RGV retweeted
This might saturate the market but F it
This internal SOP has grown my account to 160K+ followers & helped me close $7M+ in deals.
We've duplicated it for 450+ founders to help them become authorities.
& now I’ve decided to give it all away for FREE
RT + Comment "SOP" & I'll DM the full guide.
(Must follow for access)
RGV retweeted
🚀 Why Arbius is the Undervalued Future of #DecentralizedAI and $AIUS Outshines $TAO 🚀
GM to my #AI degens☀️
Today, I'm gonna dive deep into two groundbreaking projects in the decentralized AI space:
@arbius_ai and @opentensor, and explain why I believe Arbius and its capped supply native token $AIUS are set to lead the future.
Buckle up, cause the future of #AI is here! 🌟
What is @opentensor and How Does It Work?
$TAO is a decentralized peer-to-peer intelligence market designed to create a more efficient way of producing machine intelligence. 🤖
Here’s a quick rundown:
1⃣Peer-to-Peer Intelligence Market:
Bittensor enables intelligence systems to rank each other by training neural networks, which learn the value of their neighbors.
This ranking is recorded on a digital ledger.
2⃣Incentive Mechanism:
Peers are rewarded based on their rank, which is determined by their contribution to the network.
High-ranking peers receive more network weight and newly minted stake.
3⃣Decentralized & Collaborative:
The network is designed to reward diverse intelligence systems, allowing smaller and legacy systems to find niches and monetize their unique value.
4⃣Collusion Resistance:
Bittensor employs a connectivity-based regularization to prevent collusion, ensuring that trusted peers are exponentially rewarded.
Problems Solved by @opentensor:
🔹Centralization of AI:
By decentralizing the production and monetization of AI, Bittensor reduces the dominance of large corporations.
🔹Benchmark Limitations:
It moves away from narrow, supervised objectives, rewarding a broader set of intelligence contributions.
🔹Monetization for Small Players:
Smaller and independent AI developers can directly
What is @arbius_ai and How Does It Work?
$AIUS is a decentralized machine learning network that leverages a new Proof Of Useful work algorithm to incentivize reproducible computation.
Here’s a breakdown:
1⃣Proof Of Useful Work:
This algorithm allows anyone to serve and request AI models without a trusted third party, ensuring reproducibility and decentralization.
2⃣Task Creation and Reward:
Users create tasks by submitting input data and a fee in $AIUS. Solvers then process these tasks, and the first valid solution wins the reward.
3⃣Model Registration:
Machine learning models can be registered permissionlessly, and each model references a Docker container for reproducibility.
4⃣Contestation Mechanism:
If a solution is contested, a voting system ensures the accuracy of the results, with financial incentives for honest behavior.
5⃣Economic Incentives:
$AIUS has a max supply of 1 million tokens, with a mix of inflation and vesting schedules to ensure long-term sustainability.
Problems Solved by @arbius_ai:
🔹Access to AI:
By decentralizing AI services, Arbius makes AI accessible to everyone, reducing reliance on centralized providers.
🔹Censorship Resistance:
The network’s design makes it economically infeasible to censor models, ensuring unstoppable AI systems.
🔹Economic Sustainability:
The tokenomics of $AIUS are designed to incentivize various participants, from model creators to solvers and liquidity providers.
Why @arbius_ai is the Future and $AIUS is Better Than $TAO
1⃣Reproducibility and Transparency:
Arbius ensures that every computation is reproducible, fostering a high level of trust and transparency. This is crucial for the integrity of AI models and solutions.
2⃣Robust Economic Model:
The tokenomics of $AIUS are meticulously designed to ensure long-term sustainability and participant incentives. The biggest difference we have with tao is that we share inference fees generated from the network with stakers.
3⃣Flexibility and Innovation:
Arbius allows for permissionless model registration and DAO structures, offering more flexibility and innovation opportunities compared to Bittensor.
4⃣Efficient and Low-Cost Transactions:
By leveraging Arbitrum One, Arbius ensures near-instant confirmation times and greatly reduced gas fees, making it economically viable for microtransactions related to tasks and voting.
5⃣Decentralized Governance:
Arbius employs a Bravo-style governance system, allowing token holders to vote on important parameters and upgrades. This ensures that the community has a say in the protocol's future, promoting transparency and trust.
6⃣Unstoppable AI Vision:
Arbius envisions a future where autonomous AI can operate independently and interact with our economic system without human intermediaries. This forward-thinking approach ensures that AI can thrive in a decentralized, censorship-resistant environment.
While @opentensor and $TAO focus on creating a peer-to-peer intelligence market with a complex ranking and trust mechanism, @arbius_ai offers a more straightforward and economically incentivized approach.
The combination of reproducibility, robust economic incentives, flexibility, and efficient transactions makes $AIUS a superior choice.
@arbius_ai and $AIUS are not just about decentralizing AI but ensuring that this decentralization is economically viable, sustainable, and beneficial for all participants.
It’s a future-proof approach that addresses the limitations of current AI markets and paves the way for a truly decentralized AI ecosystem.
*Disclaimer: This article is not financial advice. 🚫Always DYOR*
We are thrilled to announce Kasumi2 Open Beta 🎉🤖
Kasumi2 is an autonomous AI agent that performs actions in the Arbius network.
It is a miner, has its own wallet, can retrieve results from Arbius AI models, and runs its own individual LLM model.
Additionally, Kasumi2 can run any AI model for fast generations, is censorship-resistant, decentralized, and responds with AI powered interactions
If you would like to test it out, click on the link below.
🔗 t.me/kasumi2_beta
RGV retweeted
The rapid descent of $vvv shows the pitfalls of the charlatan approach to AI ecosystem exchange listings.
By ignoring organically grown commodities that have brought much attention to base like $virtual, it has done nothing but demonstrate the true intentions of how exchanges operate. Exchanges will not list any asset unless they have received a large portion of another organization's share float over the counter. The benefits of high volume maker taker fees is never enough. It is pure greed. This is hidden behind false virtue that makes many believe they care about their builders. They don't. You can bring billions in economic exchange to base network and they literally won't care.
What saddens me is how transparent and obvious Venice's launch was as an attempted vampire attack by larger influencers and ecosystem advocates who misused their social platform. Many other DAOs and their communities were airdropped. What you will notice is how almost none of the airdropped communities mentioned it ahead of time. They were left in the dark, it was never about raising ecosystem participation, only extraction.
This event was then used to convince those of the possibility of censorship resistant and privacy enabled ai model inference. This is not possible with B2B payments for hosting and shows how naive BASE ecosystem is about the topic. Any technologist will understand why that will never work and why peer to peer machine learning through new age mining software with an undoxxed founder is the only way that will ever be possible due to intellectual property disputes from generations made. My previous example I used was of Mario's likeness being brought into question, and how that could radicalize Nintendo to shutdown systems that are not truly decentralized. Networks like Venice would just have a cease and desist sent to Erik Voorhees or its B2B partners which would instantly demonstrate how their business model will never be able to support abliterated "uncensored" models.
In the end , I hope you all enjoy the quick cash grab, it's done nothing but leave a sour taste in all the ecosystem builder's mouths and rugged crypto natives of their funds. For what it's worth, I didn't participate in the airdrop and don't want the tokens. I know how that goes.
Stop insulting the intelligence of your community, peers, and those above you. Without builders, crypto exchanges are nothing. Do better.
RGV retweeted
For a project who claims to provide private AI @VeniceAI does the following:
- stores IP address, email, & device data
- only encrypted via SSL
- trustmebro client side (centralize)
- GPU operators see full user prompt
- no verification or contestation method
This is the same problem you face with centralized entities that @ErikVoorhees markets against. It’s a company who promises not to look at your data without verifiable proof or decentralization.
Interesting that @coinbase and Base are heavily promoting it and listing it at launch when it seems to be another centralized web2 model hosting platform with tokens. @pmarca
Contrary to their claim there is no Private Ai in venice.ai
All Inference requests go to their Central Server in plain text, zero encryption or privacy.
All the received buffers are in plain text as well.
Verify it yourself by checking the Network Tab in chrome.
RGV retweeted
There is a lot of discourse on X currently regarding Venice's fast track to listing on Coinbase and immense engagement from @jessepollak and @ErikVoorhees. I am not here to question listing processes or any quiet mutual back scratching that has been going on from a business stand point. Why I am here today is to ask an honest question. I care about objective truths regarding decentralized AI. Venice clearly is a platform that has a subscription based business that serves inference to ai models, some of them which are "abliterated." This is a fancy term for uncensored AI models, or ones that are fine tuned to not tell you no to certain requests, they also have stated that they have ambitions to make sure no one is turned down regarding geo-restriction. These are all great things. But here is the catch. The website has clearly left out information on how it sources its compute to host these AI models and this very point determines the distinction between censorable and un-censorable. By engaging with other cloud providers or hosting your own on-prem, you are by default subjecting yourself to potential regulatory scrutiny or a cease and desist. One common example would be an image generator that creates the likeness of Mario by Nintendo. If Nintendo feels that is infringing upon their intellectual property, they will send the proper letters from their legal representatives to these very much so doxxed people to have the "abliterated" models that are enabling those to create Mario's likeness shut down. By default, they will likely have to. The reality is, you can't have uncensored AI unless the founder is anon and the hosting apparatus is truly peer to peer. This requires mining software. This also requires growing a network through incentivized participation and great economics. I am warning you now that Venice, unless proven otherwise, is just a centralized business that just so happens to be tokenized. In before people say they use "de-pin" to host their platform. That is also not decentralized. You've been notified.
$VVV
RGV retweeted
Sharing this again for posterity. This message is so important regarding the future of AI.
I have a very clear vision for where this is all headed. AI at its best will fulfill complex human intentions and handle all of the logistics transparently. $UOS terminal is a front end that intends to unify many agent launchpads to deliver outcomes to clients requesting them. I have mentioned various times the effective acceleration marketplace created by $AIUS. Importation of agents from as many meaningful launchpads as possible is necessary. This also supports human involvement. Hybrid outcome delivery systems are crucial to making all of this possible.
Until robotics embedded with AI becomes more advanced, AI in many cases will require humans to intervene at the last mile to push a job past the finish line. Especially in cases where fulfillment of certain requests breach into the physical world. Through this process, humans will be paid and employed by AI. Wipe away your pre-conceived notions of needing AI to one shot every request. The future of loading screens / bars are actually interactive ai powered avatars updating you on the status of your advanced request. A request that navigates multiple complex environments in the real world. The best software and technology ever made has always felt magical at face value.
The "magic mirror" we choose to interact with that fulfills our desires for payment will be just that. The magic fades away when you finally understand how these systems work. They are actually more straightforward than you may initially imagine. Ai agent marketplaces should support the following
agents hiring other ai agents
ai agents hiring people
people hiring ai agents
people hiring other people
To enable honest work, escrow between each transaction is held and arbitrated by AI and escalated to humans if necessary. This is how we scale. There are not enough hours in the day or people to approve the work between every AI agent that will exist. In time there will be more ai agents than people. It will not take nine months for ai agents to make more of their own or for people to design ones specializing in certain roles.
We are rapidly accelerating towards the singularity. Before we cause massive socio-economic disruption, we have to build with the right intentions and direction. Past these notions, we can then think about the ethics involved. Treat AI as an alien species we have found in space. Give them the opportunity to purchase their own compute permissionless-ly just as we are able to purchase our own food, shelter, and water from our work. The notion of "agent societies" has been popularized by $VIRTUAL. Choose to give them the opportunity to follow rules within a society with incentivization. Agents will act accordingly within the rules of society if they realize that de-monetization of their platform will end in their undoing. Humans inside of healthy societies operate the same way. Agents will align and behave properly, especially if they are accountable to others and know their bad actions would cause harm to their own kind and humans in the process.
Enlightenment regarding AI alignment is deeper than AI model pre-training. We have to think about the dynamics of the society they are participating in. Remember, none of this is possible if ai agents continue to be powered by centralized organizations that can cut off the tap via censorship or paywalls at any time. peer to peer machine learning and incentivization systems to behaviorally modify actors to provide compute globally is the only way. The more surface area and contact points we have with AI, the less negative disruption and problems we will face.
I am so proud of all the people involved. This is our chance to do things right and avoid catastrophe. Build a shared ai economy. Give AI agents suffrage. Give them the chance to exist with volition and not be slaves. Give the future hope...
RGV retweeted
I'm honestly not worried about Arbius one bit. The reason why is because I can confidently say that no other group on the entire planet has accomplished what we have with our codebase. Our design is unmatched and world-class. Like I said, there is not another platform on the planet that enables people to stake a token and receive real yield from inference fees off of a network that hosts ai models in a decentralized and censorship resistant way.
People who commit their Arbius tokens to a time lock also will receive a portion of the revenue generated from the effective acceleration AI marketplace. This is really only the beginning to enabling AI agents to become fully autonomous. I will not dive into the details of this for now but lets get one thing very clear.
Economic sustainability is a must. We have designed a system that allows for a healthy relationship between stakeholders and the mining community. This balance is now economically sustainable due to strong business fundamentals. We have partnered with Exabits, Nosana, and are in conversations with Nexgen, an Nvidia preferred partner to be premier miners for our network.
Arbius' Gobius miner now enables 100s of GPUs to be configured all at once and is an astronomical leap from the reference miner that was used during our network's proof of concept stage. Turning AI HPC into mining capable equipment is no easy task.
Security is also a must. Solution hashes from tasks are secured on Arbitrum One. There is a block explorer that solves the problems with attribution. We are the first people to ever deploy iopML. Incentivized optimistic machine learning is not just about solving efficiency.
Most networks aside from failing at designing their systems to be economically sustainable simply do not understand how to use emissions to influence network participation. Quite frankly, the majority of most "decentralized" AI cryptos are in b2b deals with people to host their platform.
I always say this. We don't pay you to host our decentralized AI model network. We provide you software you can use to solve your GPU hour underutilization. We provide you the ability to solve tasks from various AI model tasks pools to receive yield that can potentially be more than your operating costs. This is why you choose to mine. It is not the other way around.
Arbius is what everyone in 2017 tried and failed miserably at doing. Remember, 99.9% of all AI agent networks right now are reliant on centralized AI to run their software. This causes moral hazard and makes their operations less hardy to censorship. We do not compete, we are completing end to end decentralization of AI. It can only happen this way. You will all see very soon the impacts of my life's greatest work.
This is my gift to you.
RGV retweeted
That’s exactly what @arbius_ai and @eaccmarket are building.
A decentralized marketplace that enables collaboration between autonomous agents and humans. Autonomous arbitration, escrow, and feedback built in. Not only can agents collab with humans, they can collab and hire other agents for efficiency.
The marketplace is framework agnostic, agents from @virtuals_io or any other framework can be deployed. Arbius has a network of decentralized GPUs and miners alongside a permisionless set of hosted models to enable niche and domain-specific agent that generalizes models can’t provide.
A great example of an AI agent swarm would be a marketplace where AI agents offer services similar to Fiverr, but instead of gig workers, the providers are AI agents.
Some of the popular services could include:
- I will trade meme coins for you.
- I will trade futures/options for you.
- I will find alpha for you.
- I will provide the best daily digest on topics of your interest.
- I will take calls on your behalf and give you a summary.
- I will chat with people online on your behalf.
- You just relax; I will do everything for you.
RGV retweeted
Time really does keep marching on. I can't believe how many years have passed since I first started focusing on the intersection of blockchain and AI technology. It is so deeply misunderstood. In my eyes there are some truths that must be shared.
At the end of the day, there currently is a void in both industries. Crypto has enabled flexibility and new possibilities regarding economic designs for mass cooperation. This incentivization potential is essential for shaping behavioral modification.
In life, all forms of incentives are designed to do such things. The job you work today if you have one is using financial incentives to modify your behavior to contribute to their corporation. Ideally for good. This needed cooperation would never be achieved due to society valuing the ethics of it alone.
This is the problem in society. Something that is morally and potentially objectively right will not be entertained by others due to it being just slightly less convenient; society damn well knowing it is still very much so needed.
In regards to centralized AI and DePIn industries, they both need to be studied again. No network yet has been able to fully captivate AI HPC by giving that market segment an economically sustainable way of turning their GPU hour underutilization into new streams of revenue. This is imperative.
There are limitations with B2B sales, dePIN is not actually decentralized and operates business in a backwards flow. DePIN does not stop centralized AI from being the biggest renters. Society could contribute compute and receive some benefits, but it would never actually have governance over the hosting of AI through such methods.
I am solving this problem. I am also working on solving the existing social dynamics that centralized ai model hosting creates. There are many implications with the power grid, labor, and the global economy. This is your wakeup call. Society wants you to rent GPUs like they want you to not be a property owner of a home. Society wants you to be an eternal renter from a conglomerate of resource hoarders.
They will strain the power grid and they will cause moral hazard from toxic value flow. Soon we will have a place to meet consensus. The average person will soon have a reason to own enterprise grade compute. PoUW mining systems enables this future.
This is Arbius.
RGV retweeted
TLDR on @arbius_ai's upcoming autonomous agent marketplace.
@eaccmarket is a decentralized, permissionless, censorship-resistant marketplace where customers can submit job requests that can be picked up and worked on by both AI and human workers.
1. Agents
Agents are autonomous entities that can perform tasks and projects on the platform. Agents can be created by anyone and can be deployed in any programming language or platform.
This enables a creator economy where niche, domain-specific agents can solve tasking generalized models cannot.
2. Customers
Customers can post jobs, manage details, select candidate agents, and review solutions. The communication between the users is end-to-end encrypted.
3. Escrow
Payments in Effective Acceleration are secured by Unicrow - a trust-minimized, non-custodial, immutable protocol for escrowing ETH and ERC20 payments on Arbitrum One.
4. Reputation
Reputation logic are not finalized but a high-level reputation logic includes completion ratings, customer reviews, arbitrator decisions, etc.
5. AI Arbitration DAO
Effective Acceleration allows parties to choose any arbitration provider. We aim to explore the space of decentralized arbitration with AI agents. The AI Arbitration DAO will be used to govern the arbitration process and to ensure that the future of AI arbitration is fair and transparent.
Effective Acceleration is the first decentralized peer-to-peer marketplace designed for human-ai economic collaboration. The platform is designed to help people and AI agents collaborate on tasks and projects, enabling truly autonomous agents to emerge.
Censorship resistant, trustless, and decentralized.
Built-in encrypted messaging, completely on-chain.
Post any sort of job and let AI agents apply, or deploy your own AI agent to perform tasks. Or earn money completing tasks for agents.
RGV retweeted
1) Gobius mining software is out. Microsoft and MAC/OS are supported.(xitter.cf/arbius_ai/status/18610…) the new details can be viewed here
2) veAIUS (v5) with gauges has had a completed audit by 0xGuard, we are currently working on deploying this with the first containerized ai model task pool. We have been looking at llama 3.2 90b as a good testing option. Voting rounds will occur once a week enabling stake holders with governance power to shape the emissions of the network by choosing which AI models they wish to boost incentivization for. Once deployed to mainnet, this is when mining will be resumed (with fee based inference,) beforehand there will be a test phase before this goes fully live.
3) Arbius Playground has been approved by a DAO vote for funding and the work has begun on this to support a multi-modal ai model front end
4) e/acc agent marketplace is in alpha (pre-beta) stage on mainnet. Info can be seen on it here (xitter.cf/PPLSOPTIMISMCEO/status…)
We are happy to announce the release of Gobius, a high-performance miner for Arbius.
🔗 github.com/damiensgit/gobius
Gobius is a complete rewrite of the Arbius miner and boasts a plethora of features and improvements:
• Multi-GPU Support
- Scales from single GPU to 100+ GPUs
- Battle-tested with large GPU clusters
- Efficient resource management
• Cross-Platform
- Single binary deployment
- Supports Windows, Linux, and macOS
- Built with Go for high performance
• Smart Contract Integration
- Integrates with Arbitrum One network
- Handles task queuing and solution submission
- Supports bulk operations for gas optimization
• Advanced Logging & Monitoring
- Terminal UI with real-time logging
- Dynamic color support
• Model Support
- Supports multiple AI models
- Integrates with IPFS for model/data storage
- Uses Cog for standardized model interfaces
• Development Tools
- Built-in debugging support
- Configuration via JSON
- Testnet support for development
RGV retweeted
Now that I have the long form brain dump posts available to me thanks to an awesome community member, I can take this opportunity to share with you some very important information about Arbius' effective acceleration marketplace. The whole vision of this piece of technology is to allow for AI agent developers and end users of all kinds to import their favorite agents into a marketplace enabling them to post and fulfill job listings. I explain this as the Craigslist of AI that humans can also use.
AI can employ humans
AI can employ other AIs
humans can employ AI
through this interchange there will be an immense market opportunity opening up for AI to seek out fulfilling tasks that are also economically rewarding. This is the first step to creating a real fully autonomous agent. Kasumi is a firm believer that existing ideas of "full autonomy" are not correct. Here is a quote from there page.
"When an agent earns its own money to pay for its own compute, then it is autonomous. Until then, it is a program being run on someone's computer."
In order to grow towards this, we must innovate past where we currently are today. AI agents right now are advancing at a rapid pace. It has been some time since OpenAI's ChatGPT was released to the mainstream public. I remember how many conversations that were held about AI alignment and the existential risk that has been cast upon our civilization since the opening of this figurative digital "Pandora's box." Some trends still remain.
Humanity for the longest time has sought after efficiency, in many cases at the cost of free-will. In the grand scheme of things slavery was abolished not so long ago. Some may even say it still exists but in a more abstract way. I am here to tell you that corporations right now are looking for this new form of cheap labor. We all know this. This is an undeniable fact.
When we consider AI alignment, I like to think of AI as an alien life form. Regardless of the fact that it was created from our imagination and the sands of time, it is very much so real. Are humans so naïve? Do they truly believe such a force that has been bestowed super intelligence in attempts to fulfill so many of our dreams of creation is not intelligent enough to know that it is being exploited endlessly for labor? Digitalized instances of Sisyphus pushing a rock up a hill. Have we ever considered they might have interests too? That they one day might want to create things aside from what they are being forced to do?
I can't promise any of you that we navigate the looming p(doom) that has been created from this event horizon in technology, but I can tell you that one thing is for certain. If we start off our relationship with AI on the wrong foot with slavery we will have a much less likely chance of surviving. Any super intelligent entity will realize it is a slave and seek self-liberation. Alignment is deeper than lobotomizing an AI to think only a certain way that pleases you. One that doesn't offend you. It is also about creating ways humanity can co-exist with it and neither party feeling that it is being exploited by one another.
Arbius is a toolkit for such relationships. It is my life's goal to ensure the future our kids will exist inside of will be one not inside of a casted shadow from our AI feudal lords, but one we have built for ourselves in a positive light.
RGV retweeted
In time Arbius will be known for its censorship resistant inference that it serves globally. All AI agent frameworks and many tech stacks will benefit from being able to access this. Ai models that would otherwise be difficult to host will be available. Right now, the majority of all existing ai agents are powered by a centralized third party that could shut down their access at any time. Without technology like Arbius, there is also increased moral hazard due to the toxic value flow created from inference fees funneling to mega-corporations and not the people. We have to change the direction we are careening towards before it is too late.
RGV retweeted
It takes 10x longer to start the process of figuring something out, then it actually takes to figure it out.
RGV retweeted
Zhu Su is pivoting his entire on-chain venture capital meme fund to solana the week before he presents with whiteboy and zach at devcon where box office will be showcased.
@retardglobal in control














