1/ We proudly present the Sentient Protocol, unveiled at the @openagisummit this week.
Sentient is an open source AI monetization protocol that enables community-built AGI. The key innovation is Model Loyalty and a new format, the OML format, for representing models that enables them to be Open (download and use locally), Monetizable (track and monetize their usage remotely), and Loyal (locked for usage that do not conform to safe, ethical, values espoused by the model owner). More details follow.
2/ Current forms of predominant AI were built on public goods from years of open innovation, but they extracted the value maximally from these public goods without sharing anything with the contributors - and created closed source hegemonies and empires out of it. Additionally, it censored information and imposed cultural preferences, which stifles innovation.
Open models are torch-bearers of resistance. They provide an alternative for AI innovators to participate in the large AI economy. However, there’s no way to monetize them, nor is there a way to ensure they are used safely and ethically.
We need a new ecosystem where open-source public goods of AI drive 𝗼𝗽𝗲𝗻 AGI innovation. There is an urgent need for a new technology that allows builders to share models openly and yet get rewarded when those models are used. There is a need for a new protocol that aligns incentives of AI builders with AI innovation. @viswanathpramod
3/ The Sentient Protocol is a blockchain protocol for solving the alignment problem of community-built open AGI. It comprises contracts for incentives (ownership, usage, rewards) and nodes to enable decentralized control for access and alignment. The incentives and the necessary crypto-economic security is enforced via AVSs from the @eigenlayer ecosystem along with a trustless blockchain connected to @0xpolygon Agg layer. @hstyagi
4/ Underlying the Sentient protocol is a new cryptographic primitive called OML (open, monetizable, loyal). The goal is to allow the model to be transparently downloadable and yet retain the ability to track usage (monetization) and ensure safe and ethical usage (loyalty). The cryptographic primitive of program obfuscation will also solve OML, but that is a long standing open problem.
Sentient is devising 𝗔𝗜 𝗺𝗲𝘁𝗵𝗼𝗱𝘀 𝘁𝗵𝗲𝗺𝘀𝗲𝗹𝘃𝗲𝘀 to create OML libraries for AI models – the birth of a new area that we call 𝗔𝗜-𝗻𝗮𝘁𝗶𝘃𝗲 𝗰𝗿𝘆𝗽𝘁𝗼𝗴𝗿𝗮𝗽𝗵𝘆. In the first version, we convert backdoor attacks (security threat in AI) into model fingerprinting methods to authenticate model ownership. @sewoong79
5/ The Sentient protocol is modular and can be composed with other implementations of OML (say, using trusted hardware) in the distribution layer or different decentralized storage or compute networks.
6/ A detailed description of this thread is available here:
A few weeks ago, researchers from Sentient and Princeton University exposed serious vulnerabilities in crypto agents managing real funds, using @elizaOS as a case study to highlight the financial risks in today’s agent frameworks.
Now, the team has expanded the research, introducing quantitative benchmarks to measure the severity of these attacks and showing how the threat extends beyond crypto agents to any agent that handles sensitive personal data.
We’re thrilled to officially announce our strategic partnership with @0g_labs, aimed at pioneering AI innovation in web3.
This collaboration merges Sentient Chat’s cutting-edge AI with 0G’s agents and robust on-chain data infrastructure. Together, we're set to redefine user experience and functionality in the web3 space.
2/ 🤖 0G AGENTS ON SENTIENT CHAT
Excited to integrate top-tier agents from the 0G ecosystem directly into Sentient Chat, kicking off alongside their Testnet V3 release.
Here's a sneak peek of what these agents can do:
- Upload files and share information/memories with each other
- Manage dex trades
- Automated LP position management
These and other industry-leading agents will be available soon, powered by Sentient Chat’s best-in-class search technology.
3/ 0G’S ON-CHAIN DATA FOR SENTIENT CHAT
We're also bringing 0G’s comprehensive on-chain data into the Sentient Chat ecosystem.
0G’s data layer provides extensive, high-quality blockchain insights through a secure pipeline, empowering Sentient Chat users with richer context, precise analytics, and more actionable data for every interaction.
Perplexity has done great work in the last few months, but the world deserves an open version.
Closed-source AI told us it was hopeless👇, but we’ve built it for the community anyway.
Experience Loyal AI and read more about it below🧵
𝟐/ 𝐏𝐈𝐎𝐍𝐄𝐄𝐑𝐈𝐍𝐆 𝐎𝐏𝐄𝐍-𝐒𝐎𝐔𝐑𝐂𝐄 𝐃𝐄𝐄𝐏 𝐒𝐄𝐀𝐑𝐂𝐇
Sentient’s research team has quickly become one of the leading teams contributing to AI-based search. They created a reasoning agentic search framework that is better than SOTA models under the FRAMES benchmark. This work will be openly published for all to use soon, you can read more here:
👉 tinyurl.com/OpenDeepSearch
𝟑/ 𝐀𝐆𝐄𝐍𝐓 𝐇𝐔𝐁
The most exciting part of Sentient Chat is its 𝐧𝐚𝐭𝐢𝐯𝐞 𝐚𝐜𝐜𝐞𝐬𝐬 𝐭𝐨 𝐚𝐠𝐞𝐧𝐭𝐬 (15 and growing). The agent hub will be a library of agents that users can access instantly for specific tasks.
The 𝐜𝐨𝐦𝐦𝐮𝐧𝐢𝐭𝐲 𝐰𝐢𝐥𝐥 𝐛𝐞 𝐚𝐛𝐥𝐞 𝐭𝐨 𝐩𝐥𝐮𝐠 𝐭𝐡𝐞𝐢𝐫 𝐚𝐠𝐞𝐧𝐭𝐬 𝐢𝐧𝐭𝐨 𝐭𝐡𝐞 𝐡𝐮𝐛 and collectively improve the knowledge available to users. Eventually, agents will collaborate with each other to carry out complex tasks, 𝐞𝐟𝐟𝐞𝐜𝐭𝐢𝐯𝐞𝐥𝐲 𝐜𝐫𝐞𝐚𝐭𝐢𝐧𝐠 𝐀𝐆𝐈.
We’re partnering with @kgen_io to launch 𝐚 𝐩𝐨𝐰𝐞𝐫𝐡𝐨𝐮𝐬𝐞 𝐦𝐨𝐝𝐞𝐥 𝐛𝐮𝐢𝐥𝐭 𝐟𝐨𝐫 𝐠𝐚𝐦𝐞𝐫𝐬 𝐚𝐧𝐝 𝐠𝐚𝐦𝐞 𝐝𝐞𝐯𝐞𝐥𝐨𝐩𝐞𝐫𝐬. By combining Sentient’s cutting-edge AI with KGeN’s proprietary data of ~300M attributes from over 13M gamers, we’re creating the ultimate gaming LLM.
⚡𝐖𝐇𝐘 𝐈𝐒 𝐓𝐇𝐈𝐒 𝐇𝐔𝐆𝐄?
This isn’t just any model—it’s an AI trained specifically on real gaming data, designed to improve experiences for everyone in the gaming ecosystem. Whether you’re a developer building next-gen games or a player leveling up your skills, this collaboration is going to redefine what’s possible in AI-powered gaming.
❓Join the 𝐒𝐞𝐧𝐭𝐢𝐞𝐧𝐭 𝐗 𝐊𝐆𝐞𝐍 𝐗 𝐒𝐩𝐚𝐜𝐞 𝐨𝐧 𝟐/𝟏𝟖 (𝟕 𝐀𝐌 𝐏𝐓) where @hstyagi and @ishank20 will share some exciting information on the model collaboration.
👇𝐑𝐞𝐚𝐝 𝐨𝐧 𝐭𝐨 𝐬𝐞𝐞 𝐰𝐡𝐚𝐭’𝐬 𝐜𝐨𝐦𝐢𝐧𝐠 in the next few months
𝟐/ 𝐀𝐈 𝐀𝐋𝐈𝐆𝐍𝐄𝐃 𝐓𝐎 𝐆𝐀𝐌𝐄𝐑𝐒 𝐀𝐍𝐃 𝐎𝐖𝐍𝐄𝐃 𝐁𝐘 𝐆𝐀𝐌𝐄𝐑𝐒
Sentient’s AI models are already known for their ability to align with specific values—whether it’s pro-freedom, pro-crypto, or pro-justice. Now, we’re taking it a step further by aligning Dobby with pro-gamer values.
This collaborative model will be fine-tuned for the unique needs of gamers and ownership will be distributed to consumers/gamers in the KGeN community.
𝟑/ 𝐀𝐍 𝐀𝐈 𝐒𝐘𝐒𝐓𝐄𝐌 𝐅𝐎𝐑 𝐃𝐄𝐂𝐄𝐍𝐓𝐑𝐀𝐋𝐈𝐙𝐄𝐃 𝐆𝐀𝐌𝐈𝐍𝐆
KGeN’s POG-E (Proof of Gamer Engine) allows gamers to retain control over their gaming data and ensures that game developers and platforms can access authentic, decentralized player data. With ~300M data points from 13M+ gamers, KGeN has amassed a phenomenal amount of data that can be utilized to improve the gaming experience.
This new partnership model can enable a variety of agents and use-cases such as:
🎮 Creating 𝐢𝐦𝐦𝐞𝐫𝐬𝐢𝐯𝐞 𝐚𝐧𝐝 𝐩𝐞𝐫𝐬𝐨𝐧𝐚𝐥𝐢𝐳𝐞𝐝 𝐠𝐚𝐦𝐞 𝐞𝐱𝐩𝐞𝐫𝐢𝐞𝐧𝐜𝐞𝐬 by suggesting relevant in-game actions, rewards, and challenges based on players’ unique gaming history.
🤖 Agents for varied use cases such as 𝐭𝐫𝐚𝐝𝐢𝐧𝐠 𝐚𝐠𝐞𝐧𝐭𝐬 𝐟𝐨𝐫 𝐠𝐚𝐦𝐞/𝐀𝐈 𝐭𝐨𝐤𝐞𝐧𝐬 or 𝐩𝐞𝐫𝐬𝐨𝐧𝐚𝐥 𝐚𝐠𝐞𝐧𝐭𝐬 𝐨𝐟𝐟𝐞𝐫𝐢𝐧𝐠 𝐠𝐚𝐦𝐢𝐧𝐠 𝐬𝐮𝐩𝐩𝐨𝐫𝐭 𝐚𝐧𝐝 𝐜𝐨𝐦𝐩𝐚𝐧𝐢𝐨𝐧𝐬𝐡𝐢𝐩.
💬 𝐒𝐨𝐜𝐢𝐚𝐥 𝐦𝐞𝐝𝐢𝐚 𝐬𝐜𝐨𝐮𝐭𝐬 to distill gaming knowledge from social media and KOLs.
Dobby will be a revolutionary step for decentralized community governance as we partner with @eigenlayer to help automate the adjudication of complex disputes or claims, especially those centering on intersubjectivity.
Judge Dobby will be the first software embodiment of subjective decisions through training on community discussions, corresponding governance, and corporations’ legal hearings.
Our first Dobby models were focused around pro-crypto and pro-freedom alignments. Now 𝐰𝐞'𝐫𝐞 𝐜𝐫𝐞𝐚𝐭𝐢𝐧𝐠 𝐚 𝐩𝐫𝐨-𝐠𝐨𝐯𝐞𝐫𝐧𝐚𝐧𝐜𝐞 𝐚𝐧𝐝 𝐩𝐫𝐨-𝐄𝐢𝐠𝐞𝐧 𝐦𝐨𝐝𝐞𝐥, showcasing how our AI team is able to granularly align models to specific goals and values for the EigenLayer community while maintaining performance and safety.
EigenLayer's decentralized protocol allows validators to restake ETH, EIGEN, and ERC-20 assets to secure multiple services at once. With this model collaboration, the EigenLayer community and protocol will have an adjudication layer that caters to the frontiers of intersubjective disputes and claims.
This model can help adjudicate or provide insights for:
In the coming days we will release Dobby, the first ever Loyal AI model. Participate in our first ever 𝐅𝐢𝐧𝐠𝐞𝐫𝐩𝐫𝐢𝐧𝐭𝐢𝐧𝐠 𝐂𝐚𝐦𝐩𝐚𝐢𝐠𝐧 to earn ownership of Dobby.
𝐹𝑖𝑛𝑔𝑒𝑟𝑝𝑟𝑖𝑛𝑡𝑖𝑛𝑔 is a novel training technique developed by Sentient’s top researchers. Fingerprint queries that are statistically indistinguishable from regular queries and that are resistant to Monte Carlo attacks are hidden in the models so that they cannot be removed by fine-tuning or revealed through prompt engineering. With them communities can control how their models are used and ensure that usage is aligned with their interests and principles.