Acurast AMA Segment 8 by Numbers Official

00:02:52 Hi Elena. Can you hear me? Yes, we can hear you. Okay. Nice. Nice. Yeah. Thank you for hosting this space. Uh, I saw Apostle also there. Really happy to be talking in the same space again as you guys. I know it's just been a week, but how are things on your side? Uh things are very well on our side. Thanks for joining the spaces and thank you for having me once more Elena as one of the uh people representing Acurist uh on this space. Wonderful. Um so numbers intern, we don't really know anything about numbers protocol. Could you tell us what numbers protocol does and um perhaps explain it in ways that our new community members understand? Okay. Uh I will try to make it uh simple but uh it may take quite some time. Uh I need to tell a story for for that to be able to happen. So a bit about numbers. Um we built provenence infrastructure for human and AI. That's the short version of it. Uh now you probably wonder what provenence mean. So essentially it means the full history of something and in our context in what numbers do this something refers to digital content. So in short, we make it possible for people for you to trace the full history of your diesel asset. For example, let's say uh photos you see in social media in Instagram or X three to four years ago uh before generative AI is popular. The only thing you need to worry about when you see this image is probably Photoshop. let's say to to know if whether this image is edited or is it just a photo taken out of context but now generative AI has become so popular and so the problem becomes so much more complex AI can generate image context videos text anything very uh realistically very humanlike uh so if we are talking about image or video deep fake for example. Now it can be easily created and it will only be a lot harder to detect whether this is real or not as time goes on. So we make it possible for you to trace the full history of this media. This mean you will know who or what created an image when you see it. You will know what tools are involved in it. You will know whether this uh photos is taken with a camera by a human or is it machine generated by AI. Which AI tool is it? When it was created, the history of its edit, if it was edited like from the first very first edit to the final one. And if you license this image, you can even uh we we we have this license option. You can expand it even further like setting up a contract where this certain image or video or asset may only be distributed once or whether you can set if it's allowed to be used for AI training or not. The possibility are uh infinite I will say. Thank you for such a great explanation. Very interesting to learn more about numbers. Apostle, how would you describe what Acuras is? Yes, in simple terms uh I'll describe Acurist as a decentralized cloud that is powered by smartphones uh instead of servers. Uh but what makes us different is that all the smartphones that have been onboarded on and at the moment we have over 250,000 smartphones on boarded all of them have a trusted execution environment within them which allows for uh confidential computing especially when it comes to use cases with more uh sensitive data. But in simple terms, this means that even the hardware owners uh so the people that provide the their smartphone uh the processors cannot see what is uh the code being the code that is being run. And what has been happening in both of your protocols in the last few months? Are there any major milestones, major updates that you'd like our community to know? Uh may I go first? Should I go first? Yes. Go ahead. Okay. Uh so in the past month uh it has been more than a month I would say. Um we launched the X42 ecosystem in numbers and we actually has onboard some very notable news agency into it. Uh I'm not sure if you guys know about them but in Taiwan they are pretty big. uh there one of them are called China Times. So what this mean is agents can license or buy a license of a content directly without human intervention. In the context of China Times, the photo journalists working with China Times has uh licensed some of their works and we actually have seen hundreds of agents actively involving their photos. Some of them uh buy the buy the license of their works for a certain uh for a certain thing like uh creating an article and there are a lot more others than that. I actually have a follow-up question on X uh as you said on the X42 because because that is something we are engaging in uh on Ericen. I think we also integrated with bass around a month ago and I personally see it as the next step towards the agentic internet. So is this is this how you also see it uh as the next step is basically where multiple agents would be able to communicate with each other when an agent calls on another agent and becomes sort of a mass agent system. That makes sense. Yeah, it makes sense. I believe it is. Uh this is something uh I would like to share because uh the way we came into this is uh probably a bit different from what people has think about. So uh we we are both projects working in web three in the crypto environment, right? So uh there has been a lot of conjecture or maybe uh speculation uh about web3 and crypto. Uh yeah because it is the game pu uh projection or prediction is the game uh of web three and crypto like speculation right so there has been doubt uh I I would not say we are doubting uh that crypto or web tree will be uh will be the future but I have seen lots of doubts from people uh outside webree or in or even inside webree inside crypto So people has has been talking about how uh it might not survive this cycle or maybe uh it might be over is if somebody big like my uh sailor sold their bitcoin or something which is what happened few days ago and people panicking. So our approach to uh x42 is is a bit different. We think if if you are working on two things crypto and AI which of this do you believe is uh has future have have something that has a longerl lasting future without speculation I believe a lot of people will say uh cryptobubble may burst I don't know uh out people outside crypto outside web three may say so but people inside web three would not But that's like 50/50, right? But if you ask this same group of people the question, will AI bubble burst? Then you will see a lot more of them will think uh AI actually will not burst. AI will uh we we probably could have like 70% or 80% of them agree that AI is here to stay. So if AI is here to stay then what about crypto right I have I have a lot of talk with our founders and we both uh end up with one answer because AI will stay what's the what's the best what's the easiest way for AI to pay something for agents to interact with other contents to have their own monetary system to to be able to do transactions What's the best way for AI to do that? Of course, it's crypto. Crypto is their native uh currency and so because we believe uh AI agentic flow will be the future because of that we also believe crypto will be here to stay and that's why we uh kind of arrive to this solution. then we need to integrate actual two. I hope that answer your question. No, I believe that makes sense. So in uh X42 would basically be the MCP for AI. I mean MCP being more of a connection. Uh X42 would be the payment the easier payment layer that could that makes this that basically puts MCP on steroids as far as I understand. Mhm. Yeah. Correct. Correct. Perfect. Yes. Uh just to answer the question on what's been going on across end. Uh so the on the technical side, the main update that we're very excited about is uh the cargo and the tunnel update. So this means that you're able to deploy uh multiple workloads on Acuras. What I mean by this is in terms of uh languages. So now Acuras supports Python, Go, Ross, C++. And the best way to think of it is it basically a decentralized Linux. Previously on Acuras it was we were quite niche in terms of what can be run. It was mostly JavaScript and NodeJS. But this m but this really makes it a true decentralized computing platform due to uh that the ability to handle workloads in multiple languages. Uh and then on the commercial side uh what we've decided is what would be very interesting is actually to build products and services ourselves on top of Acuras that we can uh then sell to enterprises. So in this case we have two products. The first one is a cyber security product that already is working with enterprises in Switzerland. And what it does is it's able to it's able to produce uh realistic uh cyber security attacks. And the other one is our own scraping and data extraction service. And in our case uh what we are able to provide is bandwidth and compute from the same device which makes it much more efficient. uh when it comes to the particular scraping use cases, both of your projects emphasize um Oh, sorry. Go ahead. No, sorry. I just wanted to uh ask a bit uh about the data extraction that Aposto mentioned. Uh I'm quite interested in this because uh I think lately I I don't know about you guys but in my timeline both numbers timeline uh and my personal X timeline I have seen a lot of people uh building something uh with AI. Uh the most common thing you will see is uh in X especially is people who try to find uh workflow to automate their posting or to help them find a good angle to do to post. Uh I believe this is because they wanted to utilize the X monetization and one common problem I see that they usually have is uh scraping the data actually scraping their own personal data right so it's probably uh depends on its person it depends on how they're approaching training their agents to do something uh some of these attempt may be successful uh The agent may probably create something very similar to what the users usually create. But I think a lot of people struggles