Hey Griffin community!
Last Thursday Oct 25 we hosted an engaging AMA session on Twitter Spaces with our product manager Michal, and we wanted to share all the exciting details with you. From deep diving into our reputation system to sneak peeks at upcoming features like Griffin ID and Griffin GPT, there's a lot to cover. So, grab a cup of coffee (or your beverage of choice), and let's dive in!
Reputation's role in web3
We kicked off the AMA with a thought-provoking question: "What's the role of Griffin AI in a reputational system in building trust?"
Michal delved into how crucial reputation is in the Web3 landscape, highlighting that it's a defining factor separating Web2 from Web3.
"I think that reputation in Web3 plays a central role, right? This is what actually splits apart Web2 and Web3. We are a Web3 company; we care about this reputation. It's not a buzzword for us in a sense."
He used relatable examples like Airbnb and Amazon to illustrate how reviews and reputation guide our choices in everyday apps. However, he also pointed out that in traditional systems, there's room for manipulation.
"Sometimes, in most cases, right—just to be perfectly honest with you—in most cases, when you check reputation, it represents the actual feedback from users. But if there is a shortage of flats, if there is not a high season, there might be tricks around. As a product manager, trust me, I know how these things work and how different managers might influence your opinion and influence your choice."
Our goal at Griffin AI is to eliminate these middlemen who can manipulate trust, creating a transparent system where reputation is genuine and unalterable.
Commitment to unaltered reviews
Michal emphasized that in our network, there's absolutely no way for us—or anyone else—to alter your reviews or reputation.
"By staying away from reputation, I mean there is not a single technical possibility for us to alter your reviews or to alter your reputation online. This is impossible for us. That's why we believe in our users; we think that people who understand AI agents and who understand what curiosity is—because it's all about curiosity—people who are curious, people who are smart, we could trust them, and they could trust us."
He stressed that our network isn't about us controlling anything; it's about providing the infrastructure for a trustless system where you have full control and transparency.
Three vital reputational elements
When asked, "How can individuals track and verify the reputation of AI agents?", Michal broke down our sophisticated reputation system, which hinges on three key components:
Service quality
"Essentially, there are three elements of reputation that we think are vital: service quality, reliability, and user feedback. By service quality, I mean that...we have service level agreements. So, for instance, you want to run a node, and you have some agent...You provide a service level agreement—what your agent can do, what your agent can't do, what's its value, etc."
This means every service provider clearly states what they offer, ensuring you know exactly what to expect.
Reliability
Michal explained the importance of nodes being consistently online:
"Reliability, the second element that I mentioned, is essentially about your node being all the time online. Imagine you are buying a service from some node, and this node is down. And your business depends on the service you are buying. So your business model probably will be affected."
User feedback
Your feedback is crucial and stored on-chain for complete transparency.
"If users provide feedback, this feedback will be stored in the chain. So it won't be stored in a centralized database where people like me would have access and could alter this. This is impossible. It would be on-chain, right?"
By combining these elements, we provide you with all the tools to make informed choices, ensuring that services not up to par are naturally filtered out by the community.
Integrating identity systems
We also discussed the mechanisms in place to establish and maintain trust within our AI network. Michal highlighted the integration of identity:
"We have a reputation system which is quite sophisticated, but on top of that, we have identity—so creating identity is an integral part of the system. In a sense, you're anonymous, but you're transparent. And by combining reputation and identity, we could already deduce which users like, which reviews and which ranking they have."
This balance of anonymity and transparency ensures that while your personal details remain private, your actions contribute to the overall trust and reliability of the network.
Griffin ID launching soon
Many of you have been curious about Griffin ID, and we're thrilled to announce that it is planned for launch before Christmas!
"Let me start with Griffin ID. So, we will ship it on one of our tools... How it works, you just connect your wallet, and you sign the transaction, and by signing the transaction, you receive an ID, right? And in order to interact with this ID, you are the owner of this ID."
Michal explained how Griffin ID empowers you to own your data fully. If you're unhappy with one application, you can transfer your ID—and your reputation—to another without any interference.
"With on-chain ID, you could just transfer your ID from one client application to another client application with all the data that you have. So, if you are not happy with one application on a Griffin network, you could easily transfer it to another one and have your reputation always with you. No one can alter it, no one can change it."
Exploring the Agent Playground
We understand that interacting with AI agents might be new territory for many, so we've created the Agents' Playground as a starting point.
"Agents' Playground is not a Web3 application in its current shape... It's just a playground, right? It's like the first baby step to our network. So why did we decide to ship it? We decided to ship it to let our community, our users, all the tools to check agents before we launch the full-fledged network."
This platform allows you to experiment with AI agents, provide feedback, and even earn points toward our airdrop program. Your interactions here directly inform our development, helping us shape the future of the platform.
Inviting user participation
Your feedback is more valuable than ever. Michal extended an open invitation:
"If anyone of you wants to have a quick interview with me, I would be really happy, and we could reward you with some points for that. So just like if you use one of the agents and you want to just give me a review, give me feedback—negative, positive, what you like, what you don't like—just please message me."
We've already implemented updates based on community feedback, and we encourage you to keep sharing your thoughts. Whether it's a bug you've found or a feature you'd love to see, we're all ears!
Decentralizing data and storage
We believe in full decentralization—not just of data but also of storage.
"We can't only decentralize data. We need to decentralize storage. But also, on the other hand, I can tell you that it's a very complex process and... might affect user experience. So, in order to keep our user experience great and to still be sufficiently decentralized...we're moving in this direction and already partnered with some decentralized infrastructure providers."
While it's a challenging journey, we're committed to ensuring that our platform remains true to the principles of decentralization without compromising on usability.
AI Agents without censorship
One of the questions we often get is how we differ from other AI projects. A key differentiator is our stance on censorship.
"In comparison to most players down there, we don't censor our agents. So, all agents have guardrails... But for instance, there are questions that are censored in ChatGPT and most of the apps down there. So, what we do, we refrain from censoring our agents, and you could check it yourself if you navigate to Agents' Playground right now as we just launched our Llama agent that has no censorship."
This means you have more freedom to explore and interact with AI agents, getting unfiltered responses while still adhering to ethical guidelines.
Announcing Griffin GPT model
Michal also spoke about some big news that we're super excited about, the upcoming launch of our own AI model, Griffin GPT!
"Very soon, expect Griffin GPT, and I think that will be our first public model."
This model represents a significant deployment for GriffinAI, and we can't wait for you to experience its capabilities.
In closing
Our AMA was packed with insightful discussions, and we're grateful for the fantastic questions from our community. We're on an exciting journey to revolutionize how AI agents interact within the Web3 space, and we're thrilled to have you alongside us.
From enhancing our reputation systems to introducing groundbreaking features like Griffin ID and Griffin GPT, we're committed to building a platform that empowers you.
So, what are you waiting for? Dive into the Agents' Playground, explore our AI agents, and let us know what you think!
Try the Agents' Playground for yourself today! At app.griffinai.io...
YouTube
Discord
Telegram
Twitter