Hash Rate pod - Bitcoin, AI, DePIN, DeFi
July 10, 2025

Hash Rate - Ep 120 - SCORE Computer Vision (Subnet 44)

Max Sebi of SCORE (Subnet 44) joins the pod to break down how his team is weaponizing computer vision to build a revenue-generating business on BitTensor, starting with a Moneyball-style approach to professional soccer.

A Go-to-Market Blueprint for AI

  • “The main problem in sports at the moment is on the data annotation end. It's very hard for people to upload a video somewhere online and get data notation in a very accurate, fast, and affordable way.”

SCORE is building the "optic nerve of AI" by tackling computer vision, but its genius lies in its laser-focused market entry. Instead of boiling the ocean, they’re providing analytics to sports teams and hedge funds. Their three-step process—annotating video data (tracking), understanding events with neural nets, and predicting future frames—creates new, proprietary datasets. This allows them to:

  • Target the Long Tail: While top clubs use pricey systems from companies like Palantir, SCORE offers its service at 1/100th of the cost, making advanced scouting accessible to lower-tier leagues.
  • Create New Efficiencies: They automate the manual, human-centric process of scouting, allowing teams to analyze vast amounts of video to find undervalued players or "gems."
  • Generate Early Revenue: SCORE has a deal with a sports hedge fund, taking 20% of the upside generated. Revenue will be used for buybacks of their subnet token, directly rewarding the ecosystem.

The Unbeatable Economics of BitTensor

  • “How could you kickstart something like SCORE, literally saving half a million a month in compute and salaries... and then just move away from that? It's crazy to move away from that just from a business perspective.”

The core debate around BitTensor is whether successful subnets will eventually ditch the network and go private. Max argues the economic incentives make this a "dumb move."

  • Massive Cost Savings: SCORE saves an estimated $6 million per year on compute and R&D salaries by leveraging BitTensor’s decentralized network of miners.
  • A Perpetual Innovation Engine: The network provides a "proof of intelligence"—a globally distributed team of 256 miners constantly competing to improve SCORE’s models to earn TAO. Abandoning the network means losing this free, state-of-the-art research lab.
  • Aligned Incentives: SCORE provides investors with both equity and token warrants, ensuring the company’s success is intrinsically tied to the health and value of its subnet.

Navigating the DTO Arena

  • “The pre-DTO world was super easy... Now the game is very complex because you have to be a legit AI team, you have to be good at marketing, and you have to go on podcasts and explain yourself.”

The Dynamic TAO (DTO) system has transformed BitTensor into a high-stakes competitive arena. It forces subnets to not only build superior technology but also to market it effectively to attract staked TAO.

  • An Internal Capital Market: DTO acts as a launchpad, forcing subnets to compete for community capital and attention, creating a more dynamic and meritocratic ecosystem.
  • Fueling Ecosystem Growth: This new complexity is attracting serious attention from VCs and hedge funds, who view TAO as an early Bitcoin-like opportunity and are actively looking for ways to get exposure to high-potential subnets.

Key Takeaways:

  • BitTensor isn't just an abstract AI research project; it's a practical economic engine for building real-world businesses. The model provides an unparalleled launchpad for startups by replacing massive upfront capital expenditure with a powerful incentive mechanism.
  • BitTensor is a VC alternative. The network provides startups like SCORE with millions in free compute and R&D, allowing them to compete with giants by replacing venture funding with token incentives.
  • Revenue is the ultimate metric. In the post-DTO world, subnets that can demonstrate a clear path to revenue and token buybacks, like SCORE, are positioned to attract significant capital.
  • The economic moat is real. The argument that subnets will "go private" ignores the immense, ongoing value of a free, decentralized AI research lab that constantly keeps them at the bleeding edge.

For further insights and detailed discussions, watch the full podcast: Link

This episode reveals how a BitTensor subnet is building a real-world, revenue-generating business by targeting the sports analytics market, proving the tangible economic power of decentralized AI.

Introduction: The Optic Nerve of AI

  • Host Mark Jeffrey welcomes Max Sebi of Score Computer Vision (Subnet 44), a project focused on solving computer vision challenges.
  • Mark highlights Score's unique go-to-market strategy: instead of tackling the broad, abstract challenge of computer vision, they are "crossing the chasm" by focusing on a specific, revenue-generating niche in sports intelligence.

The Core Technology: A Three-Step Process for Video Analysis

  • Max outlines Score's three-step technical process for analyzing video footage, a workflow applicable to any industry leveraging video data.
  • Step 1: Annotation (Tracking): This is the foundational layer. Max defines annotation as the process of assigning specific coordinates to objects within a video. For example, tracking a player's movement from one side of the field to the other or pinpointing the exact location of the ball frame by frame.
  • Step 2: Event Understanding: After tracking, neural networks are used to interpret the tracked data and identify significant events, such as a player scoring a goal or sustaining an injury. This moves beyond simple coordinates to contextual understanding.
  • Step 3: Predictive Inference: By analyzing historical annotated data, the system learns to predict future frames in real-time. Max explains, "Every time an object is moving frame by frame, you can deduct what's what's likely to happen." This capability is critical for applications in sports, robotics, and autonomous vehicles.

The Business Case: Scouting, Hedge Funds, and New Data Sets

  • Max details Score's primary customer segments, explaining how the technology provides distinct value to each.
  • Sports Teams: The main use case is scouting. The system automates the analysis of vast amounts of video, making player recruitment far more efficient. It allows teams to score players, systems, and formations, providing a powerful management tool both for scouting and in-game analysis.
  • Sports Hedge Funds & Betting Syndicates: These clients are seeking a competitive edge. Score provides this by creating entirely new, proprietary data sets derived from its deep video analysis. The ability to process real-time data also opens up opportunities for sophisticated in-game betting strategies.

Democratizing Sports Analytics for the Long Tail

  • Max presents a compelling analysis of the current market asymmetry, where only top-tier teams with massive budgets can afford advanced analytics, which are often manual and human-intensive.
  • Top clubs spend between $1-3 million annually on scouting data. Max states that Score can provide a superior, automated solution at a fraction of the cost: "If you think about something like score, we would do it at like a hundredth of the price."
  • This cost advantage is because Score is automating a human-centric process, not just competing with other automated solutions.
  • Strategic Implication: Score's core thesis is to address the "long tail"—the vast market of smaller clubs, semi-pro teams, and even individuals who lack the resources for expensive computer vision tools. This strategy could be a blueprint for other subnets seeking product-market fit.

Revenue, Market Size, and Go-to-Market Focus

  • While not disclosing specific figures, Max confirms Score is generating revenue. Their initial deal with a sports hedge fund involves a 20% share of the upside generated, with revenue to be demonstrated to the community via token buybacks.
  • He clarifies their Total Addressable Market (TAM) is not the entire $600 billion football industry, but a more focused $1.5 billion market for sports analytics and data.
  • Max emphasizes a disciplined approach, focusing on the sports market for the next six months to build a solid revenue foundation before exploring other industries like the broader data annotation market, where they could compete with giants like Scale AI.

Addressing Enterprise Concerns: Data Privacy and Centralization

  • Mark raises a critical concern for enterprise adoption of BitTensor: the anonymity of miners who process the data.
  • Max explains that Score proactively solved this by partnering with a key, trusted data provider, 6TV. This ensures all data is sourced cleanly and ethically, providing customers with confidence and sidestepping the "anonymous miner" problem.
  • While this approach is less decentralized initially, it was a necessary strategic trade-off to build trust and secure enterprise partners.

The Economic Moat: Why Subnets Won't Abandon BitTensor

  • Mark challenges Max on a common critique: what prevents a successful subnet from abandoning its token and miners to operate as a centralized company?
  • Max provides a powerful, business-driven rebuttal, arguing it would be strategically foolish.
    • Cost Savings: Score saves approximately $500,000 per month in compute and salary costs by leveraging the BitTensor network. Abandoning this would be financially crippling.
    • Continuous Innovation: The 256 miners on the subnet are a global, decentralized R&D team constantly working to improve the models. Losing this competitive edge would leave the company vulnerable to disruption from centralized players like OpenAI.
  • Max's perspective is grounded in pragmatism, not just ideology: "We would be crazy not to think about it this way because how could you like kickstart something like score literally save half a million a month in compute and you know salaries... and then just move away?"

Analyzing BitTensor's DTO System

  • The conversation shifts to DTO (Dynamic TAO Allocation), BitTensor's mechanism where TAO is staked for subnet tokens, directing network rewards.
  • Max offers a pragmatic view:
    • Positive: The system is a powerful engine for attracting attention and liquidity, creating an exciting, competitive environment that draws in users from other ecosystems.
    • Negative: It adds a significant layer of complexity to an already complex system, making it harder to explain BitTensor to newcomers.
  • Mark adds that the DTO system cleverly drives value to the core TAO token and creates a level playing field, as the liquidity mechanism prevents insiders from having an unfair advantage.

Subnet Composability: A Collaborative Ecosystem

  • Max reveals that Score actively uses other subnets, partnering with Macrocosmos (Subnet 13) for data and planning to share revenue back with them.
  • He champions the idea of "subnet composability," where subnets collaborate and build on each other's services, creating an integrated and more powerful ecosystem.
  • Actionable Insight: This trend of inter-subnet collaboration is a key indicator of the network's maturing health. Investors should watch for these partnerships as they create positive-sum dynamics that strengthen the entire BitTensor ecosystem.

Conclusion

  • Score's journey illustrates a powerful blueprint for success on BitTensor: target a niche market, build a real business, and leverage the network's unique economic incentives for an unbeatable competitive advantage. For investors, this highlights the immense value in subnets that combine technical innovation with a clear, pragmatic go-to-market strategy.

Others You May Like