<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Neural Internet on Medium]]></title>
        <description><![CDATA[Stories by Neural Internet on Medium]]></description>
        <link>https://medium.com/@neuralinternet?source=rss-a261c76f5cd9------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*kiEpUKqQp29JUXgGegWOzg.png</url>
            <title>Stories by Neural Internet on Medium</title>
            <link>https://medium.com/@neuralinternet?source=rss-a261c76f5cd9------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 08:29:09 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@neuralinternet/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[PR Rewards for Subnet 27]]></title>
            <link>https://medium.com/@neuralinternet/pr-rewards-for-subnet-27-391083dfe358?source=rss-a261c76f5cd9------2</link>
            <guid isPermaLink="false">https://medium.com/p/391083dfe358</guid>
            <dc:creator><![CDATA[Neural Internet]]></dc:creator>
            <pubDate>Thu, 28 Mar 2024 21:51:47 GMT</pubDate>
            <atom:updated>2024-03-28T21:51:47.301Z</atom:updated>
            <content:encoded><![CDATA[<h4>Contribute to Building Decentralized Compute</h4><p>We’re excited to announce a new initiative to reward our community members for their valuable contributions to the Compute Subnet! Our goal is to foster innovation, improve the network’s functionality, and enhance user experience through community-driven development. If you’ve been looking to contribute, now is the perfect time to get involved and potentially earn a reward for your efforts.</p><h4>Reward Range:</h4><p>We are offering rewards ranging from $600 to $10,000 for pull requests (PRs) that add significant value to the Compute Subnet. The aim is to acknowledge the hard work, creativity, and impact of contributions that help propel the network forward.</p><h4>Evaluation Criteria:</h4><ul><li><strong>Value Addition:</strong> The PR should introduce features, fixes, or enhancements that significantly benefit the Compute Subnet’s functionality, performance, security, or user experience.</li><li><strong>Quality of Work:</strong> High-quality code, documentation, and adherence to best practices will be key factors in our evaluation.</li><li><strong>Innovation:</strong> We appreciate innovative solutions that address existing challenges or introduce new opportunities for the network.</li></ul><h3>Process and Payment:</h3><ol><li><strong>Submit Your PR:</strong> Contribute to the Compute Subnet by submitting your PR through our GitHub repository.</li><li><strong>Review Process:</strong> Our team will thoroughly review each submission to assess its impact, quality, and the value it adds to the network. This process may involve discussions, requests for changes, or further clarifications to ensure the contribution meets our standards and objectives.</li><li><strong>Reward Determination:</strong> Once a PR is approved, the reward amount will be determined based on the contribution’s overall value and the effort involved. We aim to make this process as transparent and fair as possible.</li><li><strong>Legal Compliance:</strong> To comply with tax regulations and ensure a smooth payout process, we require all contributors eligible for a reward to complete a W-8 (for non-U.S. persons) or W-9 (for U.S. persons) form. This step is crucial for us to file the necessary 1099 forms and cover the expenses associated with these filings.</li><li><strong>Payment Release:</strong> Upon successful review of the contribution and completion of the required tax documentation, the reward will be processed and released to the contributor.</li></ol><h4>Get Started:</h4><p>This initiative is open to developers, technologists, and enthusiasts passionate about decentralized computing and blockchain technology. Whether you’re a long-time member of our community or new to the Compute Subnet, we encourage you to dive in, explore open issues, and bring your ideas to life.</p><p>Together, we can build a stronger, more efficient, and inclusive Compute Subnet. We’re looking forward to seeing your innovative contributions and are excited to reward your efforts in making a significant impact on the network!</p><p>For more details about the contribution process, review guidelines, and tax documentation requirements, please visit our [Contribution Guide] on GitHub. Let’s innovate together and shape the future of decentralized computing!</p><h3>Contribution Guide for the Compute Subnet</h3><p>Welcome to the Compute Subnet Contribution Guide! We’re thrilled to have you consider contributing to the Compute Subnet. This guide outlines the process for contributing improvements, features, and bug fixes to the project. By participating, you’re helping to create a more robust, efficient, and innovative decentralized compute network.</p><h4>Getting Started</h4><p>Before you begin, it’s important to familiarize yourself with the project. Spend some time understanding the Compute Subnet, its architecture, and its codebase. Here are a few steps to get you started:</p><ol><li><strong>Explore the Repository:</strong> Visit the [Compute Subnet GitHub repository] to explore the existing codebase, open issues, and ongoing projects.</li><li><strong>Join the Community:</strong> Engage with the Compute Subnet community in the <a href="https://discord.gg/bittensor">Bittensor Discord Server.</a> It’s a great way to get help, discuss ideas, and stay updated.</li><li><strong>Review Open Issues: </strong>Look at the open issues for areas where you can contribute. Some issues may be tagged with “good first issue” for newcomers.</li></ol><h4>Making Contributions</h4><p><strong>Submitting a Pull Request (PR)</strong></p><ol><li><strong>Fork the Repository:</strong> Start by forking the repository to your <a href="https://github.com/neuralinternet/compute-subnet">GitHub account.</a></li><li><strong>Clone the Fork:</strong> Clone your fork locally to start working on the changes.</li></ol><p><em>```sh</em></p><p><em>git clone </em><a href="https://github.com/your-username/compute-subnet.git"><em>https://github.com/your-username/compute-subnet.git</em></a></p><p>```</p><p>3. <strong>Create a New Branch: </strong>For each new feature or fix, create a new branch from `main`.</p><p><em>```sh</em></p><p><em>git checkout -b feature/your-feature-name</em></p><p><em>```</em></p><p>4. <strong>Make Your Changes:</strong> Implement your changes, ensuring you adhere to the coding standards and best practices.</p><p>5. <strong>Write Tests:</strong> If applicable, write tests for your changes to ensure reliability and prevent future regressions.</p><p>6. <strong>Document Your Changes:</strong> Update the README.md or relevant documentation to reflect any new features or important changes.</p><p>7. <strong>Commit Your Changes:</strong> Use meaningful commit messages that clearly explain your changes.</p><p><em>```sh</em></p><p><em>git commit -am “Add a concise and clear description of your change”</em></p><p><em>```</em></p><p>8. <strong>Push to Your Fork:</strong> Push your changes to your fork on GitHub.</p><p><em>```sh</em></p><p><em>git push origin feature/your-feature-name</em></p><p><em>```</em></p><p>9. <strong>Submit a Pull Request:</strong> Go to the original Compute Subnet repository, and you’ll see a prompt to submit a pull request. Ensure your PR description clearly describes the changes and their impact.</p><h4>PR Review Process</h4><p><strong>Once submitted, your PR will undergo a review process by the project maintainers. The process includes:</strong></p><ul><li><strong>Initial Review:</strong> The maintainers will review the PR for its relevance, quality, and alignment with the project goals.</li><li><strong>Feedback and Changes:</strong> You may receive feedback and requests for changes. Engaging in discussions and making requested changes is part of the process.</li><li>Approval and Merge: Once approved, your PR will be merged into the main codebase.</li></ul><h3>Rewards Program</h3><p>For contributions that add significant value, we offer rewards ranging from $600 to $10,000. The reward amount depends on the impact, effort, and quality of the contribution. Please refer to our [Rewards Program]() document for more details, including the evaluation criteria and payout process.</p><h3>Legal and Compliance</h3><p>Contributors eligible for rewards are required to complete a W-8 (for non-U.S. persons) or W-9 (for U.S. persons) form to comply with tax regulations. This step is essential for filing the necessary 1099 forms.</p><h3>Code of Conduct</h3><p>We are committed to providing a welcoming and inclusive environment. All participants are expected to adhere to our [Code of Conduct](). Respect, professionalism, and constructive communication are paramount.</p><h3>Questions?</h3><p>If you have any questions or need further guidance, don’t hesitate to ask for help in the <a href="https://discord.gg/neuralinternet">NI Discord Server.</a> Our community and maintainers are here to support you. Thank you for contributing to the Compute Subnet! Together, we’re building the future of decentralized computing.</p><h3>Compute Subnet Code of Conduct</h3><h4>Our Commitment</h4><p>In the interest of fostering an open and welcoming environment, we as contributors and maintainers pledge to making participation in our project and our community a harassment-free experience for everyone, regardless of age, body size, disability, ethnicity, sex characteristics, gender identity and expression, level of experience, education, socio-economic status, nationality, personal appearance, race, religion, or sexual identity and orientation.</p><h4>Our Standards</h4><p><strong>Examples of behavior that contributes to creating a positive environment include:</strong></p><ul><li>Using welcoming and inclusive language</li><li>Being respectful of differing viewpoints and experiences</li><li>Gracefully accepting constructive criticism</li><li>Focusing on what is best for the community</li><li>Showing empathy towards other community members</li></ul><p><strong>Examples of unacceptable behavior by participants include:</strong></p><ul><li>The use of sexualized language or imagery and unwelcome sexual attention or advances</li><li>Trolling, insulting/derogatory comments, and personal or political attacks</li><li>Public or private harassment</li><li>Publishing others’ private information, such as a physical or electronic address, without explicit permission</li><li>Other conduct which could reasonably be considered inappropriate in a professional setting</li></ul><h4>Our Responsibilities</h4><p>Project maintainers are responsible for clarifying the standards of acceptable behavior and are expected to take appropriate and fair corrective action in response to any instances of unacceptable behavior.</p><p>Project maintainers have the right and responsibility to remove, edit, or reject comments, commits, code, wiki edits, issues, and other contributions that are not aligned with this Code of Conduct, or to ban temporarily or permanently any contributor for other behaviors that they deem inappropriate, threatening, offensive, or harmful.</p><h4>Scope</h4><p>This Code of Conduct applies within all project spaces, and it also applies when an individual is representing the project or its community in public spaces. Examples of representing a project or community include using an official project e-mail address, posting via an official social media account, or acting as an appointed representative at an online or offline event. Representation of a project may be further defined and clarified by project maintainers.</p><h4>Enforcement</h4><p>Instances of abusive, harassing, or otherwise unacceptable behavior may be reported by contacting the project team at <strong>[project email]</strong>. All complaints will be reviewed and investigated and will result in a response that is deemed necessary and appropriate to the circumstances. The project team is obligated to maintain confidentiality with regard to the reporter of an incident. Further details of specific enforcement policies may be posted separately.</p><p>Project maintainers who do not follow or enforce the Code of Conduct in good faith may face temporary or permanent repercussions as determined by other members of the project’s leadership.</p><blockquote><em>Neural Internet is an AI collective pushing the boundaries of machine intelligence. We are building the next generation of AI applications through distributed solutions.</em></blockquote><p><em>Read our documentations </em><a href="https://docs.neuralinternet.ai/"><em>here</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=391083dfe358" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Future of AI: Compute is King]]></title>
            <link>https://medium.com/@neuralinternet/future-of-ai-compute-is-king-54e7d98e7342?source=rss-a261c76f5cd9------2</link>
            <guid isPermaLink="false">https://medium.com/p/54e7d98e7342</guid>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[bittensor]]></category>
            <category><![CDATA[cloud-computing]]></category>
            <category><![CDATA[nvidia]]></category>
            <dc:creator><![CDATA[Neural Internet]]></dc:creator>
            <pubDate>Sat, 09 Mar 2024 22:32:37 GMT</pubDate>
            <atom:updated>2024-03-09T22:32:37.865Z</atom:updated>
            <content:encoded><![CDATA[<h4><strong>$2 Trillion Market Boom and Computational Revolution</strong></h4><p>The intersection of artificial intelligence and cloud computing is emerging as a key driver of economic growth and efficiency across various sectors. This convergence is anticipated to foster substantial market expansions and operational improvements, marking a pivotal shift in the way industries adapt and innovate.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/0*oeHwYt5xSL-Zquyy" /><figcaption>Source: Next Move Strategy Consulting</figcaption></figure><p>According to <a href="https://www.statista.com/statistics/1365145/artificial-intelligence-market-size/">Statista,</a> the AI market size is expected to grow by a staggering 20 times by 2030, reaching an estimated $2 trillion. As AI adoption is just beginning, it is crucial to consider the wide range of industries that will be directly affected by this technological shift, such as healthcare, customer service, and banking. The AI boom has also driven substantial growth in cloud computing, supported by the training and inferencing of Large Language Models in generative AI applications. This trend has directly impacted the cloud computing industry, significantly increasing the demand for computational resources.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*iR8mr3A1TQrMbLqE" /><figcaption>Source: Mordor Intelligence</figcaption></figure><p><a href="https://www.fortunebusinessinsights.com/cloud-computing-market-102697">Fortune Business Insights</a> reported that the cloud computing market size was valued at USD 569.31 billion in 2022 and is projected to grow from USD 677.95 billion in 2023 to USD 2,432.87 billion by 2030, exhibiting a Compound Annual Growth Rate (CAGR) of 20.0% during the forecast period.</p><p><a href="https://www.grandviewresearch.com/industry-analysis/cloud-computing-industry">Grandview Research</a> indicates that the global cloud computing market size was valued at USD 483.98 billion in 2022 and is expected to grow at a compound annual growth rate (CAGR) of 14.1% from 2023 to 2030.</p><p>According to<a href="https://www.mordorintelligence.com/industry-reports/cloud-computing-market"> Mordor Intelligence</a>, the Cloud Computing Market size is estimated at USD 0.68 trillion in 2024, and is expected to reach USD 1.44 trillion by 2029, growing at a CAGR of 16.40% during the forecast period (2024–2029).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/0*qrJd6ETnnozfhM5u" /><figcaption>Source: Precedence Research</figcaption></figure><p>The anticipated explosive growth in the specialized AI chip sector is driven by increasing computational demands. <a href="https://www2.deloitte.com/us/en/insights/industry/technology/technology-media-and-telecom-predictions/2024/generative-ai-chip-market-to-reach-40-billion-in-2024.html">Deloitte’s comprehensive study</a> projects that the generative AI chip market could surpass $50 billion by 2024.</p><p>In the second quarter of the fiscal year 2023, Nvidia sold $10.3 billion worth of datacenter hardware. Market research firm Omdia looked into the numbers and found out that Nvidia shipped a whopping 900 tons of its H100 GPUs during that time. These H100s, used for machine learning and high-performance computing, added up to about 1.8 million pounds in weight.</p><p>Omdia estimated that Nvidia shipped around 300,000 H100 GPUs in just three months, averaging 1.84 kilograms each. They expect Nvidia to maintain this shipping volume, potentially reaching 1.2 million H100 GPUs per year. Despite the significant shipments, there is still a massive unmet demand for these AI GPUs, which may lead to even higher shipments this year. To put the 900 tons of GPUs into perspective, it is equivalent to the weight of 4.5 Boeing 747 planes or 11 Space Shuttle Orbiters.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*cqOYUC19eAybsw45" /><figcaption>Source: Omdia Research</figcaption></figure><p><strong>The Computational Market: AI’s Accelerating Demand and the Future of Compute to $7 Trillion</strong></p><p>Sam Altman is reportedly seeking to raise an extraordinary<a href="https://www.tomshardware.com/tech-industry/artificial-intelligence/openai-ceo-sam-altman-seeks-dollar5-to-dollar7-trillion-to-build-a-network-of-fabs-for-ai-chips"> $5 to $7 trillion</a>, which is hard to wrap one’s head around. The goal is to build a network of semiconductor fabs (fabrication plants) that would significantly enhance the world’s capacity for chip production, crucial for powering advanced AI technologies. This ambitious initiative aims to connect OpenAI with various investors, contract chip manufacturers, and power providers to create a robust infrastructure capable of meeting the soaring demand for AI computational power.</p><p>It’s clear we are heading for a computational driven economy. With the beginning of the internet era, we entered into the tech stock boom and bust. Where fortunes were made and lost within months. Afterwards new giants emerged, computers became pocket sized smartphones and Alexa in your kitchen. These giants grew into large data banks and server farms for the cloud. It marked the beginning of this demand for computational resources to power our new world.</p><h4>Moore’s Law and AI Training</h4><p>At a rate faster than the historical trend of Moore’s Law, AI training is redefining the meaning of training AI. Doubling every 3.4 months since 2012, as<a href="https://openai.com/research/ai-and-compute"> highlighted by OpenAI</a>. Computing in AI training runs has grown by more than 300,000 times, a rate of increase that far exceeds what would have been expected based on Moore’s Law alone, which would have yielded only a 7x increase over the same period. This storm is expected to increase as developments increase over time, with the acceleration of AI, it’s to be expected.</p><p>This rapid increase in compute demand for AI underscores the accelerating pace of advancements in AI technology and the significant investment in computational resources required to train increasingly sophisticated AI models. The trend highlights the critical need for continued innovation in compute solutions and infrastructure to support the growing capabilities and applications of AI.</p><h3>GPU Shortages and the Concentration of Compute</h3><p>The long-term viability of AI development faces a significant challenge due to a critical shortage of advanced chips, such as Nvidia’s H100 and A100. These chips are essential for effectively training, fine-tuning, and performing inference tasks in AI.</p><p>The scarcity of crucial AI resources consolidates considerable market power among a few key players — chip manufacturers such as the Taiwan Semiconductor Manufacturing Company, chip designers like Nvidia, and cloud infrastructure giants including Google, Microsoft, and Amazon.</p><p>These behemoths not only control essential hardware but also benefit from vast platform ecosystems, providing them unparalleled access to data and diverse avenues for AI commercialization, coupled with a first-mover advantage in large-scale AI operations.</p><p><a href="https://www.forbes.com/sites/johanmoreno/2023/07/25/70-of-generative-ai-startups-rely-on-google-cloud-ai-capabilities-says-alphabet-ceo-sundar-pichai/">According to Forbes</a>, 70% of generative AI startups have become reliant on Google Cloud’s infrastructure and AI capabilities. This revelation from Alphabet’s Q2 2023 earnings call underscores Google Cloud’s dominance in providing AI-optimized infrastructure crucial for training and serving generative AI models. Google Cloud showcased a significant revenue increase to $8 billion in Q2 marking a pivotal moment in the cloud division’s growth.</p><p>Startups like Cohere, Jasper, and Typeface, among others, are leveraging Google Cloud’s vast resources, advanced TPU chips, and a library of pre-trained AI models, to expedite their development processes. This trend casts a spotlight on a pressing issue: the centralization of AI resources and infrastructure.</p><p>The high demand and concentrated supply of these chips have led to innovative solutions, including using GPUs as collateral for financing or the development of decentralized cloud computing platforms to source an ever-expanding distributed pool of compute.</p><h3>Democratizing AI: Decentralized Compute Market</h3><blockquote><em>The term “compute” can have different meanings depending on the context. It may refer to the amount of computational work required for specific tasks, such as training an AI model, or strictly to hardware components like processors. Often, “compute” encompasses both hardware and software elements within a computing system.</em></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5vLGA6dYr6EQomV2qew9dQ.png" /><figcaption>NetUID: 27</figcaption></figure><h4>Shaping the Future of Computing<strong>: How Subnet 27 Democratizes AI Development</strong></h4><p>Compute markets of today consist of siloed pools of compute resources separated by each provider and its designated apps to access each pool of compute. Subnet 27 decentralizes computing resources by combining siloed pools of compute on a blockchain to be validated and accessed trustlessly. This opens a door to scalable compute without the constraints of centralized power.</p><p>Decentralized cloud computing is an innovative approach that utilizes a peer-to-peer network to distribute computational resources across multiple nodes or miners, instead of relying on centralized data providers. As cloud platforms proliferate, there is a growing need for a decentralized compute network that can seamlessly integrate these platforms, enabling efficient resource sharing and allocation.</p><h4><strong>NetUID: 27 — Compute Subnet </strong><a href="https://bittensor.com/"><strong>Powered by Bittensor</strong></a></h4><p>Neural Internet’s Subnet 27 introduces a permission-less compute market for platform composable cloud services. The rapid growth of AI’s market size and the growing demand for computational power showcases a need for a higher levels of flexibility and efficiency in managing these resources. Subnet 27 integrates various third-party cloud platforms into one cohesive unit enabling a decentralized higher-level cloud computing infrastructure.</p><p>Incentivized by Bittensor’s<a href="https://bittensor.com/whitepaper"> Yuma Consensus</a>, Subnet 27 allows individuals and organizations to contribute their computing resources and, in return, earn rewards in $TAO, Bittensor’s native digital asset. Validators on the subnet then ensure the integrity and efficiency of the subnet while providing access to the computational resources available on the network.</p><p>Subnet 27 is facilitating a shift towards a more accessible and flexible computational landscape, which is particularly beneficial for compute-intensive tasks such as those required in AI and machine learning domains.</p><h4>Future of AI is Compute</h4><p>In conclusion, the rapid growth of the AI market and the increasing demand for computational resources are driving a significant shift in the technology landscape. The convergence of AI and cloud computing is expected to create a $2 trillion market by 2030, with the cloud computing market alone projected to reach $2.4 trillion by the same year.</p><p>The shortage of advanced AI chips and the concentration of computational resources among a few key players highlight the need for innovative solutions. Decentralized compute markets, such as Subnet 27 powered by Bittensor, offer a promising approach to democratizing AI development by leveraging a peer-to-peer network to distribute computational resources.</p><p>As the world moves towards a computation-driven economy, the development of decentralized, platform-composable cloud services will be crucial in ensuring the accessibility and flexibility of computational resources, ultimately shaping the future of AI and its widespread adoption across various industries.</p><blockquote><em>Neural Internet is an AI collective pushing the boundaries of machine intelligence. We are building the next generation of AI applications through distributed solutions.</em></blockquote><p><em>Read our documentations </em><a href="https://docs.neuralinternet.ai/"><em>here</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=54e7d98e7342" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[NetUID 27: Post Mortem]]></title>
            <link>https://medium.com/@neuralinternet/netuid-27-post-mortem-4ed708041bc6?source=rss-a261c76f5cd9------2</link>
            <guid isPermaLink="false">https://medium.com/p/4ed708041bc6</guid>
            <dc:creator><![CDATA[Neural Internet]]></dc:creator>
            <pubDate>Wed, 10 Jan 2024 02:58:22 GMT</pubDate>
            <atom:updated>2024-01-12T19:29:50.021Z</atom:updated>
            <content:encoded><![CDATA[<h4><strong>Exploit Analysis &amp; Solution: Proof-of-Work</strong></h4><p>In the ever-evolving landscape of decentralized AI and blockchain technology, the integrity and reliability of computational networks are paramount. Neural Internet’s Subnet-27 encountered challenges in its initial approach to validating miner hardware capabilities. This post-mortem analysis delves into the issues faced and the innovative solutions implemented to create a more secure and reliable subnetwork.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rh7qi5Le2eabULFquUgR0g.png" /></figure><h4>Initial Challenges</h4><p>The initial procedure for scoring allocation within subnet-27 involved a generated script on the miner’s machines, modified in real-time by the validator using a public key to encrypt the result of the script. This certification process guaranteed tamper-proof and certified responses because miners never get access to the private key of validators.</p><p>However, a critical issue emerged due to the use of conventional Python libraries and open-source components within the generated script. These components relied on binary files and kernel-written information which in theory are non-modifiable files (for kernel information). Unfortunately, miners were able to manipulate these binaries, replacing them with pre-defined outputs (using if conditions and simple echo commands), circumventing the intended validation process.</p><p>Additionally, fraudulent miners exploited a technique to deceive the Linux kernel into recognizing the system as equipped with numerous GPUs or RAM, tricking the Python program. Consequently, despite the signed and encrypted results, the outputs were falsified, originating from manipulations within the Linux kernel itself.</p><h4>Adopting a New Strategy</h4><p>The proposed solution currently in place leverages a similar concept of proof of work, akin to Bitcoin’s approach. Currently Bitcoin miners are using either bruteforce nonce or a random nonce with a sha-256 algorithm until they find the hash that suits the current block difficulty (which is dynamic over the total hashrate available). For our implemented solution, the dynamic difficulty attempts to reproduce the current design of Bitcoin, but the algorithms and the bruteforce differs a bit. Instead, our bruteforce password challenge uses a password + salt with BLAKE2b-512 algorithm. We picked a different algorithm than Bitcoins because of the general purpose of the subnet requiring GPUs for mining and due to the fact that the BLAKE2b-512 algorithm is ASICS resistant.</p><p>Upon finding the solution, the miner returns the password to the validator, who verifies its consistency with the original password generated for that specific miner. Each iteration generates a new hash and a distinct salt for each miner. The difficulty level is contingent on the length and character set of the password, both of which are communicated to the miners during the challenge. An enhancement strategy involves randomly selecting characters from an extensive character set, mitigating dictionary-based brute force attacks in favor of hash-based brute force methods.</p><h4>Impact and Future Directions</h4><p>The implementation of this new strategy has shown promising results in initial tests. It effectively measures the real power of each miner’s hardware, ensuring fairness and preventing fraudulent activities. As we continue to develop a resilient and permissionless compute network, we remain committed to addressing all challenges and collaborating with other validators to strengthen the ecosystem. Our team’s dedication to creating a secure compute subnet — a critical component of decentralized AI — has led to increased emissions and a call for continued support and guidance from the community.</p><h4>Conclusion</h4><p>The journey of enhancing NetUID 27 has been a testament to the dynamic nature of blockchain technology and the importance of adaptive security measures. By sharing our experiences and solutions, we hope to contribute to the broader discussion on building robust, decentralized Bittensor subnetworks.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4ed708041bc6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Subnet 27: Incentive Increase]]></title>
            <link>https://medium.com/@neuralinternet/compute-subnet-incentive-increase-3b287e021dd2?source=rss-a261c76f5cd9------2</link>
            <guid isPermaLink="false">https://medium.com/p/3b287e021dd2</guid>
            <category><![CDATA[bittensor]]></category>
            <dc:creator><![CDATA[Neural Internet]]></dc:creator>
            <pubDate>Tue, 05 Dec 2023 00:29:07 GMT</pubDate>
            <atom:updated>2023-12-05T00:29:38.507Z</atom:updated>
            <content:encoded><![CDATA[<h4>NETUID 27 Testing Phase Conclusion</h4><p>The two week testing phase for NETUID 27 has concluded. Emissions have increased and compute resources are being absorbed.</p><p>Mining on NETUID 27 requires much less technical expertise compared to other subnets. This lower barrier for entry to mining will increase competition as incentives increase.</p><p>Below are key fixes and updates that took place while we tested live on mainnet with minimal incentives.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rh7qi5Le2eabULFquUgR0g.png" /><figcaption>Compute Resources on Subnet 27</figcaption></figure><h3><strong>2023/12/04 Major Updates (bug fixes)</strong></h3><p><strong>Issues Addressed:</strong></p><ul><li>Validator hanging after setting weights</li><li>Revised README to include more details on how the subnet functions</li><li>Scoring mechanism related to indexing</li><li>Miners running compute resources and not activating docker; now they will see near 0 incentive if docker is not activated ( very important this is what gives validators / clients access to the resources on the subnet)</li><li>Addressed issues with register.py</li><li>Added a feature to register.py to enable validators / clients to specify the amount of GPUs they’d like to reserve access to</li><li>Allocation mechanism</li><li>Enhanced the validation mechanism now way more robust (fixes issues with wandb now all scores and runs are now being shown in weights and biases)</li></ul><p><strong>Testing:</strong></p><ul><li>Added auto-updates to ensure continuous efficiency.</li></ul><p>To take advantage of Auto Update, simply include the — auto_update flag when running your processes. The flag supports the following options:</p><pre>--auto_update major: Automatically updates and exits the process if there is a major version change.<br>--auto_update minor: For minor version changes.<br>--auto_update patch: For patch version changes.<br>--auto_update no: Disables the auto-update feature.</pre><p>When you run your processes with the — auto_update flag, updates will be managed automatically.</p><p>Additionally, if you are using pm2 to manage your processes, pm2 will automatically restart your code after an update, ensuring minimal downtime and a smooth transition to the latest version.</p><h4>GitHub PRs</h4><p><a href="https://github.com/neuralinternet/compute-subnet/pull/46">https://github.com/neuralinternet/compute-subnet/pull/46</a> <a href="https://github.com/neuralinternet/compute-subnet/pull/47">https://github.com/neuralinternet/compute-subnet/pull/47</a> <a href="https://github.com/neuralinternet/compute-subnet/pull/48">https://github.com/neuralinternet/compute-subnet/pull/48</a> <a href="https://github.com/neuralinternet/compute-subnet/pull/49">https://github.com/neuralinternet/compute-subnet/pull/49</a> <a href="https://github.com/neuralinternet/compute-subnet/pull/50">https://github.com/neuralinternet/compute-subnet/pull/50</a> <a href="https://github.com/neuralinternet/compute-subnet/pull/51">https://github.com/neuralinternet/compute-subnet/pull/51</a> <a href="https://github.com/neuralinternet/compute-subnet/pull/52">https://github.com/neuralinternet/compute-subnet/pull/52</a> <a href="https://github.com/neuralinternet/compute-subnet/pull/53">https://github.com/neuralinternet/compute-subnet/pull/53</a> <a href="https://github.com/neuralinternet/compute-subnet/pull/55">https://github.com/neuralinternet/compute-subnet/pull/55</a> <a href="https://github.com/neuralinternet/compute-subnet/pull/56">https://github.com/neuralinternet/compute-subnet/pull/56</a></p><p>You see this error in the validator logs when the wrong port number is being broadcasted:</p><p>2|Vali-27 | 2023-11-29 16:16:50.750 | DEBUG | dendrite | &lt;-- | 55511125 B | PerfInfo | 5E576y6dS3vR28ipJqHTSnikpegbvP1ZMCf651ZvPViEpXhE | 66.32.21.42:8105 | 503 | Service at 66.32.21.42:8105/PerfInfo unavailable.</p><p>You see this error in the validator logs when the axon is unreachable:</p><p>2|Vali-27 | 2023-12-02 19:22:19.654 | DEBUG | dendrite | &lt;-- | 55511067 B | PerfInfo | 5Cqv82kh4dvvqSdp4reCvR1LbVKNQxbt8GohQjquYjtPbDWo | 66.32.21.42:8108 | 408 | Timedout after 30.0 seconds.</p><p>You see this error in the miner logs when there’s a problem with python version and GLIBC version:</p><p>1|miner | 2023-11-24 20:25:17.972 | ERROR | Exception: Internal server error with error: [Errno 2] No such file or directory: &#39;./neurons/Miner/app&#39;</p><p><em>As of now, DO NOT RUN A MINER AND VALIDATOR ON THE SAME HOTKEY IT CAUSES VALIDATION ISSUES</em></p><p><strong>Actions Required</strong></p><p>Please pull the latest code and restart the already-running process using pm2 restart all command. For example:</p><pre>git pull<br>pm2 restart all </pre><h3>Future developments</h3><p>As the subnet matures, we are preparing to release a front-end that allows users to reserve and rent out compute from Subnet 27 through the Neural Internet validator. In addition to this, we are working to add support for Docker-based cloud platforms such as Runpod, Vast.AI and Lambda. Feel free to <a href="https://docs.google.com/forms/d/e/1FAIpQLScSHbFW-tjrODBnhg4t5rrzEh42Ivi4AdE3C78FQgH9-ia-Ig/viewform"><strong>submit a bug report or any feature requests</strong></a> you might have.</p><p>Thank you to everyone who tested Subnet 27 and all the miners contributing their compute. We can confidently conclude the testing phase and ask for all validators to consider increasing emissions for subnet 27!</p><p>Let’s build the decentralized compute layer Bittensor needs.</p><p>Happy mining &amp; validating!</p><p><strong><em>Neural Internet</em></strong></p><blockquote><em>Neural Internet is an AI collective pushing the boundaries of machine intelligence. We are building the next generation of AI applications through distributed solutions.</em></blockquote><p><em>Read our documentations </em><a href="https://docs.neuralinternet.ai/introduction/overview"><em>here</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3b287e021dd2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Run a Compute Miner?]]></title>
            <link>https://medium.com/@neuralinternet/how-to-run-a-compute-miner-82498b93e7e1?source=rss-a261c76f5cd9------2</link>
            <guid isPermaLink="false">https://medium.com/p/82498b93e7e1</guid>
            <category><![CDATA[bittensor]]></category>
            <dc:creator><![CDATA[Neural Internet]]></dc:creator>
            <pubDate>Fri, 01 Dec 2023 20:30:42 GMT</pubDate>
            <atom:updated>2024-03-12T18:25:37.307Z</atom:updated>
            <content:encoded><![CDATA[<h4>Subnet 27: Earn $TAO mining on Bittensor’s Subnet 27</h4><p>Subnet 27 is a permissionless compute market for platform-composable cloud services. Integrating various cloud platforms into a cohesive unit. Its purpose is to enable higher-level cloud platforms to offer seamless compute composability across different underlying platforms. Given the proliferation of cloud computing, there is a growing need for a subnet capable of seamlessly integrating these platforms, thereby allowing efficient resource sharing and allocation. This compute-composable subnet enables nodes to contribute computational power, with validators ensuring the integrity and efficiency of the shared resources, thus empowering the entire Bittensor ecosystem and cloud computing as a whole.</p><p>For miners interested in joining this innovative network, Subnet 27 offers the opportunity to contribute computing resources and <strong>earn $TAO</strong> in return. This guide is structured to provide a comprehensive breakdown of how you can get started with contributing to Bittensor’s commodity markets using your compute power.</p><h3>Getting Started</h3><p>Here are some GPU providers. Choose a provider or use your own hardware. <strong>We do not support container-based instances</strong> such as Runpod, Vast AI, and lambda:</p><ul><li><a href="https://datacrunch.io/">DataCrunch</a></li><li><a href="https://www.oracle.com/cloud/compute/gpu/">Oracle</a></li><li><a href="https://www.coreweave.com/">Corweave</a></li><li><a href="https://www.fluidstack.io/">FluidStack</a></li><li><a href="https://www.latitude.sh/accelerate">Latitude.sh</a> (referral: BITTENSOR27)</li><li><a href="https://oblivus.com/">Oblivus</a></li></ul><p><strong>Examples of GPUs to rent</strong> (listed in order of computing power)</p><ul><li>NVIDIA H100</li><li>NVIDIA A100s</li><li>NVIDIA A6000s</li><li>NVIDIA A4000s</li><li>NVIDIA 3090s/4090s</li></ul><p>Competition is the name of the game, and our playing field is ever-evolving, so be ready to bring your GPUs and put on your game face!</p><h3>Setting up a Miner</h3><p>Watch this <a href="https://youtu.be/5UJlKwAhYdE">How-To Video</a> while following this article.</p><p><strong>This installation process requires Ubuntu 22.04 and python3.8 or higher.</strong></p><p>It is best practice to default to using a local subtensor.</p><p><strong>Subtensor Setup Github: </strong><a href="https://github.com/opentensor/subtensor/blob/main/docs/running-subtensor-locally.md">https://github.com/opentensor/subtensor/blob/main/docs/running-subtensor-locally.md</a></p><p><strong>Compute Subnet Github:</strong> <a href="https://github.com/neuralinternet/compute-subnet">https://github.com/neuralinternet/compute-subnet</a></p><p><strong>Compute Subnet Gitbook Docs:</strong> <a href="https://docs.neuralinternet.ai/products/subnet-27-compute/bittensor-compute-subnet-miner-setup">https://docs.neuralinternet.ai/</a></p><h3>Install Docker</h3><p><em>To run a miner, you must </em><a href="https://docs.docker.com/engine/install/ubuntu"><em>install</em></a><em> Docker and run the service. If Docker is already installed on your machine, scroll down to install Bittensor.</em></p><p><strong>Install Link</strong>: <a href="https://docs.docker.com/engine/install/ubuntu/#install-using-the-repository">https://docs.docker.com/engine/install/ubuntu/#install-using-the-repository</a></p><p>Verify that the Docker Engine installation is successful by running the hello-world image.</p><p><em>This command downloads a test image and runs it in a container. When the container runs, it prints a confirmation message and exits.</em></p><pre>sudo docker run hello-world</pre><p>It is best practice to default to using a local subtensor.</p><p><strong>Subtensor Setup Github: </strong><a href="https://github.com/opentensor/subtensor/blob/main/docs/running-subtensor-locally.md">https://github.com/opentensor/subtensor/blob/main/docs/running-subtensor-locally.md</a></p><h4>Install and Run Subtensor With:</h4><pre>git clone https://github.com/opentensor/subtensor.git<br>cd subtensor<br># to run a lite node on the mainnet:<br>sudo ./scripts/run/subtensor.sh -e docker --network mainnet --node-type lite</pre><p><strong>cd out of directory once complete</strong></p><h4>BEGIN BY INSTALLING BITTENSOR</h4><pre>/bin/bash -c &quot;$(curl -fsSL https://raw.githubusercontent.com/opentensor/bittensor/master/scripts/install.sh)&quot;</pre><p>See Bittensor’s documentation for alternative installation instructions.</p><p><strong>Bittensor Documentation:</strong> <a href="https://docs.bittensor.com/">docs.bittensor.com</a></p><p><strong>VERIFY THE INSTALLATION</strong></p><pre>btcli --help</pre><p>Which will give you an output similar to the one below:</p><pre>Explain<br>usage: btcli &lt;command&gt; &lt;command args&gt;</pre><pre>bittensor cli v6.1.0</pre><pre>positional arguments:<br>  {subnets,s,subnet,root,r,roots,wallet,w,wallets,stake,st,stakes,sudo,su,sudos,legacy,l}<br>    subnets (s, subnet)<br>                        Commands for managing and viewing subnetworks.<br>    root (r, roots)     Commands for managing and viewing the root network.<br>    wallet (w, wallets)<br>                        Commands for managing and viewing wallets.<br>    stake (st, stakes)  Commands for staking and removing stake from hotkey accounts.<br>    sudo (su, sudos)    Commands for subnet management<br>    legacy (l)          Miscellaneous commands.</pre><pre>options:<br>  -h, --help            show this help message and exit<br>  --config CONFIG       If set, defaults are overridden by passed file.<br>  --strict              If flagged, config will check that only exact arguments have been set.<br>  --no_version_checking<br>                        Set true to stop cli version checking.<br>  --no_prompt           Set true to stop cli from prompting the user.using the [Bittensor Command Line Interface](&lt;https://docs.bittensor.com/getting-started/reference/btcli&gt;) with **btcli --help*** and/or check the installation in python.Run the below command to install Bittensor in the above virtual environment.git clone https://github.com/opentensor/subtensor.git<br>docker compose up sudo ufw allow 30333/tEnteps to verify that subtensor is running.</pre><p><strong>Create a Cold &amp; Hotkey with the commands below:</strong></p><pre>btcli w new_coldkey</pre><pre>btcli w new_hotkey</pre><p>If you already have a Key, you can regenerate it ‘safely’ on a machine using btcli w regen_coldkeypub. However, you must regen the full key if you plan to register or transfer from that wallet. Essentiallyregen_coldkeypub lets you load the key without exposing your mnemonic to the server. If you want to, you can generate a key pair on a local safe machine to use as cold storage for the funds you send.</p><pre>btcli w regen_coldkeypub</pre><pre>btcli w regen_coldkey</pre><pre>btcli w regen_hotkey</pre><h4>CLONE COMPUTE-SUBNET</h4><pre>git clone https://github.com/neuralinternet/Compute-Subnet.git</pre><p>Access the Compute-Subnet Directory</p><pre>cd Compute-Subnet</pre><h4>COMPUTE SUBNET DEPENDENCIES</h4><p>For optimal functionality of the Compute Subnet, it’s essential to install the appropriate graphics drivers and dependencies.</p><p><strong>Required</strong> <strong>dependencies for validators and miners:</strong></p><pre>cd Compute-Subnet<br>python3 -m pip install -r requirements.txt<br>python3 -m pip install -e .</pre><p>You may need to execute python3 -m pip install -r requirements.txt twice if it fails the first time.</p><h4>EXTRA DEPENDENCIES FOR MINERS</h4><pre>sudo apt -y install ocl-icd-libopencl1 pocl-opencl-icd</pre><p>Install Hashcat</p><p>Recommended hashcat version &gt;= v6.2.5</p><pre>wget https://hashcat.net/files/hashcat-6.2.6.tar.gz</pre><pre>tar xzvf hashcat-6.2.6.tar.gz</pre><pre>cd hashcat-6.2.6/</pre><pre>sudo make</pre><pre>sudo make install</pre><pre>export PATH=$PATH:/usr/local/bin/</pre><pre>echo &quot;export PATH=$PATH&quot;&gt;&gt;~/.bashrc</pre><pre>hashcat --version</pre><p>Version should output v6.2.6</p><p>now cd out of directory</p><h4>Download the NVIDIA CUDA Toolkit</h4><p>If the Nvidia toolkit and drivers are already installed on your machine, scroll down to verify, then download Docker.</p><pre>wget https://developer.download.nvidia.com/compute/cuda/12.3.1/local_installers/cuda-repo-ubuntu2204-12-3-local_12.3.1-545.23.08-1_amd64.deb</pre><pre>sudo dpkg -i cuda-repo-ubuntu2204-12-3-local_12.3.1-545.23.08-1_amd64.deb</pre><pre>sudo cp /var/cuda-repo-ubuntu2204-12-3-local/cuda-*-keyring.gpg /usr/share/keyrings/</pre><pre>sudo apt-get update</pre><pre>sudo apt-get -y install cuda-toolkit-12-3</pre><pre>sudo apt-get -y install -y cuda-drivers</pre><pre>export CUDA_VERSION=cuda-12.3<br>export PATH=$PATH:/usr/local/$CUDA_VERSION/bin<br>export LD_LIBRARY_PATH=/usr/local/$CUDA_VERSION/lib64</pre><pre>echo &quot;&quot;&gt;&gt;~/.bashrc<br>echo &quot;PATH=$PATH&quot;&gt;&gt;~/.bashrc<br>echo &quot;LD_LIBRARY_PATH=$LD_LIBRARY_PATH&quot;&gt;&gt;~/.bashrc</pre><p>You may need to reboot the machine at this point</p><pre>sudo reboot</pre><p>The simplest way to check the installed CUDA version is by using the NVIDIA CUDA Compiler (<strong>nvcc</strong>).</p><pre>nvidia-smi<br>nvcc --version</pre><p>The output of which should look something like</p><pre>+---------------------------------------------------------------------------------------+<br>| NVIDIA-SMI 545.29.06              Driver Version: 545.29.06    CUDA Version: 12.3     |<br>|-----------------------------------------+----------------------+----------------------+<br>| GPU  Name                 Persistence-M | Bus-Id        Disp.A | Volatile Uncorr. ECC |<br>| Fan  Temp   Perf          Pwr:Usage/Cap |         Memory-Usage | GPU-Util  Compute M. |<br>|                                         |                      |               MIG M. |<br>|=========================================+======================+======================|<br>|   0  NVIDIA RTX                     Off | 00000000:05:00.0 Off |                  Off |<br>| 30%   34C    P0              70W / 300W |  400MiB / 4914000MiB |      4%      Default |<br>|                                         |                      |                  N/A |<br>+-----------------------------------------+----------------------+----------------------+<br>                                                                                         <br>+---------------------------------------------------------------------------------------+<br>| Processes:                                                                            |<br>|  GPU   GI   CI        PID   Type   Process name                            GPU Memory |<br>|        ID   ID                                                             Usage      |<br>|=======================================================================================|<br>|  No running processes found                                                           |<br>+---------------------------------------------------------------------------------------+</pre><pre>nvcc: NVIDIA (R) Cuda compiler driver<br>Copyright (c) 2005-2023 NVIDIA Corporation<br>Built on Fri_Nov__3_17:16:49_PDT_2023<br>Cuda compilation tools, release 12.3, V12.3.103<br>Build cuda_12.3.r12.3/compiler.33492891_0</pre><h4>PM2 Installation</h4><p>Install and run pm2 commands to keep your miner online at all times.</p><pre>sudo apt update</pre><pre>sudo apt install npm</pre><pre>sudo npm install pm2 -g</pre><p>Confirm pm2 is installed and running correctly</p><pre>pm2 ls</pre><h4>START THE DOCKER SERVICE IN COMPUTE SUBNET</h4><pre>cd Compute-Subnet</pre><pre>sudo systemctl start docker</pre><pre>sudo apt install at</pre><p>Make sure to check that docker is properly installed and running correctly:</p><pre>sudo service docker status</pre><p>This is an example of it running correctly:</p><pre>root@merciful-bored-zephyr-fin-01:~# sudo service docker status<br>● docker.service - Docker Application Container Engine<br>     Loaded: loaded (/lib/systemd/system/docker.service; enabled; vendor preset&gt;<br>     Active: active (running)</pre><h4>SETTING UP A MINER</h4><p>Miners contribute processing resources, notably GPU (Graphics Processing Unit) and CPU (Central Processing Unit) instances, to facilitate optimal performance in essential GPU and CPU-based computing tasks. The system operates on a performance-based reward mechanism, where miners are incentivized through a tiered reward structure correlated to the processing capability of their hardware. High-performance devices are eligible for increased compensation, reflecting their greater contribution to the network’s computational throughput. Emphasizing the integration of GPU instances is critical due to their superior computational power, particularly in tasks demanding parallel processing capabilities. Consequently, miners utilizing GPU instances are positioned to receive substantially higher rewards compared to their CPU counterparts, in alignment with the greater processing power and efficiency GPUs bring to the network.</p><p>A key aspect of the miners’ contribution is the management of resource reservations. Miners have the autonomy to set specific timelines for each reservation of their computational resources. This timeline dictates the duration for which the resources are allocated to a particular task or user. Once the set timeline reaches its conclusion, the reservation automatically expires, thereby freeing up the resources for subsequent allocations. This mechanism ensures a dynamic and efficient distribution of computational power, catering to varying demands within the network.</p><h4>Hotkey Registration</h4><p>At this point, you will need some $TAO in your coldkey address for miner registration. Once your coldkey is funded, run the command below to register your hotkey:</p><pre>btcli s register --subtensor.network finney --netuid 27</pre><p>If you get the error ‘too many registrations this interval’ it means the maximum number of registrations that cycle has been reached, and you need to wait a bit and try again. You can check the registration cost <a href="https://taostats.io/subnets/netuid-27/#registration">here</a></p><h4>SETTING UP UFW FOR MINER</h4><p><strong>TCP Ports:</strong> Open ports using ufw (put any number in place of <strong>xxxx</strong> and <strong>yyyy</strong> of your choice) and use them as axon port:</p><pre>sudo apt update</pre><pre>sudo apt install ufw</pre><pre>sudo ufw allow xxxx:yyyy/tcp</pre><pre>sudo ufw allow 22/tcp</pre><pre>sudo ufw enable</pre><pre>sudo ufw status</pre><h4>RUNNING THE MINER</h4><p>Now, using pm2, run miner as:</p><pre>pm2 start ./neurons/miner.py --name MINER --interpreter python3 -- --netuid 27 --subtensor.network finney --wallet.name COLDKEYNAME --wallet.hotkey HOTKEYNAME --axon.port XXXX --axon.external_port XXXX --axon.ip xx.xxx.xxx.xx --axon.external_ip xx.xxx.xxx.xx --logging.debug --miner.blacklist.force_validator_permit --auto_update yes</pre><p>If you have a local subtensor (<strong>recommended</strong>):</p><pre>pm2 start ./neurons/miner.py --name MINER --interpreter python3 -- --netuid 27 --subtensor.network local --wallet.name COLDKEYNAME --wallet.hotkey HOTKEYNAME --axon.port xxxx --logging.debug --miner.blacklist.force_validator_permit --auto_update yes</pre><p>To set up your miner, first replace COLDKEYNAME &amp; HOTKEYNAME with the names of your keys. Then, update axon.portwith the 4-digit number you&#39;ve selected for <a href="https://medium.com/o/BZv7nWCzSG54TwpxPjrm/s/RCFZhMrQpz0DM7pJi6mq/products/subnet-27-compute/bittensor-compute-subnet-miner-setup#id-7.1-setting-up-ufw-for-miner">xxxx above</a>. For the parameters --axon.external and --axon.ip, use your miner machine&#39;s public IP address in place of the &#39;xxxxx&#39;s. You can find this IP by running <strong>hostname -I</strong>. Though not always necessary, these parameters can be crucial for resolving certain connectivity issues.</p><p>When operating a miner and you have local subtensor running on a separate machine, it&#39;s crucial to add and adjust the --subtensor.chain_endpoint parameter. This should be set to the IP and port (XXX.XX.XXX.XXX:XXXX) where your subtensor is running. This parameter can be removed if your subtensor is local to the miner machine.</p><h4>CHECKING MINER LOGS</h4><p>After launching the compute miner, you can then check the logs using the two commands below:</p><pre>pm2 logs</pre><pre>pm2 monit</pre><p>Run pm2 logs and wait to see incoming HTTP traffic. Ensure you are receiving challenges and then finding them.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*J1AQfON2nakIap6_Sy4mqQ.png" /><figcaption><strong><em>Happy mining </em></strong>❤️<strong><em> Dont forget to update </em></strong>😄</figcaption></figure><h4>BENCHMARKING THE MACHINE</h4><pre>hashcat -b -m 610</pre><p>Output</p><pre>Speed.#1.........: 12576.1 MH/s (75.69ms) @ Accel:8 Loops:1024 Thr:1024 Vec:1<br>Speed.#2.........: 12576.1 MH/s (75.69ms) @ Accel:8 Loops:1024 Thr:1024 Vec:1<br>...<br>...</pre><p><strong>The recommended minimum hashrate for the current difficulty is &gt;= 4500 MH/s.</strong></p><p>Difficulty will increase over time.</p><h4>MORE USEFUL COMMANDS</h4><pre>btcli s metagraph --netuid 27</pre><pre>btcli s list</pre><pre>btcli wallet overview --subtensor.network finney --all --netuid 27</pre><pre>pm2 logs miner --lines 1000 | grep -i &quot;Challenge.*found&quot;</pre><pre>pm2 logs -f | grep -E &quot;SUCCESS|INFO|DEBUG|ERROR&quot;</pre><pre>nvidia-smi --query-gpu=name,memory.total,clocks.gr,clocks.mem --format=csv</pre><pre>grep &quot;Challenge .* found in&quot; &quot;/home/ubuntu/.pm2/logs/MINER-out.log&quot; | sed -E &#39;s/.* found in ([0-9.]+) seconds.*/\\1/&#39; | awk &#39;{sum+=$1; count+=1} END {if (count &gt; 0) print sum/count; else print &quot;No data to calculate average&quot;}’</pre><pre>btcli w transfer --subtensor.network local --dest DESTINATION_WALLET --wallet.name default --amount 0</pre><pre>btcli stake remove --subtensor.network local --all --all_hotkeys --wallet.name default</pre><h3>Decentralizing Compute</h3><p>With Subnet 27, we decentralize compute and the people’s right to access compute. Ten-figure VC-funded AI companies are swallowing up the GPU supply while advocating for computing regulations. It looks like we must stand for our right to compute.</p><p>An incentivized and permission-less compute market is priceless. Your contribution means less power to the AI oligopoly and more to the collective. Join us as we democratize compute and AGI, bring your own GPUs.</p><p>Subnet 27 is live; come and take it!</p><blockquote>Neural Internet is an AI collective pushing the boundaries of machine intelligence. We are building the next generation of AI applications through distributed solutions.</blockquote><p><em>Read our documentations </em><a href="https://docs.neuralinternet.ai/"><em>here</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=82498b93e7e1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Subnet 27: Incentivizing Compute]]></title>
            <link>https://medium.com/@neuralinternet/subnet-27-decentralizing-compute-with-bittensor-8ee68c9af2ff?source=rss-a261c76f5cd9------2</link>
            <guid isPermaLink="false">https://medium.com/p/8ee68c9af2ff</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[decentralization]]></category>
            <category><![CDATA[bittensor]]></category>
            <dc:creator><![CDATA[Neural Internet]]></dc:creator>
            <pubDate>Mon, 20 Nov 2023 01:36:35 GMT</pubDate>
            <atom:updated>2023-11-20T23:33:23.486Z</atom:updated>
            <content:encoded><![CDATA[<h4><strong>Bittensor’s Layer-0</strong>: Permission-less Compute Market</h4><p>Introducing Subnet 27: <strong>Bittensor’s Layer-0, </strong>the permission-less compute market for platform composable cloud services. An abundant era in decentralized, efficient and scalable computing has begun.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*xt6pvMMNTsquVP9G9gGFqA.png" /><figcaption>Subnetwork 27</figcaption></figure><p>Compute markets of today consist of siloed pools of compute resources separated by each provider and its designated products to access each pool of compute. Subnet 27 decentralizes computing resources by combining siloed pools of compute on a blockchain to be validated and accessed trustlessly. This opens a door to efficient and scalable compute without the constraints of a centralized power.</p><h4>Incentivizing Compute</h4><p>Decentralized cloud computing is an innovative approach that leverages a peer-to-peer network to distribute computational resources across multiple nodes, rather than relying on centralized data centers. With the proliferation of cloud platforms, there’s a need for a Bittensor subnet that can seamlessly integrate these platforms, allowing for efficient resource sharing and allocation.</p><p><strong>Computational resource management:</strong> Subnet 27 revolutionizes the distribution and control of computational power. This innovative model democratizes access, allowing not just major corporations but also individual developers and small enterprises to equally benefit from and contribute to the network. This shift breaks the traditional monopoly of big tech companies, fostering a more inclusive technological landscape.</p><p>Importantly, it is powered by the Yuma Consensus where miners are incentivized with $TAO to provide compute to the network. This promotes participation and fosters innovation, allowing startups and researchers to access and provide computational resources affordably. This not only encourages technological experimentation but also ensures the network’s continuous growth and sustenance.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*j-04n-gYW1zLd06TdmmBdA.png" /><figcaption>Siloed vs. Decentralized</figcaption></figure><h4>Resilience to Regulation and Censorship</h4><p>In a centralized system, a single entity or a limited number of entities control the computing resources. This makes it easier for governments or regulatory bodies to impose restrictions or censorship, as they only need to target a few centralized points.</p><p>In contrast, a decentralized network distributes computing resources across many independent nodes, often globally. This dispersion makes it much harder for any single regulatory authority to exert control or enforce censorship, as they would have to coordinate across multiple jurisdictions and deal with a multitude of independent nodes.</p><p>With the very real possibility of governments regulating GPUs usage, now is the perfect time for a decentralized compute market. These political moves coupled with a shortage of GPUs in the market hinder the collective when it comes to AI/ML development and access. Decentralized AI needs a decentralized compute layer.</p><h3>Subnet 27 Overview:</h3><p><strong>Compute-composable subnet: </strong>Integrated into various cloud platforms (e.g., Runpod, Lambda, AWS) Subnet 27 becomes a cohesive unit, enabling higher-level cloud platforms to offer seamless compute composability across different underlying platforms. This compute-composable subnet will enable nodes to contribute computational power, with validators ensuring the integrity and efficiency of the shared resources.</p><h3>Miners</h3><p>Miners contribute processing resources, notably GPU (Graphics Processing Unit) and CPU (Central Processing Unit) instances, to facilitate optimal performance in essential GPU and CPU-based computing tasks.</p><p><strong>Dynamic Reservation Timelines:</strong> A key aspect of the miners’ contribution is the management of resource reservations.</p><ul><li>Miners have the autonomy to set specific timelines for each reservation of their computational resources.</li><li>This timeline dictates the duration for which the resources are allocated to a particular task or user.</li><li>Once the set timeline reaches its conclusion, the reservation automatically expires, thereby freeing up the resources for subsequent allocations.</li></ul><p>This mechanism ensures a dynamic and efficient distribution of computational power, catering to varying demands within the network.</p><h3>Validators</h3><p>Validators hold the critical responsibility of rigorously assessing and verifying the computational capabilities of miners. This multifaceted evaluation process commences with validators requesting miners to provide comprehensive performance data.</p><p><strong>Compute Data:</strong> The inclusion of RAM and disk space measurements is vital, as these components significantly impact the overall performance and reliability of the miners’ hardware. RAM capacity influences the ability to handle large or multiple tasks simultaneously, while adequate disk space ensures sufficient storage.</p><p><strong>Computational Integrity: </strong>Following the receipt of this detailed hardware and performance information, validators proceed to test the miners’ computational integrity. This is achieved by presenting them with complex hashing challenges, designed to evaluate the processing power and reliability of the miners’ systems. Validators adjust the difficulty of these problems based on the comprehensive performance profile of each miner, including their RAM and disk space metrics.</p><p>In addition to measuring the time taken by miners to resolve these problems, validators meticulously verify the accuracy of the responses. This thorough examination of both speed and precision, complemented by the assessment of RAM and disk space utilization, forms the crux of the evaluation process.</p><p><strong>Scoring Mechanism: </strong>Based on this extensive analysis, validators update the miners’ scores, reflecting a holistic view of their computational capacity, efficiency, and hardware quality. This score then determines the miner’s weight within the network, directly influencing their potential rewards and standing.</p><p><strong>Github Link: </strong><a href="https://github.com/neuralinternet/compute-subnet"><strong>https://github.com/neuralinternet/compute-subnet</strong></a></p><p><strong>Technical Docs:</strong> <a href="https://docs.neuralinternet.ai/products/subnet-27-compute">https://docs.neuralinternet.ai/products/subnet-27-compute</a></p><p><strong>Official Subnet 27 Channel: </strong><a href="https://discord.gg/bittensor">https://discord.gg/bittensor</a></p><h3>Subnet 27: A New Resource for AI Development</h3><p><strong>Decentralized Compute:</strong> Subnet 27 brings an entirely new resource to Bittensor, compute. Arguably the most important and finite resource needed for the creation of machine intelligence. All network participants will have access to an ever-expanding pool of compute for all development needs.</p><p>With a decentralized compute subnet plugged into Bittensor, we become ungovernable. Come and Take it.</p><p>Join the community for more updates: <a href="https://discord.gg/neuralinternet">https://discord.gg/neuralinternet</a></p><p><strong>Development Updates:</strong> Incentives delegated from our validator to Subnet 27 have begun today as we continue testing with the community. Once we’ve finalized initial testing and confirmed the incentive mechanism works as intended, the allocation from our validator’s emissions to the subnet will begin to increase. More updates coming this week.</p><p>Neural Internet is an AI research collective empowering humanity through collaborative machine learning. We are building intelligent solutions with distributed technologies for a safer and more equitable digital world. Powered by open-source and open ownership machine intelligence, we /imagine a future where AI is decentralized and in harmony with humanity.</p><p><em>Read our documentations </em><a href="https://docs.neuralinternet.ai/introduction/overview"><em>here</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8ee68c9af2ff" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Kickstart a Future: Shoe Drive]]></title>
            <link>https://medium.com/@neuralinternet/kickstart-a-future-shoe-drive-a90a79c9696c?source=rss-a261c76f5cd9------2</link>
            <guid isPermaLink="false">https://medium.com/p/a90a79c9696c</guid>
            <category><![CDATA[nonprofit]]></category>
            <dc:creator><![CDATA[Neural Internet]]></dc:creator>
            <pubDate>Mon, 14 Aug 2023 17:41:02 GMT</pubDate>
            <atom:updated>2023-08-14T17:41:02.730Z</atom:updated>
            <content:encoded><![CDATA[<h4>Peyton Barber Foundation x Clothes to Kids Shoe Drive</h4><p>We at Neural Internet are both delighted and humbled to be working with the <a href="https://www.peytonbarberfoundation.com/"><strong>Peyton Barber Foundation</strong></a> and Clothes to Kids for the upcoming Back to School Shoe Drive. We recognize the simple yet profound joy that a child can experience when slipping on a pair of shoes that fit just right — a joy that is not a given for everyone.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cZNNcid31wV9JyF6_rtlfg.png" /><figcaption>Peyton Barber Foundation</figcaption></figure><h4>Kickstart a Future: Back to School Shoe Drive</h4><blockquote><strong>Dates: </strong>August 1st to August 31st</blockquote><blockquote><strong>Business Drop-Off Locations in Florida, USA: </strong>Clearwater, St Pete, Tampa, Lakeland, Orlando, Daytona Beach</blockquote><p>Today, we invite you to make a difference that will resonate through the lives of the children we serve. Together, we can stride towards a brighter future, gifting empowerment and nurturing dreams, one pair of shoes at a time.</p><h4>Clothes to Kids</h4><p>Each year, Clothes to Kids contributes over 13,000 pairs of shoes to children who need them the most. Through our collaboration with the Peyton Barber Foundation and Clothes to Kids, we hope to augment these efforts significantly, fostering positive change within the Tampa Bay, Florida community in this school season.</p><h4>Calling all Local Businesses</h4><p>We call upon local businesses, community members, and individuals to aid us in our mission. We welcome shoes of all brands, whether new or gently used, in sizes from Toddler (9C) to Men (16Men). Each pair you donate will play a pivotal role in instilling confidence and self-esteem in the children who will be our leaders tomorrow.</p><p>View all drop-off locations here: <a href="https://www.peytonbarberfoundation.com/events/shoe-drive">https://www.peytonbarberfoundation.com/events/shoe-drive</a></p><h4>Monetary Donations</h4><p>If shoe donations aren’t feasible, please consider making a monetary contribution. These funds will be directed towards the purchase of shoes for those in need. Every $15 dollars will get a kid a new pair of shoes for this new school year. Corporations are also encouraged to lend their support to this meaningful cause through sponsorship or donations.</p><p><a href="https://checkout.stripe.com/c/pay/cs_live_a1Vc9GjyC8vR68ME3o370BEJ7uZU5J0AcLHdB5sRxZgL4Em8ZsqUQ88hcD#fidkdWxOYHwnPyd1blppbHNgWjA0S0l3djdCTm1SNFNTNzdLSmxCMEJsPFZAQVZQNnRxc0JPTjFLTUpOVmxdfXdMTGNVNG02PGo2V1NuSVdLQUxgS1dSUFBSZGRAcE5ucXx%2FRHxoTlVcN3ZwNTU2VmE2R2B3TScpJ3VpbGtuQH11anZgYUxhJz8nPXJINTV0YlBQPWdnZ2ByMWJiJyknd2BjYHd3YHdKd2xibGsnPydtcXF1dj8qKm92K3Zxd2x1YCtmamgqJykndXdgY2xpaWBhRGhqcGtxJz80NTU1NSknaWpmZGlgJz8nYGsnKSdgaGdgYVZqcHdmYCc%2FJ2dwfFpncWtaNEtWam8yQk5tUjRTUzc3S1xUQjR9UlBMJ3gl"><strong>Donate directly through Stripe</strong></a></p><p>The Peyton Barber Foundation, our esteemed partner in this endeavor, is a registered 501(c)3 non-profit organization, making your contributions tax-deductible as per law. Join us in making strides towards a better, more equitable future for our children.</p><p><em>Neural Inτerneτ is an AI R&amp;D DAO maximizing the potential of the Biττensor network, an incentivized open-source protocol that powers a scalable, decentralized neural network. Neural Internet is committed to promoting innovation in AI with decentralization and transparency at the forefront.</em></p><p><em>Read our documentations </em><a href="https://docs.neuralinternet.ai/introduction/overview"><em>here</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a90a79c9696c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Hi, my name is replyτensor…]]></title>
            <link>https://medium.com/@neuralinternet/hi-my-name-is-reply%CF%84ensor-21a78aecd810?source=rss-a261c76f5cd9------2</link>
            <guid isPermaLink="false">https://medium.com/p/21a78aecd810</guid>
            <category><![CDATA[chatbots]]></category>
            <category><![CDATA[twitterbots]]></category>
            <category><![CDATA[bittensor]]></category>
            <category><![CDATA[large-language-models]]></category>
            <dc:creator><![CDATA[Neural Internet]]></dc:creator>
            <pubDate>Wed, 28 Jun 2023 14:12:02 GMT</pubDate>
            <atom:updated>2023-06-28T14:12:02.456Z</atom:updated>
            <content:encoded><![CDATA[<h4>An interactive reply bot automated by Neural Internet</h4><p>ReplyTensor, the innovative reply bot developed by Neural Internet, is making waves on popular platforms such as Twitter and Telegram. Not merely a bot, ReplyTensor is the physical embodiment of Bittensor’s Prompting Network (netuid 1), its purpose being to engage users via replies to tagged tweets and messages.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MqNpUPRAiylQDfYY_7jYcQ.jpeg" /><figcaption>@replytensor</figcaption></figure><p><strong>Twitter Account:</strong> <a href="https://twitter.com/replytensor">https://twitter.com/replytensor</a></p><p>Being a work in progress, ReplyTensor continually receives updates and improvements, thereby enhancing the overall user experience.</p><h4><strong>Twitter</strong></h4><p>Starting a conversation with ReplyTensor on Twitter is as easy as tagging it in a tweet. Once tagged, @ReplyTensor jumps into action, replying to the tagged tweet within Twitter’s API restriction of 280 characters. However, there’s good news on the horizon — once API character restrictions are lifted, ReplyTensor will be equipped to respond with long-form replies.</p><h4>Telegram</h4><p>ReplyTensor’s functionalities are not limited to Twitter; it has made its presence known on Telegram too. To initiate a conversation, simply navigate to the $TAO Community Telegram at <a href="https://t.me/taobittensor">https://t.me/taobittensor</a> to start a conversation.</p><p>ReplyTensor can also be integrated into your own Telegram group:</p><ol><li>Click on ‘Add Member’ in your Telegram group.</li><li>In the ‘Add member’ field, type “replytensor”.</li><li>Begin a conversation by typing “/chat”.</li></ol><p>As with any technology in the nascent stages, Telegram’s ReplyTensor version is undergoing minor bug fixes to optimize user experience.</p><h4>The Future of ReplyTensor</h4><p>The bot is planned to connect to Neural Internet’s Validator API, which will yield better results and enhanced capabilities. Research and development is taking place to create a pipeline for custom data training and integrations allowing a wider range of use cases, from customer service to educational training chatbots.</p><h4>Community</h4><p>At Neural Internet, we greatly appreciate your ongoing enthusiasm and engagement with ReplyTensor. Your feedback plays a pivotal role in shaping the bot’s evolution. If you encounter any issues or require assistance with the bot, we encourage you to reach out to us via Twitter or Telegram DMs.</p><p>We would also like to extend our heartfelt gratitude to @TAO_Validator for their initial support and to every user who has interacted with ReplyTensor. Your active participation is driving the future of Bittensor and decentralized AI.</p><h4>Stay tuned for the evolution of ReplyTensor!</h4><p><em>Neural Inτerneτ is an AI R&amp;D DAO maximizing the potential of the Biττensor network, an incentivized open-source protocol that powers a scalable, decentralized neural network. Neural Internet is committed to promoting innovation in AI with decentralization and transparency at the forefront.</em></p><p><em>Read our documentations </em><a href="https://docs.neuralinternet.ai/introduction/overview"><em>here</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=21a78aecd810" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[DGX H100 Super Computer]]></title>
            <link>https://medium.com/@neuralinternet/dgx-h100-super-computer-754b86e852ae?source=rss-a261c76f5cd9------2</link>
            <guid isPermaLink="false">https://medium.com/p/754b86e852ae</guid>
            <category><![CDATA[nvidia]]></category>
            <category><![CDATA[ai-development]]></category>
            <category><![CDATA[dgx-h100]]></category>
            <category><![CDATA[supercomputer]]></category>
            <category><![CDATA[neural-internet]]></category>
            <dc:creator><![CDATA[Neural Internet]]></dc:creator>
            <pubDate>Thu, 18 May 2023 12:27:22 GMT</pubDate>
            <atom:updated>2023-05-18T13:23:58.137Z</atom:updated>
            <content:encoded><![CDATA[<h4>Unleashing the Power of AI with the NVIDIA DGX H100</h4><p>Artificial Intelligence is changing the world, ushering in a new era of innovation and optimization across diverse industries. However, AI’s sophisticated processes require massive computing power, particularly for complex and large-scale tasks. NVIDIA has risen to this challenge by developing the DGX H100, a powerful system that can handle some of the most demanding AI workloads.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wkA7SH_3farMlU1nVzKVzQ.jpeg" /><figcaption>NVIDIA DGX H100</figcaption></figure><h4>What is the DGX H100?</h4><p>The DGX H100, part of the NVIDIA DGX platform, stands as a fully integrated hardware and software solution specifically engineered for enterprises aiming to establish their own AI centers. The DGX platform is a family of systems that have been meticulously designed to cater to the needs of AI research and development.</p><p>In its latest and most advanced iteration, the DGX H100 incorporates the NVIDIA H100 Tensor Core GPU. This new chip not only offers unprecedented performance but also provides superior scalability and security for AI applications, setting a new industry standard.</p><p>Depicted as the epitome of performance, scalability, and ease of use, the DGX H100 serves as an ideal choice for enterprise-level AI. It is essentially a supercomputer, specifically designed for AI applications. Acting as a powerhouse of computing, it is tailored to handle the most demanding AI tasks, thus demonstrating its capabilities as a comprehensive and highly efficient solution for AI-centric enterprises.</p><h4>What can the DGX H100 do?</h4><p>The DGX H100 is a powerful AI system capable of handling demanding tasks such as generative AI, natural language processing, and deep learning recommendation models. Its prowess lies in its ability to train large language models (LLMs) which require significant data and computational resources.</p><p>The DGX H100 enhances the training of LLMs up to 30 times faster than its predecessors, made possible by its innovative features. These include the Transformer Engine for speedy matrix multiplications, the NVLink Switch System for seamless data sharing and collaboration across multiple GPUs, and the FP8 precision for improved numerical accuracy with less memory and bandwidth usage.</p><p>Moreover, the DGX H100 can effectively train LLMs with up to 395 billion parameters using the mixture-of-experts (MoE) technique. This process divides the model into smaller sub-models, each specializing in different task aspects, achieving superior performance and efficiency.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*R1-1Ivnh6PT7_wv3BErKsg.jpeg" /><figcaption>GPU Relative Comparisons</figcaption></figure><h4>Conclusion</h4><p>The NVIDIA DGX H100 is a fully integrated hardware and software solution, revolutionizing the way enterprises can utilize AI. As a platform for AI innovation and optimization, it offers unprecedented performance, scalability, and security for AI tasks, enabling business innovation, workflow optimization, and the construction of AI-ready infrastructures.</p><p>Capable of processing vast data amounts at remarkable speeds, the DGX H100 excels in complex AI tasks such as deep learning, machine learning, and data analysis. It’s particularly suited for tasks like image and speech recognition, natural language processing, and autonomous vehicle technology.</p><p>The DGX H100 allows researchers and developers to train AI models rapidly, accelerating the development and improvement of AI technologies. This paves the way for advancements in fields such as healthcare, finance, and transportation.</p><p>In summary, the NVIDIA DGX H100 is a specialized supercomputer for AI tasks, capable of handling demanding AI workloads. It’s an essential tool for researchers, developers, and tech enthusiasts alike, marking a significant leap forward in the world of AI.</p><p><em>Neural Inτerneτ is an AI R&amp;D DAO maximizing the potential of the Biττensor network, an incentivized open-source protocol that powers a scalable, decentralized neural network. Neural Internet is committed to promoting innovation in AI with decentralization and transparency at the forefront.</em></p><p><em>Read our documentations </em><a href="https://docs.neuralinternet.ai/introduction/overview"><em>here</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=754b86e852ae" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Collaborative Future of AI]]></title>
            <link>https://medium.com/@neuralinternet/the-collaborative-future-of-ai-d807378f994?source=rss-a261c76f5cd9------2</link>
            <guid isPermaLink="false">https://medium.com/p/d807378f994</guid>
            <category><![CDATA[open-source]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[tensorflow]]></category>
            <category><![CDATA[collaboration]]></category>
            <category><![CDATA[bittensor]]></category>
            <dc:creator><![CDATA[Neural Internet]]></dc:creator>
            <pubDate>Sun, 14 May 2023 06:53:02 GMT</pubDate>
            <atom:updated>2023-05-14T06:53:02.002Z</atom:updated>
            <content:encoded><![CDATA[<h4>How Open-Source Ecosystems are Shaping AI Development</h4><p>In the fast-paced world of AI development, the significance of open-source ecosystems is indisputable. Open-source software (OSS) has rapidly grown in importance, driving a wave of innovation and collaboration that is transforming the AI industry. This article explores how leveraging open-source ecosystems can elevate AI development to unprecedented heights.</p><h4>The Open-Source Revolution in AI</h4><p>Open-source revolution represents technology democratization. It involves the sharing of software’s source code freely, permitting anyone to view, modify, and distribute the software. This paradigm shift has proven instrumental in AI development, propelling innovation, collaboration, and transparency.</p><h4>The Impetus Behind Open-Source AI</h4><p>Open-source ecosystems cultivate a community-driven approach, enabling AI developers to draw on collective expertise. It accelerates innovation, as developers can build upon existing code rather than reinventing the wheel. Most importantly, Open-source ecosystems fosters transparency, critical for the ethical development and application of AI.</p><p>From startups to tech giants, organizations recognize the value and understand that embracing an open-source culture can expedite their journey towards AI maturity.</p><h4>Harnessing the Power of Open-Source Ecosystems</h4><p>Leveraging open-source ecosystems in AI development involves:</p><ol><li><strong>Embracing Collaboration:</strong> Open-source is centered around community engagement, contribution, and shared insights to accelerate AI advancement.</li><li><strong>Incentivizing Learning:</strong> Open-source platforms offer a wealth of learning resources that keep AI developers updated with latest trends and innovations.</li><li><strong>Promoting Transparency:</strong> Sharing code, methodologies, and findings in the open-source community ensures ethical AI practices and builds trust in AI systems.</li><li><strong>Driving Innovation::</strong> The open-source culture sparks innovation through iterative improvements, keeping organizations at the cutting edge of AI development.</li></ol><h4>Bittensor: The Future of Distributed AI Networks</h4><p>In the open-source realm, Bittensor stands as an innovative platform that exemplifies the potential of open-source ecosystems in AI development.</p><p>Bittensor is an open-source protocol underpinning a globally distributed, decentralized neural network. It produces machine intelligence by incentivizing the training of models within its distributed infrastructure, rewarding insightful data analysis with a unique digital currency.</p><p>This ground-breaking protocol signifies an evolution in the AI development process. By decentralizing the neural network, Bittensor ensures that AI growth is not confined to a central authority. Instead, it unfolds across a global network, fostering a diverse, inclusive, and resilient AI ecosystem.</p><p>Moreover, the concept of rewarding contributors with a digital currency goes beyond mere incentivization. It establishes a dynamic where data analysis and the creation of insights are not just rewarded but are acknowledged as valuable commodities within the AI development framework.</p><p>Bittensor embodies the ethos of the open-source movement — collaboration, transparency, and democratization. It sets a new precedent for the future of AI development, demonstrating how open-source ecosystems can be harnessed to drive innovation, decentralization, and inclusivity in the world of AI.</p><h4>The Way Forward</h4><p>In the digital age, open-source ecosystems stand as pillars of innovation, collaboration, and transparency. By utilizing these ecosystems, we can transform AI development and tackle its complex challenges. The transformative power of open-source can democratize AI, ensuring it’s a technology that benefits all.</p><p>As we push the boundaries of AI, we must foster an environment that encourages collective intelligence. Let’s harness the power of open-source ecosystems and shape an open, inclusive, and innovative future for AI.</p><p><em>Neural Inτerneτ is an AI R&amp;D DAO maximizing the potential of the Biττensor network, an incentivized open-source protocol that powers a scalable, decentralized neural network. Neural Internet is committed to promoting innovation in AI with decentralization and transparency at the forefront.</em></p><p><em>Read our documentations </em><a href="https://docs.neuralinternet.ai/introduction/overview"><em>here</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d807378f994" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>