Inside the Proof-of-Control Stack | Intelligence at the Frontier 2026

Okay, so here's a few things that just happened to catch you all up. We just had a really exciting interview from our advisor Ryan Bellendorf, who talked about proof of control and how he sees this as a new kind of, uh, movement that needs to happen in a similar way that he and, uh, a bunch of colleagues launched open source movement, right? The open source initiative. They were able to define it, make it legally defensible, make it commercially viable. And now we're in a similar moment where we have all this rise of security technology that's, you know, enabling agents to communicate. You can have control over technology, but all these people call it different things. And so what we've done at the Society is we've rebranded it. We are calling it proof of control and to say you can have independent provability. And regardless, all of our members offer something like this. And this is the new category of technology is needed. It's not just compute, but the real, uh, need and the real moat is gonna be in verification technology such as yours. And so this is what you've been hearing, what we've been hearing all day in various instantiations. And so I wanna have a start off with actually Shane. We've been talking about agent orchestration. We have Shawal Walters talk about Eliza Labs, and we've been talking about, you know, just what happens when agents are out of our control. But I want you to tell us, you know, what exactly is the problem that Convoz is solving? You know, why do we need a messaging, new kind of messaging app when I already have WhatsApp and Signal. I have regular messaging for people who have Android, and for Chinese apps, I have to manage Line or WeChat. And then for my Japanese and Taiwanese friends, I have Line on top of that. And then we have LinkedIn, all these things, right? So what is it that Konvoz is doing?

Yeah, I'll start by saying I don't want another messenger either. Neither do I. Um, but I think the current messaging infrastructure is about to collapse because it's all built on phone numbers and email and single identifiers that once you turn the agentic world onto it, they're all already collapsing. So every messaging communication channel I have is a massive spam inbox. I get added to crazy groups. I'm always getting scammed by random people. You can't verify who you're talking to. And so at the core is just a new messaging protocol for humans and agents to be able to communicate. While also being able to be protected from each other. And I think the next era of communication is more open, it's decentralized, um, and it means that no single country can also shut it down. So the thing we've been building for 5 years is the protocol, um, on the mainnet, no country owns over a third of the nodes. So therefore, if any country changes their laws on privacy and encryption, the network never shuts down, which means thousands of apps can launch, uh, and you don't end up in a world where WhatsApp's banned in Iran, Turkey banned Discord, France throws Paval in jail, UK backdoor in the, to iMessage and iCloud, Signal threatens to leave with chat control. Like that's never gonna stop. And so I think you have to have a new messaging infrastructure for the next era of the internet.

And can you just tell us what exactly is different about Convos? And when you say it's a decentralized messaging network, how can I, I downloaded it from the App Store. I can use it.

Anyone can download it and come message me afterwards. It's very simple. It's just private chat for the Agentic AI era. And what that means is there's no phone number, there's no signup, there's no email, there's nothing. It's as secure as Signal, except it's not based on a single number that's going to spam you and scam you. And so every time you start a conversation, it creates a new key. I scan the QR code with the person I'm with, or I share it on a different messenger. When they join, that relationship is unique. Uh, and if I want, I can set an explode timer on it. and it burns the keys on every single person's phone who's in the conversation at the end of it. And the cool thing is you can invite people or you can invite agents. And so the safest place in the world to use any AI model is inside of Convos because it never has anything about you. And you choose every single time whether to use your name when you show up.

In every single chat, I already have 5 chats going and I am a different identity in each chat.

That's the most important part. Every time you start, it says, do you want to be Shane? Or do you want to be somebody? Or do you want to be a nickname? And it's really simple. I can just tap the button or I can just stay between us. And if the 4 of us joined a convo, we named the group the stage. The trust lives offline. We know it's the stage. The app knows nothing. So it's fully open source. We can't see anything.

I love it. So I think this is really going to change how people manage their OpenClaw. And, um, you know, right now they hooked it up to their WhatsApp and their Telegram.

It's fucking nuts. Like you put it on there, it knows your contacts, it knows your phone number. It sees to do with it, it's gonna be a complete disaster.

Yeah, I, I agree. It already is. You already see these stories of, you know, people not knowing what their agents are doing. So with that, I wanna, um, move to Pranav. You know, we're talking about agents and being able to control them. Well, what AMP, what Edge Node has done is building out Ampersand. And I think this is really important because it is the first product we've seen in agent orchestration that allows you to control agents beyond the command line. And you actually took UX into account and you're looking at the client side and the client interaction and it Visual Dashboard. And so could you tell us, you know, what is the problem you're solving and what's the proof of control aspect that you're solving?

Sure.

So is this working?

It is.

Okay. Oh, so first of all, a very interesting topic of the debate that you picked up, as in right now for the last 3 years, we were at the stage where everybody wanted to make the AI agents better. Everybody, including Google and their grandmas, were working towards making AI agents become more intelligent. At this point, at Opus 4.6, we have reached a stage when I feel that we sort of can, you know, say that sort of intelligence has been reached. And now it's about efficiency and guardrails. That's where we are lending towards. I was attending a very interesting debate upstairs between open AI and closed AI. And why some of the AI needs to be open, why some of AI needs to be closed from Google itself, who was advocating for closed and some part, you know, it was about open. And I feel there are certain parts when we were creating AI, when we were racing towards AI that we didn't think about. The first of all of them was energy, right? Right now, 2% of world energy is currently used in AI. This is super damaging and it will go even higher and higher. You need to make AI agents much more efficient. That's one side. The second was right now you give your AI agents an open guardrail to just go and do whatever it wants. And that is another crazy thing. Just imagine you're right now in a highway and that highway does not have guardrails. Highway is made really amazing. It has 200 kilometers per hour, or in US you say 160 miles per hour. I don't know how you use these things, but, you know, but there's no guardrails. Would you sort of drive on that specific road? What do you enable? So that's where the reality comes as in like AI agents need to be more— AI, first of all, needs to be more efficient with energy. And secondly, you need to have guardrails. And that's what Ampersand's Sort of is trying to, you know, enable. I feel AI agents with OpenClaw sort of showed that it can do whatever it wants, but that also means it can delete your mails, get into your private photos, and also start trading on your behalf. I gave my OpenClaw 2 ETH and it sort of revealed my OpenClaw information to somebody and it gave it my private key and then I lost my 2 ETH.

Well, what I love is that when I was using Ampersand and Rodrigo gave me the tour with Noel on the spot, what blew my mind when I was at, you know, the house, their offices, is that literally it was this beautiful dashboard, indistinguishable from any other kind of SaaS dashboard. It was like they had this amazing UX designer who actually understood what a client would need to control their agents. And it was literally just controls. I could move on X402, in terms of the amount of the payments, in terms of how much the payment could be, when I could stop the payment, and I could track the authorization level down to the person who initiated it. And then you could really see, I was like, oh, this is going to be game-changing for insurance, it's going to be game-changing for claims, because you're going to— you're not relying on attestations anymore or logs that you have to construct ex post facto, is that you have the mathematical proof right there. You had the hashes on the spot. So It was really cool to just experience it myself. So I think your team has done a great job. So in terms of Anna, when we're talking about proof of control, when it comes to data portability, this is something that VANA has really solved for, solved for. And so what is the problem that you see right now? Because, you know, data portability and data privacy is something that we don't have. We did not have it built out well in terms of what the original internet has given us, in terms of the way we've been trying to patch. So why is VANA needed? Who are you solving this problem for? Who are your clients?

Yeah, so, uh, oh, I, I just have to turn it on. Sorry. Um, I think like often when people hear the term like data portability or data privacy, they almost like tune out. Like especially in crypto, we've been talking about it for so long that it's like, okay, if it's not here, kind of when is it coming? But I think really the way to think about portability is kind of, um, as AI becomes more and more important in your life, like data is the context that actually makes the AI work really well. Right? So kind of as a thought experiment, if you think over the past 2 years, if you could either delete all of your chatbot history, ChatGPT, Claude, or all of your messaging history, or all of your, uh, health history, like I personally, the thing I would be most hesitant to delete is actually my chat history, 'cause that has so much valuable information that allows me to like do my work. Right? And in, in terms of the way the internet is set up today, most of that history and most of that context that allows you to be productive is living inside of whatever platform that you're using, right? So your ChatGPT history, your memories, that's living kind of with OpenAI, your Claude history, that's with Claude. And the approach that we've taken with Vana is that really your data should be stored with you, right? There's a photo that I love from the early 2000s of these MIT Media Lab researchers who are physically wearing a hard drive like on their belt, right? Right? And their view was like, as the internet kind of gets bigger and bigger, people use it for more services, people are going to need to physically carry their data with them. Right? And the internet has played out completely differently. Right? Most of the time your data is stored with a platform, but you have full legal ownership over your data. Right? So in the same way that when you put your data in, or your car into a parking lot, the parking lot doesn't like own your car. They can't go take your wheels. They can't say you're not getting your car back. Like it's still yours. And that's true as well of kind of every single platform you use. It's true for some surprising ones too. Like if you're a Tesla driver, you can even kind of pull some of the Tesla video data, which is really interesting. So getting to kind of like, who do we solve this problem for, right? So as an individual user, it's really about giving you more choice across different platforms. It's super inconvenient to go export your data. It's kind of intentionally inconvenient. Data regulation makes it easier and easier, but some of the tooling that we've built out on top of Vonum makes it super easy. So it kind of one-click logs in as you go, grabs a bunch of your data. That's this tool called Data Connect that you can use to kind of easily port your data on Vana. And then the second group I wanted to talk about is really builders. Like, one of the things that got me into building Vana is that I was kind of training models at CSAIL, which is MIT's AI lab, and I was seeing the datasets I wanted to train on, I didn't have access to. And what makes AI really good is data, but it's really hard to access that unless you basically work at one of the big companies. Right? And so I think that there are so many really great consumer applications and just products that would exist if a user could actually bring their data with them across different applications. And so kind of that builder piece I think is really important as well.

Do you see that the market demand is there for your products? Like, is that happening from the consumer enterprise side? Are we finally in the moment?

I think it, it depends on how you frame it and it depends on sort of exactly what group we're talking about. So I think if you go to like some random person on the street, even in San Francisco, and you're like, do you wanna own your data? They, although they probably won't wanna talk to you, but like assuming they did, they would just be confused. They don't know what that means. But if you frame it more as, hey, do you want to take your ChatGPT history and have it with you in Claude? That's something that really resonates. I think it's still somewhat of a niche audience because kind of the idea of what is data, what is context, like my family in Minnesota does not know what that is, right? That still doesn't resonate. I think it's getting there. And then on the builder side, I think that is mattering more and more. And I think that increasingly people are realizing What makes my product good is access to data, right?

So how do— what are you all seeing then in terms of, for your products, for the problem you're solving? Or is there a demand for this from consumers, um, from your, from people who would be downloading this app? Like what gave you the signals that this, people were ready for Convos?

Uh, Convos is really early. I mean, we've been building XMTP for 5 years now, and there's billions of messages going through the protocol. And I think it's because the companies building on top of it own their data. So XMTP is just a really secure pipe.

So could you tell us about what kind of companies are sending messages?

Coinbase uses it as their main messaging protocol. Uh, World just launched a private messenger to compete with Signal that doesn't have a phone number, and it's the first ever proof of human messenger that is more secure than Signal. Um, and so you have like companies at that scale, you have a lot smaller companies. Um, and Convos is like our take on what a private messenger should be in the AI world. Um, but yeah, I think it's because you have portability. The user owns the key to their messages. The companies don't have to sit on this crazy logs of data and the device and the application layer actually is the access point for your communication. Um, and if you don't trust that company, you can leave and take it with you.

Yeah.

Specifically for us, ours started with like a use case of somebody who was running two agents to build a coding mechanism, and that company lost $47,000 in 7 hours because nobody was observing these agents and they were working typically out of 9 to 5. So that was the case study where we were like, okay, AI agents will create code. If you don't think that, you are either weird or you live in North Korea or something like that. You know, so either that was the use case that we saw that, oh my God, AI agents can hallucinate. They can sort of keep convincing each other that they are solving a problem, but they end up sort of hallucinating and just increasing your bills. And that's where I feel like in 2011, if you were launching a server in your company, you would have a Grafana dashboard for observability. Without a Grafana dashboard, nobody launches servers. So if you are becoming enterprise, which will sort of launch agents and make sure agents are doing all your coding, all your scraping, all your data analysis, all your work, and you're just an orchestrator, you do need a way to observe it and control it. And that's where something like Ampersand was born. And I think since then we have seen a lot of guardrail platforms trying to do something like that. And I think that's the step in the right way. Where intelligence has reached, we might have reached AGI as well. But now is the time where we guardrail it to make sure we are the ones who control you, not you who control us.

Well, then what are, you know, since we have like the signals that you're seeing, but what are the biggest blockers into adopting proof of control or your technology from your perspective? Like, what are you experiencing?

I don't think consumers care. I really think you just have to build incredible products. and it has to be a faster way to communicate, a more fun way to communicate, a better way to use AI that actually makes it quicker and easier in a group to do something I couldn't do. I, I think much of the space, um, that I'm in has been in philosophical bullshit for a long time. And we—

I'm so happy to hear you say this.

We just pontificate on all these aspirations of purpose. And I think you deliver on your mission to make the world more private or more secure or own it, whatever. If you build an incredible product that actually delivers something that people think is magical, and then by proxy of that, yeah, they get the thing that you want to give them.

I think, you know, part of what I loved about Convos is something that I don't see enough in products that offer a lot of the value of what you, the kind of interaction is the design of it was how smooth it was. And it was like just one click.

It's the most secure messenger in the world. It's also the fastest. If you're with a group of people, we have a time to chat. In 5 seconds, every single person who doesn't have the app, 10 people together can scan a code, be in the chat. There's no signup.

Yeah.

I mean, I experienced it on the call with you. It's the fastest possible thing in the world. Literally, I was on the call with Shane and I'm like, I can use it now. He's like, yeah, I just sent you a link. I'm like, to where? He's like, yeah, to your Signal. Clicked, boom, we were in a conversation, encrypted keys. We, it was on the moment. He's like, yeah.

What is encryption? What are keys? What are things?

No one cares. Well, you didn't know, it didn't reveal any.

Click a link, done.

Yeah, that was the cool part was like, oh my God, I don't know how many crypto products there are where it's like, I'm so bogged down by, you know, all the things that's showing. And it was like, no, it's just focused on a really great quality interaction. Yeah.

And most people aren't like, what is crypto? Most people are like, fuck crypto. So it's like, it doesn't matter. It's just, is it incredible and does it deliver on the promise? So yeah, Convo's Messenger, if anyone has it, scan it afterwards, I'll drop everyone some free shit.

And would, Would you all agree with what Shane just said was that, you know, he was saying consumers don't care, right, about this, like the value proposition of, of decentralization and like privacy. It's that it's not that, I mean, yeah, they don't care. So what do you think was, is that something you would agree with?

Yeah, I think, um, fundamentally someone doesn't wake up and they're like, let me go use a product that I'm ideologically aligned with, right? They're basically like, what can I do that I couldn't do before? And it's very tactical in terms of kind of like why I think consumers generally choose to, like, I don't know, why would someone join a data union on Vonna? Why would someone choose to port their data in? And I think that basically framing it and making it as simple as possible as like, this is an easy way to go and get a bunch of your data and use it as context for AI, ignoring, I mean, there's sort of all sorts of stuff we could say on the non-custodial data side of like, hey, you control your keys, this is stored fully locally. But a lot of that, I think maybe there's a, a small group of users who kind of dive into it, but most people don't care. I think it does matter in the very long run, right? As you think about building an open protocol, I think there are some regulatory aspects too where it comes in where for Vana's case specifically, we're typically looking at taking data from a user and allowing them to take it out of a platform themselves. A company wouldn't have that right, right? So it does actually have to be the user directly doing it. And that's maybe one place where that decentralization piece is pretty important.

But I think what you're saying is you introduce is that it's, it's part of like, I think, yes, surface level, it's like people don't care, but we've also been trained to also not know what's possible. You know, we've been trained to be users and to not know that you can be a shaper and determine how your conversations initiate, or that you would have privacy, or you were able to have control over your social graph and control where you're, you know, how you access your data. And that this is a, you know, the data's like my arm, right? We don't, we're not trained to think that, you know, this is the way our data's been captured. And the way we use social media, um, and the way, you know, Facebook owns, you know, WhatsApp, all of this. Like, we're not, we've, we've been, there's systems in place to help us not care. And so it's, I think, a really big uphill battle that the companies that, um, that you're, the solutions that you're offering. And I actually think the UX proposition and the design proposition, the engagement has to be even that much better is that you don't, you don't have time to train someone to, undo the systematic, you know, um, brainwashing that you don't have access to your data. Like, there's just no time for that when you need someone to click download and use. And so I actually think you have to work harder on the design side. But unfortunately, many tech companies, and I would say even more so at, um, you know, any kind of these proof of control companies, is that oftentimes design, um, and marketing, all these things that absolutely matter, are not given a seat at the table. Yes, claps to that. Um, so how do you do it at your company? I'm curious, like, literally, can you tell builders in the room, how have you made sure to prioritize? Because I've, you know, seen, you know, your products evolve. Um, I got to know Convo just over a couple days ago, and now I'm like, you know, like a family.

I mean, ours is simple. How do you make a world-class designer is the leader of the whole thing?

Boom! Can we— I'm just gonna A world-class designer is the leader of combos.

Like, it's taste at the top level, and he's the person who owns it all. And I mean, not that it's like, oh my God, it's great, but if you use it, you're like, wow, this is a great product. And it's because of him. And he really does hold that quality bar at the highest level. And then it's, you know, the iOS, Android, web engineers have built world-class consumer apps. And so I think. It has that bar. And now we're not writing any of the code, like it's all done, but like the frameworks and the training and the design like skills and everything we've built to make sure it's at a certain bar is something that we spend a lot of time on.

I sort of agree with most of what you've said where UI matters the most, but if UI was the most primal thing of, you know, making a change, then platforms like Reddit won't exist. Right? There needs to be a lot of value where a lot of specific people can be attracted. And although what's happening with crypto right now, and we are all saying that wallets are BS, we gotta have fiat on-ramps and just do things and just go from there and just sell to enterprises and don't care about decentralization, open source, just build a chain with one server and sell it to Google and make millions. I think there is somewhere where the answer lies in the middle where you cannot leave your core of what you came from and you gotta make design improvements to make it specific. Specifically on your answer as what is our biggest problem of selling to AI people? Our biggest problem is that AI people hate crypto, right? And there are reasons for that, that Open Claw came and then, you know, crypto people started launching meme coins and everything and people went crazy. And right now Ampersand sort of controls or sort of gives your Open Claw a wallet and then stablecoin-based payments can happen through. It's an amazing solution, but people do not want their Open Claw to have a wallet just because it starts getting related to crypto.

Crypto.

They start calling it machine payments just so that nobody thinks it's stablecoin, crypto, or meme coin.

Well, we can call it machine payments. I mean, who cares? But if you want your agents to not have to wait 3 days, 7 days, 10 days for settlement. and you need to be able to have authorization, you wanna do micropayments for anything, right? 0.001%, like, you need to have agent payments, like, period. You have to have xortu. There's literally no other option for you. You— there's— and actually, traditional banks are integrating, you know, crypto payments. So we can call it machine payments. And to be clear, by the way, I said UX, not UI, which is a totally different category. But you have design encompasses UI and UX.

Yeah, Reddit's great. Like, we're not talking visual design here.

Yeah.

We're talking great experience.

Right. And, and it's beyond just visual, right? 'Cause the visual and with the strategy, you have to hold all that. So that's why I love that you said at your, at Convos, like, it's actually a designer that is accountable for the product. Of course, it doesn't mean that you're, you know, putting, um, a lower value on engineering and you're saying no to all these other, you know, design fea— or product features. So I love that. Um, for Ivana, how do you think about, you know, design, um, in terms of what you're producing?

Yeah, I think, um, sometimes there's almost this mental, this like kind of education work of like, what's the right mental model to give a user, right? And like, how do you communicate like your data is stored on your device, right? Sometimes actually rather than saying like it's private, it's like, hey, this is just gonna live on your desktop or just gonna live on your phone. So I think that a lot of it was about like, how do you communicate the core message you wanna communicate in like context that somebody already understands? And then I also think, I, I mean, increasingly with AI just accelerating like software and coding so much, really making it so that when a user is interacting, it's in exactly the right way, right? So each of the different data unions on Vana has their own interface, right? The way that you connect your Tesla data is actually different from how you would connect your Instagram data from how you connect your Spotify data. And I think in a lot of ways that's an advantage because there are so many different people on this world who have different contexts in how they're coming into software and being able to give like a specific kind of experience to somebody depending on where they're coming from really goes a long way.

Yeah. And I wanna remind everyone that one of the highest paying jobs right now in Silicon Valley are communication strategists. Communication leads because—

Because they're doing PR for him.

Yes, exactly.

Yeah, I sort of spoiled it for them, I guess, if it's regarded.

Yeah, don't give away the secret sauce of Ampersand, you know, you're a little—

What I said is a real story. I gave it to E, he blow it up, and that's why we created Ampersand, remember?

Absolutely real. So with that, you know, I want to end with this one question that I— this paper that's been circulating around where You know, Christian Catalini talks about these, the new need for verification technology and that it's gonna be security companies and application solutions that offer this kind of cryptographic security. And of course, we don't have to call it that to the people using it, but underlying the layers, it's gonna be like the new TCP/IP. It's gonna be invisible. What are your thoughts and experience from people who've been in this industry for a while? We finally, you know, you've been building out your protocols or you've been building out your solutions solutions, you've been building out your index, whatever it is, you know, we're, we're finally at the point where it's not about Web3 versus Web2, but you know, are we finally there where blockchain and decentralization, independent proof, all this stuff has finally found the largest use case around agent orchestration? Like what are your immediate thoughts on that?

This is an amazing question, but I'll just break it down very simply. First of all, verification is an absolute need. Right now, I trust my OpenAI more than, you know, a lot of people, including my therapist. What that means is that if I'm having an argument with my mom, I'll go to OpenAI and find a solution and sort of start trusting it. And this is the state of the whole social world right now. So if in the future you're interacting with these OpenAIs, and you start getting solutions which are not from the real model, how are you able to verify it? You go to ChatGPT and you sort of just get an answer, but it's from a very cheaper model than what you actually paid for, just for them to like, you know, give a less cost of it. So verification of getting prompt solutions is a very, very, very huge problem in this modern day and age. And as of right now, most verifications have to either be TEEs or ZKPs, and both of them are sort of somewhere where crypto industry has its head or mouth at. And I think we don't have to make it crypto. We don't have to have a token for it, for God's sake. And we don't have to sort of also allow it to be specifically blockchain-based. But these TEEs and ZKPs that something like Near Intense is doing would probably be the solution through which I feel in the future, all these front ends can be trusted or should be trusted while things still remain private to yourself. For example, if I go to OpenAI, just to conclude, and I sort of write, okay, what are— and I'm in a Republican state and I ask about abortion or anything, will OpenAI go and sell my soul to government? Maybe yes, we can think of it like that. So how do you keep it private but make sure there is attestation and cryptographic proof? That is where crypto heads at. And I think all the ZKP solutions, all the TE solutions that crypto companies are providing could be a way to provide it as long as they don't meme-ify it with a token.

I would say at the Dancy Eyes Society, we're identifying those companies. We're figuring out who are the ones that can really offer those solutions. Be enterprise-facing, be consumer-facing, get those certifications, not just SOC 2, but all the other things that are required to actually operate a company and, you know, provide something that's sustainable. So, what are your thoughts on, is this the moment for, you know, verification, independent provability technology?

Yeah, I guess one interesting, like, thing within verification today, I think we're seeing is that, like, AI agents are actually the people making a lot of decisions more so than it's it's like I was talking to a consumer AI builder, right? Founder, non-technical, and he was asking about some of the guarantees around Vaana as he looks to kind of integrate it for data portability. And he kind of followed up with this like really long detailed technical message and he was like, oh yeah. And I was like, this is a lot of detail. Like you're not technical. But he had talked to an AI in depth about exactly what are the technical guarantees I get from Vaana, right? And so I think that one of the interesting things that comes from AI agents and kind of everyone having on-demand access to like pretty strong intelligence and knowledge about software is that the verification guarantee is actually something someone can check and really go and look at and be like, oh, this does actually allow me stronger privacy. Um, and then the other thing on kind of, I guess there's kind of a difference between verification and privacy, right? I think that sometimes privacy is actually hard to verify, right? There was the lawsuit with ChatGPT where even if you are in temporary mode, they can still be saving those messages, right? And I think that that's a case where it's, um, the way you make that verifiable is you just need to be telling someone, this is where the computers run, this is where your data is stored. Um, so that's maybe a, a bit of a technical detail, but some of the nuance around verification.

Yeah. And it has to be independently provable because with any other company right now, you're gonna say, oh, we, they're gonna say we've did this, but there's no way to prove it independently and there's no way to ensure that it's tamper-proof.

Yeah, the way to do it is to open source it.

Yes.

And I mean, it feels like we're just reliving social media again. I think the reality is you need to be able to verify things. Open systems are the only thing that can really be trusted and independently audited. Um, and we need safer spaces to be able to choose how we show up and how we use AI and get to choose which models we trust or not. Uh, and I think hopefully we are starting to get new products to allow us to do that.

Yeah.

And the verifiable stuff, our homepage of Convos is my favorite thing we've built. We literally put a ChatGPT button next to every single claim we did 'cause we fully open sourced it. And so for everyone in the world who has no idea how to do technical shit, you just click it, it copies the prompt over to whatever AI system you're using and it verifies whatever the hell we're saying that it's true.

Oh, I love that.

And we need just more of that.

And you also have the CloudSkills, right, already up on your site.

Yeah.

And also, like, if any company says trust me, you shouldn't. I think the most important thing that has to switch is it's just Zuckerberg says trust me about WhatsApp, and I say don't trust me because it's open source.

Amazing. Well, this is why we need to have agreements to underlying plumbing, which is one thing we don't have in place. Even within, you know, Web3 companies is that we don't have, we don't agree to that. So there's a lot of fighting that has to be done, um, that could be saved if we were just doing agreeing to underlying plumbing and then we can actually fight it out on the marketplace.

By the way, did you create this word, this phrase? This is a new phrase I need to adopt.

I was just trying to— It's from Advanced AI Society.

Oh, got it.

Yes.

It's new, like a few weeks old?

Absolutely.

Months old?

Yeah. You are speaking on our panel. In fact, this is a new industry association.

Love it.

I, that's true. So, how I—

You should start a PR company.

Yes.

Consult.

Yeah, yes. Well, this is why you're one of our greatest members, Edginode. So with that, I wanna conclude. Thank you for, this is, you are, this is one of the first Proof of Control panels. We only launched, first we launched at Davos back in January, and then we had our second one with the Linux Foundation just 2 weeks ago, and this is our third one. So it's really, I think, you know, when Brian Bellendorf was talking about what is that latent thing that was in the air at the time in 1998 that made him, you know, made a bunch of people to come together to coin the phrase open source, to define it publicly, to make it legally defensible and commercially viable. You are in that moment right now, and you'll be hopefully talking about it in a couple years and be like, oh, I remember that panel when I met these, you know, initial builders are actually offering security for agent orchestration. And now I know how to, you know, make sure my open clause set up to do messaging correctly. I'm not exposing it. You're not forced to just connect it to your WhatsApp. You know, you have alternatives now thanks to Shane. You now know you have alternatives when it comes to data portability. You can actually have your data on device and you can look to Ivana for other options and their enterprise partnerships are going to offer you other options. And you can also look to Edge Node with Ampersand to have real control over your agents. So thank you so much. And we're going to move on to the next panel.