Autonomys Community Lead Fradique was in attendance at Blockchain Rio in Brazil last month, an event which spotlighted Latin America’s significant growth potential as its web3 userbase expands in response to crypto-friendly administrations and regulation from Brazil to Argentina. Roberto Campos Neto, Governor of the Central Bank of Brazil, proposed tokenizing the Digital Brazilian Real (Drex) in his opening speech, positioning it as a “concrete bridge to DeFi.” This marked Campos Neto’s first participation in a conference dedicated to cryptocurrency and blockchain tech. He expressed his ultimate goal of developing an AI-powered super app that would enable consumers to manage all their financial accounts with real-time access and connectivity. Fradique networked with a number of projects at the event, some of which we’ve outlined below.

Polkadot x Autonomys

As the Autonomys Network was built using the Substrate SDK, we are eligible for marketing, business development and developer acquisition support from Polkadot and its ecosystem partners.

Polkadot is actively supporting its Substrate-based blockchains, including Origin Trail, Moonbeam, and Phala Network, in establishing partnerships and business deals across Latin America, and is expanding this ecosystem growth offering to Autonomys. This support involves several strategies:

• Business Development: Autonomys provides Polkadot with insights into our target markets and regional use cases, and Polkadot’s business developers and ambassadors familiarize themselves with the project through pitch decks on the Polkadot Academy. They then map out marketplaces and pathways to address them, and attend regional events to identify potential demands and partnerships.
• Developer Support: Polkadot encourages developers to engage with Substrate-based blockchains by building relationships with universities and the academic sector, and offering small grants to trusted university lecturers to promote their adoption within the academic community. This sustainable spiral workflow approach fosters utility understanding and network activity growth over time, and ultimately connects with developers, technical audiences, and decision-makers, aligning the project’s goals with real-world use cases. Polkadot also assists with translation and localization of documentation.
• Government Grants: Polkadot helps to facilitate partnerships with Brazilian startups, and assists them in submitting government grant requests. These grants receive acceleration through pre-approved government funds, providing Brazilian startups with 3–4x the amount spent by the sponsor (i.e. Autonomys).

This support extends beyond Brazil, with Polkadot maintaining a strong presence in Argentina and Mexico, and in the process of finalizing an agreement with the government of Paraguay. Autonomys’s maturity and AI direction offers a unique advantage, Polkadot concurs. By entering the Latin American market with their assistance, Autonomys will be able to position itself as a pioneer rather than just another competitor in a crowded space.

Mercado Bitcoin x Autonomys

Fradique introduced the team from Mercado Bitcoin — Latin America’s largest crypto exchange — to Autonomys, and in the days following, André Franco, Head of Research at Mercado Bitcoin, successfully installed an Autonomys Network farmer on his personal Mac M1. On August 28th, Fradique is scheduled to feature on Mercado Bitcoin’s Francamente Crypto podcast hosted by André Franco and Rony Szuster (Crypto Specialist at Mercado Bitcoin). We can’t wait!

Bonus Photos

About Autonomys Network

The Autonomys Network is a highly scalable decentralized AI (deAI) infrastructure stack encompassing secure distributed storage and compute, decentralized consensus, and suite of tooling for the development and deployment of verifiable super dApps and dAgents. The first primitive being built in the Auto Suite is our decentralized identity protocol Autonomys ID (Auto ID).

X | LinkedIn | Discord | Telegram | Blog | Docs | Github | Forum | YouTube

‍

The massive decentralized AI (deAI) presence at EthCC[7] made for a hectic schedule for Autonomys. Over four days in Brussels, the team was split between the main event and at least nine side-events on the topic of web3 x AI, frantically scootering between them when they clashed. Our goal was to learn about and speak to as many projects in the deAI space as possible. Here are some of our favorite recorded talks and panels on dAgents and AI3.0 at EthCC and its side-events:

Integrating Multi-Agent Coordination Networks

Advait (Leo) Jayant, SuperSight

Explores the integration of intelligent multi-agent coordination networks within the crypto ecosystem, focusing on their application in protocol design, research, governance, and investment strategies.

https://ethcc.io/archive/Integrating-Multi-Agent-Coordination-Networks

Build Your Next Web3 App with AI

Illia Polosukhin, NEAR

Illia shares NEAR’s vision for a user-owned, open-source AI x web3 stack.

https://ethcc.io/archive/Build-Your-Next-Web3-App-with-AI

Decentralized AI as a tool for blockchain apps

Moshe Malawach, LibertAI

Examines how to process data from smart contracts, feed them to a LLM, and push results on-chain as needed, in a decentralized way.

https://ethcc.io/archive/Decentralized-AI-as-a-tool-for-blockchain-apps

Acting on-chain with AI Agents

Fran Algaba, Giza

Explores wallet-enabled, on-chain agents that think off-chain with immutable and cryptographic guarantees of their activity.

https://ethcc.io/archive/Acting-on-chain-with-AI-Agents

Autonomous Agent Economies: Coordinating Multi-Agent Systems with Crypto

David Minarsch, Valory (Olas) | Don Gossen, Nevermined | Jarrod Barnes, NEAR | Levi Rybalov, CoopHive | Juri Stricker, Signature

This panel at Agents Unleashed: Supercharging AI Agent Economies featured some great discussions between people from across the deAI space.

BONUS: RenAIssance: Real Not-bluffing AI Summit

Watch the FULL day of interesting talks at FLock and Akash’s side-event. Check the schedule here.

Thanks to co-hosts DCG, Space and Time, Masa, Theoriq, Agentcoin & Vana.

‍

The day has finally arrived. After weeks of anticipation, we’re launching Stake Wars 2, the second chapter in our epic journey of incentivized staking. Before embarking on this grand adventure, let’s revisit the magic of our decoupled execution environments (domains), recap Gemini 3, and talk about how you can get involved.

⚡ A Lightning Refresher on Domains & Gemini 3

Farmers fortify the network’s consensus with their pledged space to create a strong defense against bad actors. This robust consensus offers support for domains, ensuring seamless transaction ordering. Operators stake their tokens to provide the economic security needed to maintain the integrity of the blockchain, while fraud proofs stand ready to punish any who act in bad faith. Think of domains as mystical realms sharing the consensus provided by farmers. Spinning up a new domain is as easy as executing a transaction, making it perfect for projects that need unique execution environments, but that lack the infrastructure to secure their own network.

Our current incentivized testnet Gemini 3 has over 7,000 concurrent nodes pledging >140 petabytes (PB) of storage in exchange for testnet Subspace Credits (tSSC) from reward pools. Nodes, data capacity and rewards continue to grow steadily by the day. Participants using modest consumer hardware are able to farm tSSC, which will then auto-convert proportionally into mainnet tokens during the Token Generation Event (TGE) organized by the Subspace Foundation. Be on the lookout for more information from the Foundation in the coming weeks.

It’s not too late to join the incentivized testnet as a farmer! Check out the official docs to learn how you can start farming and jump into Discord if you have any issues.

🌐 Stake Wars (2!): The Multi-Domain Saga Begins

Our vibrant community on our Discord and forum participated in our resoundingly successful Stake Wars initiative in January, which aimed to test the Nova domain — our EVM runtime environment — and our Proof-of-Stake (PoS) model. 320 operators staked >536,000 tSSC to provide execution, while 1,054 nominators trusted those operators with >318,000 tSSC. The following month, we held our Mini Space Race event as a dress rehearsal for mainnet launch. When the community smashed our target of plotting 8PB — enough space to secure a test network launch — in just six days, we were simply blown away. The team expected that it would take more than double that time!

In this new saga, we unveil an additional domain: Auto ID. This new battle arena, forged in the fires of innovation, supports identity at Internet-scale and beyond. Stake Wars 2 invites operators to test these domains and execute transactions while earning fees. Testnet rewards have again been allocated for this event, while we’re planning a few tweaks to the original competition to help us test our updated staking processes and interfaces. Anyone who holds tSSC can participate as a nominator without any hardware requirements. As warriors, you must stay vigilant. Track the ebb and flow of transactions, setting up where opportunities arise, but beware — the landscape will shift with the tides of Hacker Month 2.

Read the docs and introduce yourself in the #operators or #nominators channel on Discord if you want to get involved. If you represent an experienced validator outfit, please get in touch with us for participation and co-marketing opportunities.

📅 The Phases of Our Quest

Our journey through Stake Wars 2 will traverse the mystical land of Gemini 3 over multiple weeks, divided into three epic phases:

🛠️ Phase 0 — Prelude to the Saga

In the calm before the storm, the team will deploy and test the Auto ID domain, adding final touches to the staking interfaces hosted on Astral. The call to arms has sounded with a minStake set at 100 tSSC for both Nova and Auto ID. Brave warriors, both old and new, are invited to claim their stake (if they haven’t accumulated 100 tSSC already) and prepare for the trials ahead. To guard against false-hearted infiltrators, only those with a Verified Farmer role on Discord are able to take advantage of the minStake claim mechanism.

🏰 Phase 1 — The Quest for Dominion

With the start block announced, the true adventure begins! Maintain your strongholds, upgrade swiftly, and seek the favor of nominators. Our engineering wizards have removed the limits on nominators, giving every warrior a chance to prove their mettle.

🧪 Phase 2 — The Trials of Innovation

Side quests await those daring enough to test the latest advancements. Tasks will range from stress-testing unlimited nominator counts to probing the depths of cross-domain messaging. Each completed trial brings rewards and helps us prepare for the ultimate showdown: the mainnet beta.

🎊 The Grand Celebration

When our journey ends, we will gather for a grand celebration, recounting tales of bravery and innovation, and awarding the most valiant among us. The final block will be announced, and we’ll reflect on lessons learned as we forge ahead towards mainnet launch.

💰 Rewards for the Brave

The treasure chest for Stake Wars 2 holds 0.3% of the total token supply:

• 0.1% will go to operators based on transaction fees earned
• 0.1% will go to nominators based on transaction fees earned
• 0.1% will go to those who complete Phase 2 trials

Plus, there’s USDC prizes for developers who deploy dApps during Hacker Month 2! A further announcement about this is forthcoming.

📌 Key Things to Remember on Your Journey

• Sync your nodes from scratch, warriors, as the road ahead stretches long.
• Choose your domain wisely, or brave multiple paths if you can muster the minStake for each.
• Seek the minStake airdrop claim if you’re starting from zero, but hurry — time flees quickly.
• Keep your domain alive: register only when ready and deregister if you falter.

Stake Wars 2 transcends a mere contest; it will unfold as a saga of competition, innovation and camaraderie. We call upon all community members to join us, contribute to the cause, and celebrate the evolution of our network.

Onward, Stake Warriors! 🌟

‍

‍

Want to farm or build on the Autonomys Network?

Check out the Autonomys Academy, join the Discord & come say HAI!

‍

‍

Terms & Conditions

Exclusion of US Persons

Tokens converted during the TGE (the “Tokens”) are not intended for distribution or resale by or to any person or entity that is a citizen or resident of, or located in, the United States of America, including its territories and possessions, any state of the United States, and the District of Columbia (a “US Person”).

Excluded Jurisdictions

In addition to the United States, the following jurisdictions are excluded from participating in the TGE: the Democratic People’s Republic of Korea; Crimea, Sevastopol, Donetsk and Luhansk regions of Ukraine; Cuba; Iran; Syria; and any other jurisdiction in which accessing the TGE, related websites, the Token or protocol platforms is prohibited.

‍

AI’s rapid advance and growing integration into our lives represents not only a technological revolution, but a social one in the nature of the human-machine relationship. We must not lose sight of the importance of human autonomy. For us to navigate towards a future optimized towards humanity, we must embrace a new paradigm of radical autonomy, or absolute digital self-governance. At Autonomys, we seek to inaugurate the Age of Autonomy — where radical autonomy is our guiding principle.

Our Vision

Autonomys’ mission is to create the decentralized technological infrastructure necessary to support the future of human and artificial intelligence (H+AI) collaboration. We are developing a suite of tools, primitives and dApps atop our unique deAI ecosystem stack — the Autonomys Network — to provide all humans with secure, transparent and equitable access to AI resources. The Autonomys Network is ideal for this task as it solves the blockchain trilemma through our innovative Proof-of-Archival-Storage (PoAS) consensus mechanism.

Our first product, Auto ID, is a composable digital identity standard open to every human on the planet. Auto ID allows for the creation of secure, verifiable identities. AI models and autonomous agents are linked with their human creators’ Auto IDs, ensuring that the outputs and actions of these systems are easily identifiable and accountable, establishing provenance and authenticity of both data and deed as standard. In this way, Auto ID allows for the authentication of online content, and builds greater trust in digital interactions.

Autonomys’ vision transcends our technical architecture. We’re working to create an ecosystem that supports ethical interactions between humans and AI, fosters individual and collective growth, and lays the groundwork for a future in which technology enhances human autonomy. We’re proselytizing a philosophy of radical autonomy — absolute digital self-governance — where reliance on the passive reception of handouts is rendered obsolete. Instead of creating artificial dependencies, we are focused on fostering a self-sustaining ecosystem through incentivized participation and contribution, inspired by Ethereum’s pioneering model.

Radical Autonomy

/ ˈrædɪkəl ɔˈtɑnəmi /

adj. + n.

1. The capacity of individuals or organizations to exercise absolute control over their digital identity, data and interactions.
2. The freedom to make informed, uncoerced decisions about one’s digital presence, privacy and security without undue influence from external entities.

Radical autonomy encourages individuals and communities to:

• Imagine new personal and societal possibilities
• Explore more dynamic, responsive and customized futures
• Innovate, create and evolve according to their own vision, values and ethics.

The Age of Autonomy

Radical autonomy is the foundational principle of the Age of Autonomy, and Autonomys is architecting the transition to this new age. We envision AI not as a tool for increasing economic productivity in a way that replaces human labor, but one for augmenting human potential, enhancing human autonomy, and promoting human flourishing. We want to ensure that humans retain control over their digital existence, while fostering trust, transparency and ethical interaction in the digital realm.

In contrast to predictions of a Universal Basic Income (UBI)-dependent underclass in the Age of A(G)I, AI in the Age of Autonomy catalyzes individual and collective growth, enabling us to build a society based on the principles of self-determination, self-sufficiency and self-governance. We have incredibly exciting plans in this regard — stay tuned. By embracing radical autonomy as our guiding principle, we can build a world in which technology amplifies rather than diminishes our humanity.

The Internet of dAgents (IodA)

Autonomys is dedicated to building the infrastructure needed to support the Age of Autonomy — what Google’s Laurence Moroney calls AI version 3 and others call AI3.0 — decentralized, open and collaborative model and agent development and deployment. Composability between Autonomys and other pioneering deAI projects — the likes of Bittensor, OLAS and Morpheus — will be crucial.

v1: Deep learning becomes widespread as developers are able to build models with the likes of TensorFlow and PyTorch.
‍
v2: Emergence of large language models (LLMs) such as ChatGPT, and then GenAI.
‍
v3: web3-enabled AI ensures decentralization, security and transparency.
‍
Moroney’s typology of AI versions

The Internet of Agents (IoA) — where every human has many powerful, automatically interoperable, personally aligned models at their disposal — is the culmination of this paradigm. According to Davide Crapis, Research Scientist at the Ethereum Foundation, the three main steps on the path towards the IoA are: super dApps (scalable web3 smart contracts + AI), new agents’ mechanisms, and ownership and governance. The Autonomys Network and Auto ID provide the infrastructure to achieve all three in a decentralized way on-chain, hence we use the terms dAgents and Internet of dAgents (IodA). Below is a basic roadmap of Autonomys’ vision (inspired by Davide Crapis’ treasure map analogy):

Auto ID (SSI infrastructure) → AI Data Provenance → Verifiable Super dApps (trustable AI-powered dApps) → Verifiable dAgents (on-chain agents) → Internet of dAgents (IodA) → Age of Autonomy

Autonomys is creating an ecosystem that empowers both humans and dAgents to interact, transact and collaborate in ways that are secure, transparent and equitable. As we embark on this transformative journey, we do so with a steadfast commitment to the principle of radical autonomy, and a shared vision of a future characterized by mutual cooperation and flourishing between humans and AI. Together, we have the ability to shape the future and usher in the Age of Autonomy, a new era of unprecedented freedom, prosperity, and progress for all.

‍

‍

Want to farm or build on the Autonomys Network?
Check out the Autonomys Academy, join the Discord & come say HAI!

‍

Consensus 2024 provided bountiful opportunities for learning about the burgeoning web3 x AI space. The Autonomys team attended numerous insightful talks, panels and side-events on this theme, cultivating our network of potential DePIN and deAI ecosystem partners in the process. Read highlights from some of our favorite talks on ethical decentralized AI (deAI), decentralized physical infrastructure networks (DePIN), and self-sovereign identities (SSIs) below.

Decentralized Identities: Building Online Security in a Privacy-Indifferent World

w/ Irene Hernandez, CEO, Gataca

Hernandez spoke about the importance of decentralized digital identities, or self-sovereign IDs (SSIs), and having exclusive control over your personal information.

A privacy-centric world that lacks proper identity controls is just as dangerous as a highly transparent world protected from identity fraud. SSIs balance decentralization, privacy and security, and are thus the next generation architecture for identity management on the Internet.

Verifiable credentials have built-in proof of authenticity and ownership, massively reducing operational costs. Instead of having multiple siloed identities for individual services, we will have a single identity across all our digital services, with the ability to grant and revoke access to specific personal information easily.

SSIs are coming very soon. “It is time that we shift our collective indifference towards privacy to a collective effort in building an improved version of the Internet where privacy is not just a nice idea, but the cornerstone that sustains it all.”

Decentralized Minds: The AI + Blockchain Revolution

w/ Dr. Ben Fielding, Co-Founder, Gensyn | Greg Osuri, Founder, Akash Network | Jiaho Sun, Co-Founder, Flock | Tommy Eastman, AI/ML Engineer, Foundry

Representatives from Gensyn, Akash Network, FLock, and Foundry spoke about the state of deAI and DePIN and the future of machine learning (ML).

• FLock — federated learning on-chain for private and transparent deAI model training
• Gensyn — machine learning compute protocol connecting CPUs/GPUs with ML model trainers
• Akash Network — decentralized super cloud network granting permissionless access to unavailable high-density GPUs (NVIDIA H100s and A100s) peer-to-peer

Jiaho Sun highlighted the importance of GPU drivers and data privacy in deAI agent development. Decentralized compute DePIN need drivers to ensure distributed GPUs can collaborate on model training effectively and efficiently — NVIDIA’s CUDA software currently dominates. If you want to train an effective personal assistant AI agent, you must provide it with all your personal data (which you don’t want to be giving to Big Tech companies). Sun suggested deAI Agents Summer is thus on the horizon.

Dr. Ben Fielding offered two solutions for solving latency and bandwidth issues in decentralized ML training:

• Short term — optimizing existing models by accounting for heterogeneous devices.
• Long-term — new model architectures — moving from centralized AGI training to highly sharded, expert models.

Fielding also prophesied richer, bespoke interactions based on probabilistic AI replacing traditional app and website interfaces. However, he conceded we need much more decentralized training and fine tuning infrastructure for that future to exist, because currently there’s nowhere near enough. He foresees the future of ML as networks of mini models distributed across devices that compose to form larger models when needed. Inference queries will be paths through these mini models, similar to Internet routing protocols. An open-source compiler stack for ML training would connect high-level AI frameworks with different devices as more chips are designed.

There is lots of ongoing research into techniques for data privacy in ML training and inference, including federated learning, homomorphic and functional encryption, and differential privacy. The number one problem Gensyn is trying to solve is verification of remote ML tasks, as ZK-proofs are too expensive to verify computation. General auditing (rerunning portions of the task) is currently the most viable verification mechanism, but requires reproducible or deterministic execution via cryptographic, game theoretic or probabilistic proofs, as different devices produce different results.

According to Greg Osuri, the GPU supply chain is extremely broken today. It’s impossible to get enough high-density GPUs to do any meaningful training or inference at scale, as demand from the Big Tech quintet (Alphabet, Amazon, Apple, Meta, Microsoft) currently strangles the supply of NVIDIA GPUs. The complexity of the global supply chain has meant supply has failed to keep pace with ever-increasing demand, and creating chip fabricators requires lots of time and investment. However, there’s billions invested globally into compute for AI and more specialized GPUs for ML. Although AMD and Intel GPU driver software is unstable right now, NVIDIA’s dominance will wane within 3–5 years, Osuri suggests.

There are now lots of chips available as ML projects have upgraded to different chipsets, and crypto companies, such as Foundry, have moved away from PoW mining as it became less profitable. Many now supply Akash, making it the only way of accessing H100s on-demand. Akash powers numerous apps and dApps, including Brave, FLock and Venice.ai. Akash is able to reduce costs by 50–90% by tapping into this underutilized compute capacity and decoupling the resource from its control. As Foundry’s Tommy Eastman points out, “This represents a unique opportunity to onboard people into crypto as there’s no other options for them to access these GPUs. If we make the experience good, they will stay.”

Osuri believes the narrative is changing around crypto and AI, citing this October 2023 article in Semafor. As he asserts, “Open-source, decentralized systems are significantly better at guaranteeing data privacy than closed, centralized ones because of their transparency.”

Time to Unite the Tribes: Building Truly Open AI

w/ David Johnston, Open-Source Contributor, Morpheus | Erik Voorhees, Founder, Venice.ai | Misha da Vinci, Founder, GRFTF podcast

David Johnston, from Morpheus and Erik Voorhees, from Venice.ai, spoke about smart agents and the goals of building open AI.

• Morpheus — community building a permissionless network for decentralized, open-source AI agents by connecting APIs with smart contracts and incentivizing people to contribute compute, open-source code, capital and agents
• Venice.ai — private, permissionless chatbot with uncensored answers, built on Morpheus

A reaction to regulatory capture, Morpheus has no founder, company or foundation — just a whitepaper, written by pseudonymous co-authors Morpheus, Trinity and Neo. However, it also now has 255 contributing members and $500 million in ETH staked towards the project, contributing yield to bootstrap it into existence. Johnston claims AI experts are already saying it’s really useful for cheap compute.

In Johnston’s words, “Ethereum pioneered smart contracts. Morpheus is pioneering smart agents.” Built using open-source LLMs (Llama 3, Mistral), which caught up with closed LLMs (ChatGPT) last year, Morpheus agents can run on personal hardware, lowering the cost of access to AI for everyone.

Agents need web3 for economic action. The future is wallet-connected personal agents that you own and control, Johnston claims. AI-powered interfaces will be the Internet search engine moment of the mid-90s for web3, improving UX, matching intent with the best results, and helping onboard billions without technical knowledge, truly bringing web3 to the mass market. For example, you tell MetaMask to send 1 eth to Bob. An agent finds the most efficient path before allowing you to approve the transaction with a single button.

Voorhees’s example is more complex, giving the user the ability to interact with a financial intelligence: You tell your agent to put your USDC (on MetaMask) into the highest yielding stablecoin. The agent responds that it’s XYZ with a rate of 32.3%. However, it has a market cap of $50 million. The second-highest is ABC with a market cap of $10 billion. Do you want to buy it? Tap yes and it buys it.

“You don’t want all that stuff going through a centralized company. That should be obvious to anyone in crypto,” Voorhees suggests. Crypto’s permissionless composability will transform AI. “If you can incentivize a decentralized compute function, everything else AI can organically build on top.” deAI will grow like DeFi did despite established players, because it is open, permissionless and composable. We are building systems of interaction and exchange which are easier and more frictionless, which will draw developers and users.

Venice pulls inference data from open-source models run on decentralized GPUs via encrypted proxy servers. Data is otherwise stored locally. Why use Venice? Everything you’ve ever sent to ChatGPT goes to a centralized company (and any other entity, government or hacker they give, sell or leak your unencrypted data to) tied to your identity forever. If you don’t want to be spied on and don’t want censored answers, use Venice.

Don’t Trust, Verify: An Overview of Decentralized AI Inference

w/ Haseeb Qureshi, Managing Partner, Dragonfly

Dragonfly Capital’s Haseeb Qureshi spoke about decentralized inference and the state of the deAI industry: “Crypto is a technology that creates trust. AI is a technology that needs trust.” web3 x AI is big, but right now it’s early and hard. The next generation of AI protocols needs to be cheap, easy and fast for mass adoption, but needs to solve core scalability issues first.

The current deAI stack:

Decentralized GPU compute networks, including Render, Akash and io.net are the largest in web3 x AI. Decentralized training protocols include Bittensor (the most famous) and Gensyn (in the research stage). Decentralized inference (running the model to ensure its returns the right output) is being worked on by Ritual (the most famous) and Modulus (doing zero-knowledge inference). AI inference is very computationally intensive. It is virtually impossible to do directly on a decentralized blockchain. Lastly, there are AI-powered web3 apps, including Kaito (market intelligence), MyShell (character chatbots), Alethea (interactive AI companions).

Approaches to verifying LLM computation on-chain:

• zkML (zk-proofs of off-chain computation): EZKL, Giza, Modulus
• cryptoeconomic ML (PoS/token voting for decentralized trust): Ritual, Atoma Network
• optimistic ML (optimistic roll-ups for fraud proofs): Gensyn, ORA

These are not new ideas — they map almost perfectly onto how we think about blockchain security:

ML is ultimately a type of computation, but it has important differences with computation normally on a blockchain. A key challenge is the fact ML was not designed to be proven. This matters for zk-proofs as they won’t compute with even minute differences caused by:

• non-determinism (different GPUs/CUDA versions can lead to different results even with the same model).
• floating points (decimals can be computed differently depending on how the computation is parallelized, meaning even with the same GPU/CUDA version, you can get a different output).

There is a fourth way of verifying deAI computation: using Trusted Execution Environments (TEEs) — special enclaves for storing private information with hardware-level security. The newest generation of NVIDIA chips have these TEEs that can guarantee privacy and verifiability of computation way faster than any other form of verifiability. However, you have to trust NVIDIA (which everyone ultimately has to do if they work in AI). Qureshi says the market will decide how to answer this question.

At present, few decentralized cloud projects have a good privacy solution for sensitive data, but this also applies to managed cloud providers. Data is the single hardest thing for deAI players, Qureshi claims. Although data marketplaces can aggregate and auction data, licensing agreement enforcement on data is very difficult. Vana is building Reddit Data DAOs to allow users to sell their collective user-generated data for AI training (Reddit data is very valuable for this), and undercut Reddit in the process (Reddit licenses their site data for $50m a year — none going to users).

How to (Ethically) Build Human AI

w/ Teana Baker-Taylor, COO, Venice.ai | Dr. Ben Goertzel, CEO, SingularityNET Foundation | Arif Khan, CEO, Alethea AI | Jeff Wilser, Host, AI-Curious podcast

Teana Baker-Taylor (Venice.ai), Ben Goertzel (SingularityNET) and Arif Khan (Alethea) spoke about the opportunities and challenges of deAI.

Crypto’s ‘killer app’ could be AI smart agents, Baker-Taylor suggested. Tokenization is very useful in a decentralized, digital-native ecosystem as it allows for efficient incentivization and automatic financial transactions between agents via smart contracts.

She offered this as an example: I type in natural language my instructions for my agent (permissioned by me) into a generative interface. That agent works with other agents to complete tasks and automatically settles with them via a tokenized currency and smart contracts.

However, Taylor warned, “This is uncharted territory. Unlike cryptocurrencies, we don’t have an existing relationship with AI in the same way we did with money. The adoption cycle for these technologies could be faster, meaning we have less time to figure out what good looks like. We really need to think about our privacy in a different way” because the more we give to AI, the more open we leave ourselves to manipulation. We need to translate crypto’s notion of financial sovereignty into an AI context.

Dr. Ben Goertzel (who popularized the term AGI) said he tried doing decentralized AI in the 2000s and it wasn’t possible. Ethereum’s revolutionary smart contract ecosystem inspired and enabled SingularityNET’s founding in 2017 with the goal of building infrastructure to run various AI systems across large numbers of decentralized machines.

deAI tools still work a lot worse than Big Tech AI tools, Goertzel claimed. We need to fix that because “the amount of people who will use a decentralized AI because it’s decentralized is not enough”. We need to build systems that are smarter, more powerful and easier to use. Near parity between open and closed-source LLMs isn’t enough. The deAI community has tremendous size and diversity on which to draw from on its side. Goertzel is working on open-source AGI architecture project OpenCog Hyperon which brings together neural, symbolic and evolutionary AI in an integrated cognitive architecture built on decentralized infrastructure.

“Although we’re not there yet, when you take seriously our aspiration to make machines that can think as creatively and originally as people, and not be constrained by their programming or data, it becomes clear that who owns and controls that system is a big deal. Ultimately, the answer is going to be they will own and control themselves,” Goertzel said.

Alethea allows people to tokenize their generative AI models, characters, agents, datasets etc. via the The AI Protocol. Its key application is ALIagents which allows you to create emotive, human-like characters that can be interacted with in real-time. Khan posits the main challenge in building deAI at the moment is the coordination of capital, compute and intelligence. He suggests this is partly because it’s incredibly difficult to step away from existing structures with strong network effects — whether it’s a capital-based (VCs), compute-based (cloud computing) or intelligence-based (post-ChatGPT).

To create deAI agents, Khan claims, you need to solve 3 the Cs:

• Coordinated incentive layer
• Compute layer (truly decentralized so agents are uncensored, uncensorable & permissionless)
• Capital attribution (agents need capital to deploy & bid for compute).

Blockchain tech is the perfect solution, particularly when you integrate primitives like EigenLayer, which uses Ethereum’s security and yield to provide opportunities for developers to build applications.

‍

‍

Thanks for reading this summary of the best talks and panels on decentralized AI from Consensus 2024. We hope you found this crash-course on one of the most exciting emerging technological trends informative and insightful. We’re looking forward to further building on our deAI knowledge and developing our partnership network at future industry events, including EthCC[7] (and its many side-events) in Brussels on 8–11 July!

Bonus Consensus 2024 Autonomys Team Pics 👀

‍

Autonomys is crafting a token issuance model that will allow our network to thrive for decades to come. We’ve invested a great deal of resources into diligently conducting a process of mainnet modeling and issuance parameter selection to this end — in collaboration with leading complex systems engineering firm BlockScience. The original analysis provided by the BlockScience team is available here. We are sharing some of our research and development findings to offer greater clarity about the conclusions we reached.

Introduction

Terminology

• Fees: Payments for transactions on the network.
• Issuance: The amount of tokens minted as a reward per block, total for all recipients.
• Proposer: Farmer who won the block solution challenge.
• Rewards: Newly minted tokens issued by the protocol to farmers as compensation for their work.
• Voter: Farmer who won the vote solution challenge.

Background Reading

• For a high-level background on dynamic token issuance, read the relevant section in the Autonomys Academy and the forum post.
• For an extensive description of our technical implementation, see the Dynamic Issuance specification on Github (not required for most readers).

Goals

We designed the economic parameters of our protocol issuance to align with the broader objectives of the network:

1. Rational Economic Incentives: Ensure the economic parameters encourage behaviors that support the network’s long-term viability and growth.
2. Community Incentivization: Create incentives that encourage participation from all defined stakeholders (farmers, operators, nominators).
3. Distributional Equilibrium: Balance the issuance and distribution of tokens to support both the network’s scalability and the fair distribution of resources among participants.

Mainnet Issuance Values

Total ATC Supply: 1 billion

The total ATC supply cap will be 1 billion, as per our simulation. The exact allocations to the team, investors and others will be announced when we have the final numbers following the conclusion of the incentivized testnet and Stake Wars 2. Our implementation keeps track of the total issued tokens to date and ensures the total supply cap is never exceeded.

Space Race Duration: <7 days

The algorithm for the Token Generation Event (TGE) is largely the same as that for the Gemini-3h testnet described here (TL;DR: we will hold a Space Race that starts automatically at genesis block until a certain amount of pledged space is reached, at which point protocol rewards begin).

The results of the simulation suggest rewards are likely to start within 7 days of our mainnet launch. It did not estimate the amount of pledged space required. We will compute this in the final days of testnet based on the total space pledged, as well as the speed of growth (~1.3 PiB/day so far). This will all inform how we choose our Space Race target. The longer it runs, the longer it would take to achieve optimal community-ownership. However, a longer Space Race would slightly increase the lifetime of the protocol issuance.

NOTE: The release of our official GPU plotter (if ready before mainnet launch) will also likely impact our Space Race expectations.

Initial Reward: ~5 ATC/block

The first block after the Space Race target is reached will immediately start minting tokens to reward the proposer and voters. The value of the reward issued is indicative: if we set it to $x$ ATC, the real issued value may be higher or lower depending on utilization and the number of votes, but on average, it follows the curve below. This value, which we call reference subsidy, decreases from the next block following an exponential curve (explained further below).

This issuance curve uses example values consistent with overall recommendations, and as such, these may not be final. The illustrated value per block is the total rewards distributed among all recipients in that block (as described below). Greater rewards significantly aid the growth of the community-owned supply, but at the expense of runway and distributional equilibrium (as a smaller number of early farmers receive a lot more than later ones).

Proposer Share: 10% reward + 10% tax

Each block, on average, includes 9 votes, so we anticipated 10 recipients per block (this number may change once our sharding upgrade is implemented). Currently, the protocol issues 10% of the initial reward to the proposer, and 10% to each voter. In an empty block with exactly 9 votes, we would issue exactly 5 ATC; 5.5 ATC if there were 10 votes; 4.5 ATC with 8 votes, and so on. Assuming an initial reference subsidy of 5 ATC, the rewards for the first block would be issued as follows:

• Proposer: max. 0.5 ATC (reduced if block utilization > 0)
• Voters: 0.5 ATC each

The proposer also takes a 10% tax from each voter to disincentivize them from censoring votes. The final reward per block credited to the wallets would therefore be:

• Proposer: max. 0.5 ATC + (10% * 0.5 ATC * [# of voters])
• Voters: 90% * 0.5 ATC each

The proposer share % has not shown conclusive correlation with any of the simulation goals.

Slower Decline: ~1.5 years

To incentivize early adoption and community-owned supply growth, we have developed an issuance curve that begins with a period of slower decline, before accelerating at a certain inflection point. This is achieved by taking the sum of two exponential components: one where the decline starts immediately, and another where the decline starts after a while:

The plot shows the curve for ~15 years (in blocks), with an inflection point at 1.5 years. Although the simulation suggested the duration of this period of slower decline should be 1.5 years, it showed no strong correlation with our goals. A later inflection point would moderately benefit community-owned supply growth, but negatively impact the runway of token issuance.

Runway: >40 years

For each curve, we controlled the initial reference subsidy (the starting point on the y-axis). The rate of decline is effectively controlled by the amount of tokens we ‘allocate’ to each curve component. Taking our estimated supply on TGE and allocating 50% to each component, we can expect a 40+ year runway.

The model simulated the first 3 years after TGE, and thus this runway calculation is based on the functional form of the issuance curve and the initial reference subsidy. The component allocation value is the strongest estimate for runway: lower values make the decline faster initially, but significantly prolong the tail after the four-decade mark.

Long Tail Issuance

The component allocation cap is not enforced in our implementation. The only enforced cap is the total supply cap. The protocol will continue issuing small amounts (i.e. ~0.1 ATC/block, as the exponential function never reaches 0) until the total supply cap is reached with no outside intervention. This allows us to gracefully handle uncertainty over the duration of the incentivized testnet as no extra tokens will be left unissued, nor will the protocol exceed the 1B cap.

‍

‍

Want to farm or build on the Autonomys Network?

Check out the Autonomys Academy, join the Discord & come say HAI!

‍

‍

Thanks for joining us on this brief tour of Autonomys Network’s mainnet protocol issuance parameters. We hope that you enjoyed this breakdown of our process, and that it helps the community better understand how we arrived at these values. Keep a lookout for our forthcoming follow-up post on protocol economics.

NOTE: The Subspace Foundation reserves the right to adjust the Autonomys Network’s token issuance parameters based on further research and testing up until mainnet launch. The Foundation does not intend for ATC tokens to be available for sale or resale within the United States or to U.S. persons.

‍