What's what's new?","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1427s"},{"start_seconds":1436,"text":"Yeah. Um, I'll I'll get the presentation up because I do have some um exciting news to share. So, let's see here.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1436s"},{"start_seconds":1449,"text":"Um there we go. Share. So very briefly myself um Jordan here been active in the","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1449s"},{"start_seconds":1456,"text":"web 3 space coming on six years now and got involved in the space through um my","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1456s"},{"start_seconds":1464,"text":"interest in Dows and decentralization and the new ways that web 3 can be used to capture value um and all those","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1464s"},{"start_seconds":1471,"text":"exciting opportunities. And over the years having um advised hundreds of projects in their launches um and having","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1471s"},{"start_seconds":1479,"text":"seen a lot of different takes on how tokconomics and these tools can be used um I'm I'm really excited about what","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1479s"},{"start_seconds":1487,"text":"we're doing at public AI which is making it equitable for everyday people to earn","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1487s"},{"start_seconds":1493,"text":"with their experience. Um, and what we're doing, the way we're making it equitable is, uh, we're powering","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1493s"},{"start_seconds":1502,"text":"exceptional AI with the data that we get. So, one of the big bottlenecks for AIS. Um, of course, there's the the","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1502s"},{"start_seconds":1511,"text":"hardware and um, everything else that uh, we've already been talking about on this call today. Um, but one of the","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1511s"},{"start_seconds":1518,"text":"biggest bottlenecks is data. And without fresh new data um there's not a lot of","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1518s"},{"start_seconds":1525,"text":"point in training new models. Uh so what we do is for our clients who have","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1525s"},{"start_seconds":1531,"text":"already exhausted the existing data in the world um they've already synthesized data uh etc. They've gotten as far as","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1531s"},{"start_seconds":1539,"text":"they can via those means. Um when they need niche on demand data they can come to us and we can source it from our","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1539s"},{"start_seconds":1547,"text":"community of close to a million now. uh data builders who participate in uploading and voting and everything else","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1547s"},{"start_seconds":1557,"text":"that's required to take crowdsource data and get it to uh high quality so that it","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1557s"},{"start_seconds":1564,"text":"can be applied to production grade models. Um and what I'm going to be sharing today is updates with our latest","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1564s"},{"start_seconds":1574,"text":"client and earning opportunities for everybody that's listening. Um, this platform is literally for everybody. Um,","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1574s"},{"start_seconds":1582,"text":"we think that we've gotten it to a balanced place where no matter where you are in the world, um, it's exciting to participate in. So, let's dive right in.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1582s"},{"start_seconds":1591,"text":"Um, the big news is we've onboarded a new client that is doing real-time translation devices. And for that we","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1591s"},{"start_seconds":1601,"text":"need people speaking in their native languages um from all over the world in all different dialects um different","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1601s"},{"start_seconds":1608,"text":"accents etc. So to get that data we're launching season two of our native","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1608s"},{"start_seconds":1615,"text":"language audio recording campaign. And the way for everybody to participate um is that you first when you log on to the","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1615s"},{"start_seconds":1623,"text":"website you set your mother tongue. So the the native language that you've spoken uh since birth and then we verify","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1623s"},{"start_seconds":1632,"text":"that people actually have that expertise through a quizzing system. And then you can upload up to 30 minutes a day um to","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1632s"},{"start_seconds":1640,"text":"maximize your earnings. And then as you vote on each other's uploads, you build your reputation and can earn even more.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1640s"},{"start_seconds":1650,"text":"And this campaign is huge. Uh the rewards are over $500,000. So um we","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1650s"},{"start_seconds":1657,"text":"think this is a pretty exciting thing to be getting out to the community and we're looking forward to everybody's participation. So to go into some of the","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1657s"},{"start_seconds":1666,"text":"more details about how you actually participate and walk you through that.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1666s"},{"start_seconds":1671,"text":"Um when you hit beta.publicai.io, the first thing you'll see is a popup to set your native","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1671s"},{"start_seconds":1678,"text":"language. and we have a pretty good spread of the different languages and the countries where they're spoken. But","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1678s"},{"start_seconds":1685,"text":"if for some reason you don't see your country listed here, prioritize getting your language right in this because","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1685s"},{"start_seconds":1693,"text":"that's what's going to unlock the quizzes for you. So um prioritize language. Uh get your country uh","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1693s"},{"start_seconds":1700,"text":"selected if it is there. Um if it's not, you can feel free to open a Discord ticket which is our official support channel. Um, having had this uh live for","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1700s"},{"start_seconds":1709,"text":"over a month, we think we've got them all covered. And one note on this uh native language selector when it pops up","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1709s"},{"start_seconds":1716,"text":"is that it can only be set once. So, be sure you get it right. Um, and that's to prevent people from jumping around and saying that they've spoken several","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1716s"},{"start_seconds":1723,"text":"languages since birth. Um, and then after that, you can set the UI on the app to whatever you like, whether it's","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1723s"},{"start_seconds":1732,"text":"um, English and use your machine's local translation or whatever you like. But be sure that for this um, native language","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1732s"},{"start_seconds":1739,"text":"expertise, you set it correctly at the start to get your mother tongue locked in. And then after that step, uh, you","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1739s"},{"start_seconds":1746,"text":"can hit up the quiz section. And we're putting new quizzes live every week. Um the important one at the moment is the","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1746s"},{"start_seconds":1754,"text":"language quizzes. So um when you do your language setting, you can then pop into quizzes, uh find your language, and","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1754s"},{"start_seconds":1763,"text":"after answering a 100 questions, you'll get a badge, and that's what unlocks the uploads and voting process for that","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1763s"},{"start_seconds":1771,"text":"particular language. So it goes really fast. Um, you don't need to cheat or stress about like how how smart you are","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1771s"},{"start_seconds":1780,"text":"in your language because the questions themselves are fairly obvious. Like if you're a native speaker of a language or for like the the dental campaign, if you","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1780s"},{"start_seconds":1789,"text":"work in the dental industry, then the questions are all kind of common sense.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1789s"},{"start_seconds":1793,"text":"And it's really just a tool for us to identify actual expertise. So when you start the quiz, you'll get uh 10 questions at a time. So you'll get 10","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1793s"},{"start_seconds":1802,"text":"questions. there's a timer so that um people aren't cheating and gpting the answers. Um and then after you've gone","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1802s"},{"start_seconds":1810,"text":"through that 10 times, you'll have answered 100 questions total. And as long as you're above like 60% uh quality or correct answers on those","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1810s"},{"start_seconds":1818,"text":"quizzes, um then you are qualified to participate in the campaigns. And again, that threshold is really just to make sure","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1818s"},{"start_seconds":1826,"text":"that um people are actually authentically like from where they say they are or speaking the language they","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1826s"},{"start_seconds":1834,"text":"say they are. Um the questions aren't hard. Uh most people have no problem with it. And then after you've gotten that","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1834s"},{"start_seconds":1842,"text":"quiz out of the way, then you can go over to the upload section. Um be sure to check between the live and registered. Um, some people have been","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1842s"},{"start_seconds":1850,"text":"having trouble finding the actual campaigns. So, check both the live and the registered there. And, um, from","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1850s"},{"start_seconds":1858,"text":"there you can just upload audio. So, um, the phrases that come up are pretty short. It's between like probably five and 15 seconds of recording audio each","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1858s"},{"start_seconds":1866,"text":"time. And you can upload up to I think we have it set at like 360 uploads a","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1866s"},{"start_seconds":1873,"text":"day. Um, about 30 minutes of audio you can upload a day. And then after you've uh hit that target for uploading, you","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1873s"},{"start_seconds":1883,"text":"can go and vote on everybody else's uploads. And this is yet another earning opportunity. So um you get to earn","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1883s"},{"start_seconds":1891,"text":"points by voting on each other's submissions. And the important thing to note here is that you get rewarded","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1891s"},{"start_seconds":1899,"text":"whether or not the sample is good or bad. What you're getting rewarded for is your discretion in deciding if a sample","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1899s"},{"start_seconds":1906,"text":"is good or bad. Um the reward comes when you are in consensus with the final vote","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1906s"},{"start_seconds":1914,"text":"on any piece of data. So if something is like obviously an AI voice, vote no. If the audio quality is bad and you can't","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1914s"},{"start_seconds":1922,"text":"hear or understand somebody, vote no. Um because that's what we would expect other people who are voting to also do.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1922s"},{"start_seconds":1929,"text":"So you want to reach consensus with the other voters. That's when you get your stake back and that's when you get the","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1929s"},{"start_seconds":1935,"text":"uh rewards for participating in the voting. And then after that, you can go","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1935s"},{"start_seconds":1944,"text":"back to the mission page. Uh kind of track your progress, see how your uploads are doing as far as like their","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1944s"},{"start_seconds":1952,"text":"validation process, how far along they are, and if you don't feel like the voting is going fast enough, then go ping your friends and let them know that","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1952s"},{"start_seconds":1960,"text":"there's an opportunity uh to earn over at Public AI just by participating in audio uploads and voting.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1960s"},{"start_seconds":1967,"text":"And then if that wasn't all enough, uh we also have bonuses for continued participation. So at like the one hour","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1967s"},{"start_seconds":1976,"text":"audio upload mark, 10 hour, 50 hour, 100 audio upload marks, um there are","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1976s"},{"start_seconds":1983,"text":"significant rewards um up to $255 at the 100 hour audio mark. So, um, this is our","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1983s"},{"start_seconds":1990,"text":"way of letting the community know, uh, it's worth showing up multiple days and to continue, uh, working on these tasks,","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1990s"},{"start_seconds":1999,"text":"uh, to meet some long-term goals. Um, not just, uh, check it out once and forget it, but, there's like some significant rewards if you stick around","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=1999s"},{"start_seconds":2007,"text":"and participate in the campaign. And then there's also an opportunity as you increase your reputation through the voting and uploading to join higher","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2007s"},{"start_seconds":2016,"text":"tiers of contributors inside of our system. So after you've participated for a couple days in this campaign, you got your reputation up, you can apply to","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2016s"},{"start_seconds":2025,"text":"become a guard in our community, which means that your earnings are multiplied.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2025s"},{"start_seconds":2030,"text":"Um, and there's more opportunities. You get to see campaigns before anybody else. Um, so there's uh like a meaningful way to level up inside of our community.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2030s"},{"start_seconds":2041,"text":"And that's it. Uh especially if you speak any of the languages on the screen right now. Um these are the 12 that are","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2041s"},{"start_seconds":2047,"text":"live for the upload and voting task. So um come check it out and we'll be","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2047s"},{"start_seconds":2054,"text":"releasing a total of 44 languages uh over the next few months. So keep your eyes peeled across our channels uh","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2054s"},{"start_seconds":2062,"text":"Twitter, Telegram, Discord. Um, and if you run into any trouble or have feedback, Discord is the best place to","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2062s"},{"start_seconds":2069,"text":"connect and open a ticket. That's our only official support channel. So, yeah, keep your eyes peeled and looking forward to hearing everybody's audio.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2069s"},{"start_seconds":2079,"text":"Thank you, Jordan. This is incredible.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2079s"},{"start_seconds":2081,"text":"An amazing earning opportunity. Do you say $500,000? That's right. Yep. Oh my god. Okay, cool. That beats airdrop","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2081s"},{"start_seconds":2088,"text":"farming, that's for sure. Um, so this is super exciting. Um, when does this campaign launch? I'd like to live.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2088s"},{"start_seconds":2096,"text":"Unite. You're live now. And it's going through the 22nd of next month. Um, that's when the season cuts off and then","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2096s"},{"start_seconds":2105,"text":"we'll check out how everything's going, tune parameters. Um, we're always iterating on our system and look at the next season. But yeah, now is a great","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2105s"},{"start_seconds":2113,"text":"time to get in. Cool. I'm just hitting up the Near Foundation team to start promoting this. So, um, fantastic. Thank","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2113s"},{"start_seconds":2122,"text":"you. Thank you. Be sure to check Telegram and coordinate with them as well and getting them on this. Uh, so awesome. Jordan, thank you again. This","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2122s"},{"start_seconds":2129,"text":"is super exciting. Uh, I will also be checking this out. Is there anything else that you want to say before we hop off to Zahed?","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2129s"},{"start_seconds":2136,"text":"Nope. That's it. Thanks so much for the time. Take care, everybody. Awesome, Jordan. Thank you. All right, Zahed, how's it going? Welcome back. You are muted.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2136s"},{"start_seconds":2148,"text":"Thank you for the invite. Yeah, I'm doing great. Uh so and excited to share some of the stuff that we have been","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2148s"},{"start_seconds":2156,"text":"working. So uh how should Yeah, you were teasing it to me a little bit in our in our message","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2156s"},{"start_seconds":2164,"text":"saying it's it might be built for me. So I'm excited to hear. Yeah. So uh this is","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2164s"},{"start_seconds":2172,"text":"the same process. So hey everyone um I'm Zah, founder of Jutsu. we have uh but","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2172s"},{"start_seconds":2179,"text":"today I'm not going to talk about jutsu I'm going to talk about post uh a AI","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2179s"},{"start_seconds":2186,"text":"social media manager agent that can write um social media content adopting","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2186s"},{"start_seconds":2193,"text":"your your writing style um so the main difference between post and other tools","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2193s"},{"start_seconds":2200,"text":"uh is post can adopt your writing style so uh I will show share some demo","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2200s"},{"start_seconds":2207,"text":"And then uh you can you can also create provide a brand guideline for uh for post to uh","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2207s"},{"start_seconds":2215,"text":"start adopting uh your style. So I think it is always better to show versus tell","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2215s"},{"start_seconds":2223,"text":"uh let me share my screen. Share screen. This","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2223s"},{"start_seconds":2231,"text":"one. Okay. So this is post uh you can um sign up today like by going to","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2231s"},{"start_seconds":2240,"text":"app.post.ai. Uh but the main diff uh difference is uh today the thing that I","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2240s"},{"start_seconds":2246,"text":"want to show is the brand kit first of all. So the brand kit is the way to","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2246s"},{"start_seconds":2255,"text":"personalize your each and every post. So you can see I have I have several uh writing style. So let's say this person","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2255s"},{"start_seconds":2265,"text":"uh write their content in in LinkedIn in in a format like like showing one two","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2265s"},{"start_seconds":2272,"text":"three numbers and I know that uh she have lot of follower and every uh post she make have a lot of uh impression. So","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2272s"},{"start_seconds":2282,"text":"this is uh Josh is another person who whose writing style is you can see like a little bit different. Uh he uses","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2282s"},{"start_seconds":2289,"text":"arrays and stuff. Uh and then imagine like if you are already pro or if you have a social media manager who already","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2289s"},{"start_seconds":2298,"text":"have a writing style you can kind of create few writing styles. So the way you create a writing style is you add","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2298s"},{"start_seconds":2305,"text":"example. So you can just add few example give it a name then you have a writing style and the way you would use a","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2305s"},{"start_seconds":2314,"text":"writing style is uh when you create a post so you come here and then let's say","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2314s"},{"start_seconds":2321,"text":"I want to write about uh I want to have a post about uh 402 uh payment protocol","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2321s"},{"start_seconds":2334,"text":"u so because I was curious about this protocol all uh but let's let's ask post to generate a uh post using","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2334s"},{"start_seconds":2344,"text":"Josh writing style. So what it is doing is it is","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2344s"},{"start_seconds":2351,"text":"going to uh search in the uh web and then it will find about the protocol uh and then you can see like it's followed","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2351s"},{"start_seconds":2360,"text":"his writing style. Now imagine like a uh another post I want to adopt different","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2360s"},{"start_seconds":2368,"text":"writing style. So let's create another post. Same post uh x","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2368s"},{"start_seconds":2376,"text":"402 payment protocol but this time uh let's say I want to adopt different writing style.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2376s"},{"start_seconds":2385,"text":"So if I generate now you will see it will follow it will create the post and it will follow that specific persona or","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2385s"},{"start_seconds":2394,"text":"writing style that I choose. So you can now you can use your imagination let's say if you are a big","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2394s"},{"start_seconds":2401,"text":"fan of Gary Vee or or someone else you can create a writing style like them. So now you can see like it adopted like","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2401s"},{"start_seconds":2408,"text":"this one two three that style not not the previous one. So this is one great","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2408s"},{"start_seconds":2415,"text":"way to kind of a uh without putting a lot of effort you can kind of a clone someone's writing style and kind of a","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2415s"},{"start_seconds":2424,"text":"start posting like them and hopefully grow your network. The second","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2424s"},{"start_seconds":2430,"text":"uh thing important feature that we added is a co-pilot. So there uh I was","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2430s"},{"start_seconds":2438,"text":"presenting it in in Amazon and then when I was talking with uh Cameron. So like one of the question is like okay so this","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2438s"},{"start_seconds":2447,"text":"tool is great but like a I still love cloud or I still love uh chat GPD can I","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2447s"},{"start_seconds":2453,"text":"use it? Uh so now the answer is yes. So we added a copilot where it is fine-tuned for social media related","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2453s"},{"start_seconds":2462,"text":"question. Um so let's say uh if I want to create the same thing I can say like um hey I want to use plot son 3.7","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2462s"},{"start_seconds":2472,"text":"uh create a LinkedIn post","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2472s"},{"start_seconds":2480,"text":"on x 402 protocol. So we don't need this. Um","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2480s"},{"start_seconds":2490,"text":"so now if I ask so what it will do is like it will use plot and it is doing the reasoning so you can see it's it is","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2490s"},{"start_seconds":2499,"text":"doing the all the stuff it need to do and then in few second so it's","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2499s"},{"start_seconds":2509,"text":"writing like a one type so you need you need a little bit more so can you create","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2509s"},{"start_seconds":2516,"text":"a text post. So you can see like all of a","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2516s"},{"start_seconds":2526,"text":"sudden you can use uh different kind of models. So not just one model that we","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2526s"},{"start_seconds":2532,"text":"selected. Um so so that's kind of the so now then you can um you can even ask","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2532s"},{"start_seconds":2540,"text":"like okay how often should I","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2540s"},{"start_seconds":2549,"text":"post in LinkedIn. So then it is kind of","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2549s"},{"start_seconds":2559,"text":"answering you okay you should you should probably post at least one or twice per day and three uh two to five time uh per","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2559s"},{"start_seconds":2567,"text":"month. So this is post uh where the goal is to take away all the uh mental block","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2567s"},{"start_seconds":2575,"text":"or writer's block and all the time you need to uh spend on content creation and","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2575s"},{"start_seconds":2583,"text":"give you back the time so that you can focus on your product. Um uh we are","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2583s"},{"start_seconds":2590,"text":"finally officially launching beginning of June. Uh so and probably we'll share","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2590s"},{"start_seconds":2597,"text":"a um kind of a um uh offer for uh near community. Um so if","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2597s"},{"start_seconds":2607,"text":"you're interested uh you can already start uh kind of a um trying but like a","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2607s"},{"start_seconds":2613,"text":"official launch will be uh June 1. So yeah thank you. Yeah this is","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2613s"},{"start_seconds":2620,"text":"great. Um I would love to get entire teams to use post to I kind of think of","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2620s"},{"start_seconds":2627,"text":"it like team le marketing. So anytime that there's a big announcement or something like the uh brand guidelines everything is like automatically","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2627s"},{"start_seconds":2636,"text":"followed. you know, a lot of like especially engineers on the teams frequently don't like to post on social media and so uh I think this is great","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2636s"},{"start_seconds":2644,"text":"and super eager to try to get the entire Near Foundation on boarded uh because everyone should be posting at least once","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2644s"},{"start_seconds":2652,"text":"a week is what I'm getting at. So one two things that will come soon. So one","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2652s"},{"start_seconds":2659,"text":"is kind of a building that um kind of employeeled or community-led uh kind of","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2659s"},{"start_seconds":2666,"text":"social media uh amplification. So where uh as a uh team you can post and you can","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2666s"},{"start_seconds":2673,"text":"invite all your uh colleague to kind of a upboard. So that's something will come soon. Uh then the next thing will come","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2673s"},{"start_seconds":2681,"text":"is like now it's more like a co-pilot but like a post will evolve to a stage where you will come and hire a intern","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2681s"},{"start_seconds":2690,"text":"and then give them some parameter or some goal and then it will just do execute it. So the way a real intern","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2690s"},{"start_seconds":2698,"text":"would do. So that's the next step. So first we'll go after a intern level uh kind of a skill and then it will replace","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2698s"},{"start_seconds":2708,"text":"hopefully not replace like a augment uh like a social media manager role in the future.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2708s"},{"start_seconds":2715,"text":"Social media would you do SEO eventually? Could you do you know like kind of move up all the way to CMO is","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2715s"},{"start_seconds":2721,"text":"the goal. So the main project is Jutsu which is a AI power AI works uh","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2721s"},{"start_seconds":2731,"text":"workspace which will become something called software defined workforce. So where you will have different AI agent","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2731s"},{"start_seconds":2741,"text":"that complete different job. So one of post is one of these tools right. So then if you remember we had code","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2741s"},{"start_seconds":2748,"text":"generation and now code generation is becoming so much better. becoming like a AI engineer. So the future will be you come to this platform you h you need a","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2748s"},{"start_seconds":2757,"text":"social media person you hire a agent you need some code generation or code review you have bunch of engineer in Bangladesh","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2757s"},{"start_seconds":2765,"text":"and then you still want some kind of a mid-level employee to review their code so you hire a AI code reviewer or","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2765s"},{"start_seconds":2774,"text":"something like that and you put together all of this and those idea of like a threeperson unicorn or so that's where","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2774s"},{"start_seconds":2782,"text":"it it is going in the future. Uh post maybe SEO, I don't know, but maybe SEO will be a different agent.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2782s"},{"start_seconds":2790,"text":"SEO is also kind of broken in the world of AI anyway. So if yeah, if you can figure out how to influence the AI to give you the proper summary, then like","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2790s"},{"start_seconds":2798,"text":"you're on to something big. Um Zah, this is super exciting. You said a launch is coming in early June. Do you have a date yet? Yeah, June 1st. June 1st. Okay.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2798s"},{"start_seconds":2809,"text":"Yeah, eight days. Cool. Well, fantastic.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2809s"},{"start_seconds":2811,"text":"like I will follow up with you and and DM. But uh is there anything that people can do to sign up for this thing in the short term? Is there a form just like","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2811s"},{"start_seconds":2819,"text":"get early access? You just you just go app.ai and keep using it. So there is","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2819s"},{"start_seconds":2827,"text":"no there is nothing. It's already kind of a soft launch. We have soft launch but the new feature and kind of more official launch will be uh beginning of","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2827s"},{"start_seconds":2835,"text":"next month. But for the community near community like you can just go and start using today. Cool. Well, Zahed, thank you so much. Thank you. Y awesome, guys.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2835s"},{"start_seconds":2847,"text":"Uh this is all we have for today. Uh next week there's going to be a bigger update for the Near AI team, which I'm very excited about. Um I work with all","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2847s"},{"start_seconds":2855,"text":"of these teams sort of separately. Uh and the whole point here is well, we're building an ecosystem. Uh Near AI is not a single organization that's pushing","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2855s"},{"start_seconds":2864,"text":"everything ourselves. There's no way in hell we are going to be able to build decentralized AI independently. Uh we need everyone to do our do their parts","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2864s"},{"start_seconds":2873,"text":"and uh you know it's this is the manifestation of it all. Everyone kind of working together. So thanks to everyone for joining. Uh hope to see you","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2873s"},{"start_seconds":2881,"text":"guys next week. We do this every Thursday at 11 a.m. Pacific, 2 p.m.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2881s"},{"start_seconds":2886,"text":"Eastern. And see you next time. Thank you.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2886s"},{"start_seconds":2891,"text":"Thanks for watching. Subscribe to Near Week.","url":"https://www.youtube.com/watch?v=D0mI_jbwkuI&t=2891s"}],"youtube_url":"https://youtube.com/watch?v=D0mI_jbwkuI","url":"https://near-pedia.xyz/videos/D0mI_jbwkuI"}