TL;DR: AI-powered productivity, or "Botpower," promises infinite output, but what happens when we push it to its limits? Flawed data becomes amplified chaos, judgment-free execution produces bugs and waste at scale, and complex problems create errors and confusion. Add in massive energy costs and overstretched human oversight, and it’s clear: Botpower has issues.  Understanding them reveals where human ingenuity is still absolutely needed. Every action has its pleasures and its price - Socrates Introduction: The Other Side of Infinite We’ve written about Botpower  before – the shift from horsepower to manpower to AI-driven productivity that now defies traditional limits. First, we introduced the concept:  artificial intelligence as a multiplier that rewrites the rules of output. Then, we tried (somewhat optimistically) to quantify it with a Botpower equation , because humans love measurements. And here we are now: Botpower is happening . AI tools are generating code, diagnosing diseases, designing ads, and producing content faster than we can process it. The productivity ceiling isn’t just higher –  it is disappearing. But every system, no matter how powerful, has a breaking point. If you push AI hard enough and scale it wide enough, you start to see critical cracks. Bias becomes exponential. Small errors snowball into system-wide failures. Complexity overwhelms logic. So before we get enamored with the promises of Botpower, let’s stress-test it. What happens when AI is pushed to its limits? Where does it stumble, and what does that teach us about its potential and its risks? Garbage In, Catastrophe Out: Scaling Bad Inputs AI is not magic.  It is a machine trained on data. That’s it. And while good data yields brilliant results, bad data scales brilliantly bad ones. A  study published in Oxford Review of Economic Policy  discusses how AI’s reliance on flawed inputs can lead to amplified consequences, significantly impacting productivity and trust in AI systems. This emphasizes that the challenge isn’t just adopting AI.  It’s ensuring robust data integrity to avoid building on a shaky foundation​. Example 1: Microsoft’s Tay Chatbot Microsoft’s AI chatbot “Tay”  was designed to learn from conversations with humans on Twitter. Within 24 hours, Tay had absorbed the worst of the internet, spouting offensive, racist content. Microsoft had to shut it down. At a small scale, this looks like a glitch. But at AI’s scale, bad inputs don’t just fail quietly.  They become amplified with significant business and reputational risks.. AI doesn’t understand why  it’s wrong; it just executes faster and louder. Example 2: Predictive Policing Systems AI-driven policing tools like COMPAS  promise to predict crime patterns and optimize law enforcement resources. But many of these tools have been shown to replicate and amplify racial biases embedded in the training data. Biased inputs lead to skewed outcomes, which get implemented as policy, reinforcing systemic inequality. 💡 The Insight : AI is only as good as the data we feed it. Scaling bad inputs doesn’t just waste time.  It creates consequences that ripple through society. The challenge for businesses and institutions isn’t just adopting AI; it’s stress-testing the inputs  to make sure we’re not building on sand. Infinite Output, Limited Judgment AI can scale work infinitely, but it doesn’t judge what’s useful, meaningful, or even correct. It just does stuff . Take code generation , for example: Tools like GitHub Copilot generate code at astonishing speed, but they also produce bugs, redundancies, or code snippets riddled with security vulnerabilities. In one study, 40% of Copilot’s suggestions contained security flaws . The bot doesn’t know what makes “good” code—it’s focused on volume, not quality. Now, scale this problem across industries, like these two: AI-generated marketing campaigns flood the internet with ads that no one sees or cares about. AI-written legal contracts include nonsensical or risky clauses that go unnoticed until they cause a lawsuit. 💡 The Insight : AI isn’t self-correcting. The faster it scales, the more oversight it requires. Botpower excels at execution, but it’s still humans who define success and/or clean up the mess. The Complexity Ceiling: When Problems Are Too Messy for AI AI thrives in structured environments where patterns are clear and outcomes are measurable. But in messy, ambiguous, multi-dimensional problems, it can really struggle. Consider climate modeling : Predicting global climate outcomes involves countless interdependent variables: ocean currents, atmospheric conditions, human behavior, feedback loops, and unknown tipping points. AI can simulate patterns, but the sheer complexity of the system makes prediction a moving target. Or take economic forecasting : AI models analyze vast datasets to predict financial trends, but they often miss black swan events —low-probability disruptions (like the 2008 crash or COVID-19) that break all historical patterns. 💡 The Insight : AI is a powerful tool for structured problems, but the world doesn’t always play by structured rules. Complexity forces us to ask: Where does human reasoning, creativity, and gut instinct still matter most? The Infrastructure Tax: Botpower Isn’t Free We love to talk about AI as a force multiplier. But we rarely talk about what it costs to keep the system running. Training GPT-3 required over 1,287 MWh  of energy—the same as 120 U.S. homes  use in a year. Scaling AI systems requires immense compute power, data storage, and electricity. This infrastructure doesn’t scale cheaply or sustainably. Companies adopting Botpower need to ask hard questions: Can this scale affordably? What’s the environmental cost of our AI infrastructure? What happens when compute resources become bottlenecks themselves? 💡 The Insight : AI feels weightless, but its footprint is real. Scaling Botpower responsibly will require smarter infrastructure, not just smarter models. Stress-Testing Humans: The Real Bottleneck Here’s the paradox: Botpower can scale execution to infinity, but human judgment is finite. The faster AI moves, the harder it becomes for humans to keep up.  A recent article in Minds and Machines  highlights how AI accelerates productivity but simultaneously stretches human oversight to its breaking point, raising questions about maintaining autonomy and control over these systems.  Content Moderation : AI generates fake news, deepfakes, and misinformation faster than human teams can detect it. A 2023 study by Europol   highlighted that AI can produce convincing disinformation up to 10x faster  than traditional methods, overwhelming detection systems and platforms tasked with managing content integrity. Cybersecurity : AI accelerates both attack and defense. A security flaw that would have been buried in the noise five years ago can now be exploited at machine speed.  This arms race between AI-accelerated attacks and defenses has reshaped cybersecurity timelines, shrinking the “time-to-exploit”  window from weeks or months to mere minutes  or hours . 💡 The Insight : The more Botpower we deploy, the more we stretch our ability to oversee, correct, and adapt. At some point, humans – not AI – become the bottleneck. Conclusion: Botpower Has Limits – And That’s a Good Thing If you stress-test AI hard enough, you’ll see where it breaks – in its reliance on flawed inputs, its lack of judgment, its struggle with complexity, and its very real infrastructure costs. But these limits aren’t just problems; they’re reminders of where people - and businesses - still matter. AI doesn’t solve for meaning. It doesn’t understand context, ethics, or human nuance. It doesn’t know where to focus or when to stop. That’s our job. And maybe that’s the ultimate stress test: Not just how far AI can go, but how well we can guide it. This exploration is far from complete – let’s keep digging. Where do you see the cracks in Botpower? Share your thoughts at hello@sentinelglobal.xyz .

TL;DR: At Sentinel Global, we have five core beliefs shaping our vision of Interoperable Commerce: AI should amplify human decision-making, not replace it. Agent-based systems are the future, revolutionizing supply chains with real-time intelligence. While blockchain gets the hype, we prioritize practical trust through robust data governance. As costs fall, intelligent systems will become accessible to all, unlocking unparalleled efficiency. And at the heart of it all, data integrity is non-negotiable. Do you agree? “To accomplish great things, we must not only act, but also dream; not only plan, but also believe.”  — Anatole France At Sentinel Global, what sets our views apart from those you might find from an investment bank, consulting firm, or think tank? As a global VC, we are on the front lines speaking and working with founders every day, actively identifying and investing in the pioneering technologies that will shape our digital future and envisioning the world they will create. Our position gives us a unique vantage point: we place early bets on transformative innovations, but we also uncover early flaws. We see firsthand how ideas that seem brilliant in theory can falter when faced with the challenges of scaling or integrating into the “real world.” Through our direct work, we’ve developed a set of core beliefs that guide our approach to the key technologies and methodologies in Interoperable Commerce.   Here are five of these beliefs: The Crucial Role of AI enabled Decision-Making in Supply Chains We believe that artificial intelligence is essential for supply chain transformation, not to replace humans, but to enhance their decision-making capabilities with real-time insights Artificial intelligence holds tremendous potential for transforming supply chain operations. Supply chains, by nature, are labor-intensive and slow to adopt new technologies. This is why we believe AI’s role in supply chain management is essential —not for automating processes, but for enhancing decision-making capabilities . AI can process vast amounts of real-time data and provide intelligent insights based on contextual changes. This is particularly important in environments where decisions need to be made quickly, such as in warehouse operations or distribution centers. The ability of AI to analyze data and provide actionable insights in real time makes it a critical tool for optimizing supply chain performance. This is not about replacing human workers right now but rather augmenting and enhancing their capabilities with better intelligence, allowing them to focus their time on making more strategic decisions vs. manual tasks.   Autonomous Agents: The Future of Supply Chains We believe that agent-based systems will play a pivotal role in creating more efficient, resilient, and responsive supply chains by making real-time, intelligent decisions that surpass human limitations. One aspect of Interoperable Commerce we are most excited about are agent-based systems in decision-making. Agents – intelligent systems designed to perform specific tasks - can significantly enhance decision-making in supply chains, particularly through analyzing and interpreting real-time data entering into the companies purview through various streams.  Rather than depending solely on vast amounts of pre-processed, easy-to-query data, agents can now analyze data in real time as it is generated, make immediate decisions, and seamlessly integrate it with stored data. This approach provides a more comprehensive and holistic view for the company. For example, in a typical warehouse environment, orders are processed and distributed by various companies. Historically, a human operator would monitor these orders and make decisions based on somewhat real-time demand and availability. With agent-based systems, the same, or better, decisions can be made more efficiently and in real time, reducing the need for constant human oversight. Agents are able to operate around the clock, processing real-time orders, adjusting to changing conditions in a way that humans simply cannot and making far less mistakes doing so. This shift from human decision-making to agent-based decision-making is inevitable. While humans will still play a role in overseeing these systems for the time being, agents will gradually take over more tasks, leading to reduced labor costs and improved profit margins. The result is a more efficient, resilient, and responsive supply chain, with agents positioned to make better decisions based on real-time data. The ROI and Cost Curve of Intelligent Systems We believe that the cost and complexity of intelligent systems will decline significantly over time, making them accessible and impactful for businesses of all sizes. A key consideration when adopting intelligent systems is the return on investment (ROI). Although the upfront costs of deploying AI and agent-based systems may be significant, we expect these costs to decline sharply as the technology evolves and becomes more widespread . For innovation to be truly impactful, particularly for small and medium-sized businesses, these systems must be affordable and simple to implement. Establishing standards for data sharing, encryption, and formatting will be crucial, as consistent data structures enable seamless integration and easier downstream application. As these technologies become more accessible and cost-effective, the enhanced decision-making, operational efficiency, and responsiveness they provide will far surpass the initial investment. We also anticipate that intelligent systems will eventually reach a point where they can deliver consistent answers without needing to rely on large, resource-intensive models like LLMs (Large Language Models). This shift will further reduce costs and improve the scalability of intelligent decision-making in supply chains. One benefit that supply chains hold over other industry verticals looking to adopt and deploy AI is that often the nature of the information that is needed to supply to an agent or LLM is less sensitive / full of trade secrets vs. other industries. Law firms, financial institutions and healthcare organizations, for example, have proprietary, sensitive, and personally identifiable information that they don’t want ingested and held by outside organizations. This has the potential to create a feedback loop that will allow better models and agents to be trained based on real data created every day and will eventually result in more performant, accurate systems. The Limitations of Blockchain in Complex Supply Chains We believe that blockchain, while useful in certain contexts, is not the solution for managing trust in complex supply chains; instead, robust data governance and encryption provide more practical and effective safeguards. Despite enthusiasm for blockchain, we believe its role in managing complex supply chains is limited. While the academic consensus often suggests that blockchain enhances trust, this view remains debatable—particularly in supply chains , where its effectiveness has fallen short of expectations. At Sentinel Global, we have over a decade of experience investing in blockchain, so we understand its theoretical appeal as a secure and transparent technology. However, blockchain solutions alone do not address the trust challenges inherent in large, multifaceted supply chains. Instead, we find that trust is more effectively established through robust data pipelines, data governance and advanced encryption techniques, which offer tangible safeguards. As a result, our approach prioritizes these data governance methods over blockchain to ensure the integrity and reliability of supply chain data. The Challenge of Data Integrity We believe that maintaining data integrity is foundational to building trustworthy and reliable systems, ensuring that decisions are always based on accurate and secure information. Finally, we recognize that one of the most significant challenges in Interoperable Commerce is maintaining data integrity. Disinformation can spread quickly from one data point to another, and even a single corrupted source can compromise an entire decision-making process. Ensuring the integrity and safety of data at every step of the supply chain is critical to building a trusted and reliable system. By adopting advanced security solutions and having a focus on understanding where your data lives, how it affects or exposes your company, and how it best can be utilized to make better decisions internally and externally will separate the new class of successful enterprises from those of the past. Combining this with strong data governance and policies, we can minimize the risks associated with data sharing and ensure that decisions are based on accurate and reliable information. In Summary: Our beliefs about Interoperable Commerce center around the transformative role that advancements in AI and agent-based systems will play in the decision making process of enterprises. We see the future of supply chains as being driven by intelligent systems that enhance, rather than replace, human decision-making. As the cost curve of these technologies continues to fall, we believe that the widespread adoption of AI and agents will lead to more efficient, responsive, and trustworthy supply chains. Our focus on data as the core driver of these advancements, intelligent decision-making, and the careful adoption of emerging technologies will guide us as we navigate the future. Do these beliefs about Interoperable Commerce resonate with you? What would you add, challenge, or approach differently? We’re eager to hear from startup founders and institutions shaping the future. Share your perspective—let’s exchange ideas and push the boundaries of what’s possible together. Reach out and join the conversation. hello@sentinelglobal.xyz

TL;DR: Sentinel Global's David Renne and Andy Reed reflect on their findings after months of speaking with key leaders across supply chain technology. They assess how the biggest barrier to supply chain transformation is tech adoption. Despite AI-driven solutions, poor implementation and organizational inertia are preventing companies from realizing their full potential. They explore the reasons for this and think of the paths ahead for the industry. “ The difficulty lies not so much in developing new ideas as in escaping from old ones.” - John Maynard Keynes Over the past several months, we have traveled across the U.S., attending a variety of supply chain technology conferences. Throughout our tour, analysts, executives, and technology providers consistently reinforced one theme: adoption remains the greatest barrier to transformation.  According to an analyst giving one of the opening keynotes, 41% of a company’s return on investment (ROI) in supply chain planning is lost due to poor implementation and change governance.  The reality is stark.  Despite the promise of AI-driven solutions, most companies are struggling to integrate them effectively. Data quality remains a persistent challenge , with one executive stating bluntly, “If we could leverage AI to clean, enhance, and standardize data, then we would have something. As it stands right now, models are learning from useless data, and inherently, themselves, become useless to us.” At Sentinel Global, our role as investors compels us to distinguish hype from reality.  To assess the state of supply chain innovation, we engaged in conversations spanning multiple industries, from Fortune 500 companies to emerging technology providers, and across multiple conference settings. These discussions provided a first-hand view of the challenges facing the sector and helped us evaluate the credibility of different players' perspectives. All of these conversations underscored a fundamental truth: organizational inertia is the most formidable obstacle to widespread adoption. At one conference, ParkourSC  captured the challenge of supply chain planning with a simple phrase: Legacy Limitations . While the alliteration, or perhaps the simplicity, drew us in like moths to a flame, the message resonated deeply.  Frankly, existing approaches to supply chain planning and supply chain technology broadly, no longer provide competitive advantage to their customers. Our research at Sentinel, has indicated that  data management, procurement, demand planning, demand shaping, and global trade management   are areas ripest for disruption.  Multinational Fortune 500 companies, in particular, are eager for advancements in demand planning, and we intend to monitor this space closely. Here are our key takeaways from the conference circuit: Adoption is the Biggest Hurdle At Sentinel, we set ourselves apart by engaging directly with adopters—the key drivers of change and the most critical force in shaping our investable universe. On our trips, we listened to their stories, pain points, and outlook on supply chain technology. Our conversations spanned org charts, from analysts to CEOs and co-founders, across industries and roles. A common thread emerged: their eagerness to be heard.  While we pride ourselves on being approachable, we sensed widespread frustration with the industry and a hunger for game-changing solutions. The next wave of supply chain technology must be built with an adopter-centric approach. In our experience, adopters are ready to talk. Multiple speakers and attendees of one conference highlighted that only 40% of required users actually engage with purchased supply chain planning tools. This means that in a planning team of 50, only 20 people are effectively using the technology as intended. A manager of US supply chain at a pharmaceutical shipper remarked to us in an extemporaneous conversation, “I recently departed [a large, multi-national pharmaceutical company] for [a smaller pharmaceutical manufacturer of vaccines and insulin], and my whole day was spent in Oracle’s supply chain planning module at my previous company. Oracle worked great there, because people bought in. Now, at my new company, I have little use for it because the organization broadly doesn’t incentivize adoption.”  Narrowing down on AI and machine learning solutions, supply chain planning providers noted only a ~20% adoption rate of AI/ML forecasting tools.  Tools and technologies that have already revolutionized other areas of our economy are being ignored by 80% of existing users of supply chain planning. Again, we are befuddled by this reality, particularly given one case study shared by Gartner:  one large international consumer goods shipper, with proper planning and organizational buy-in, achieved a 35% reduction in waste and 40x faster proposal generation by implementing AI/ML forecasting tools. Opportunity obviously exists, but we are struck by its dependence on adoption and consensus building within organizations. Dominant Tech is Stale and Vulnerable Innovation in supply chain technology and supply chain planning, as we’ve acknowledged in our recent insights post , has been lacking. Current solutions from pioneers like Kinaxis  and  o9 Solutions – whose trailblazing spirit we acknowledge – were constructed ~30 years ago. Additionally, several of our conversations have further enforced one key element of our existing thesis: Legacy supply chain planning technologies still, after decades-long head starts, haven’t embedded themselves deep enough into their adopter base to create durable barriers to entry – beyond the simple fact that better alternatives do not exist. While recent unfruitful investments in this space might suggest otherwise, we spoke with several impressive teams who have successfully implemented new-age, AI-driven software in place of large incumbents. Indeed, we’re excited to see the traction in this space, and we believe that as adopters continue to find their voices, the best will rise to the top.  GenAI is Met with Skepticism At these conferences, we were confounded by the conversations around AI. In many other industries, AI, ML, neural networks create buzz and excitement for the future.  A simple search of just about any public company’s earnings calls with analysts’ questions reflect this reality. Public and private company CEOs are constantly peppered about how their companies are adapting to GenAI and the rise of human-like agents.  But in our experience at supply chain conferences, these conversations were met with metaphorical eye rolls, with most participants taking a “I’ll believe it when I see it” stance about the potential impact of AI throughout the supply chain landscape. We were also struck by the unsparing and arguably dour tone that our contacts had when the topic arose. As noted earlier, their collective outlook remains shaped primarily by one key factor: data quality. Lastly, one striking takeaway from our discussions was the body language of industry veterans when AI came up . There were folded arms, skeptical expressions, and a general “prove it to me” attitude. This mirrors a classic pattern in disruption theory – that established players often dismiss transformative innovations until it’s too late. Taxi companies dismissed Uber. Retailers underestimated e-commerce. Supply chain incumbents may be falling into the same trap. At Sentinel, we look for inflection points – those moments where new technology crosses the threshold from theoretical to indispensable. While widespread AI adoption in supply chains isn’t here yet, the companies that are solving the adoption challenge will define the next era of supply chain efficiency. Conclusion The key takeaway from our conference tour is that AI, next-gen supply chain planning tools, and supply chain technology innovations are not yet sweeping the industry. However, for companies that  successfully drive adoption, the potential gains are significant. The US freight recession that has persisted since COVID-19 is expected to abate by 2025 or 2026, which could unlock budgets and accelerate investment in next-generation technology.  However, at one conference we attended, there was a clear bifurcation of thoughts on the state of the US freight economy. Technology providers are seeing green shoots emerge, while practitioners’ outlooks were far more muted.  In a sense, the coming year could be a referendum on whether technology can, indeed, provide a backstop to cyclical forces. At Sentinel, we continue to bridge the gap between founders and adopters, investing in companies that not only develop innovative products but also solve the adoption challenge. If you are working on the future of supply chain technology, or have views that differ from our findings, we are ready to talk.  Reach out to us at david@sentinelglobal.xyz  and andrew@sentinelglobal.xyz

TL;DR: Open computing, which emphasizes collaboration and transparency, is vastly more efficient and scalable than closed networks. DeepSeek's recent success, similar to past innovations like Linux and Android, highlights again how open systems drive smarter, more cost-effective solutions and challenge traditional proprietary models. Knowledge is like a garden: if it is not cultivated, it cannot be harvested - African proverb Since our beginnings at Sentinel Global, we have championed the advantages of open computing  – our belief that openly developed technologies will drive efficiencies through better collaboration, transparency, and innovation. Sentinel’s Managing Partner, Jeremy Kranz, was one of the first venture investors  to actively specialize in open source in the 1990s. Now, the recent developments surrounding DeepSeek , a Chinese AI startup, serve as a compelling case study.  Few examples better illustrate how open computing not only disrupts outdated practices in technology and venture funding but also highlights the undeniable advantages of open-source development. DeepSeek: A Success Story in Actually Open AI DeepSeek spun out of the Chinese hedge fund High-Flyer  in 2023 with the goal of building highly efficient AI systems. Their flagship platform, DeepSeek-R1, released in early 2025, has gained attention  for delivering competitive results on a purported budget of $6 million. Instead of following the traditional tech VC playbook – that is, building massive, proprietary systems requiring immense hardware resources and blitzscaling – DeepSeek embraced open AI frameworks and focused on maximizing software-driven efficiency . Their achievements are, on the surface, very impressive. By prioritizing algorithmic optimizations over hardware dependence, DeepSeek-R1 has demonstrated faster training cycles (and a lower carbon footprint ) than many proprietary competitors. For instance, early benchmarks suggest their models achieve up to a 30% improvement in efficiency over industry standards.  This success challenges the assumption that innovation in AI must rely on some sort of trade embargoed, hardware arms race. DeepSeek’s software-focused, open approach demonstrates that it’s possible for an ambitious, nimble team to achieve remarkable results by optimizing algorithms. Of course, questions remain. For instance, while DeepSeek benefited from an unknown quantity of NVIDIA H-100 GPUs , it’s unclear how they secured access to these advanced chips amid export controls. Similarly, while their stated $6 million training cost is eyebrow-raisingly impressive, replicating that level of efficiency could depend on factors not yet fully transparent. Still, their work proves that success in AI increasingly depends on smart strategies and open ecosystems – not just deep pockets. The Principles of Open Computing in Reality At Sentinel, we identified early on  that that the bulk of future enterprise value will be built on top of open computing systems and networks,  DeepSeek’s story embodies several core tenets of our thesis: Global Collaboration Drives Innovation By leveraging open-source frameworks, DeepSeek tapped into the collective expertise of the global developer and research community. This collaborative model allowed them to build on existing innovations, reducing redundancy and accelerating development. For example, DeepSeek’s adoption of Hugging Face ’s transformers libraries saved them years of foundational work. Efficiency Over Exclusivity Rather than pouring resources into proprietary systems, DeepSeek focused on optimizing what was already available. Their approach underscores that the best solutions often come from working smarter and not just spending more. Adaptability Through Openness Open-source frameworks allowed DeepSeek to remain flexible, updating their models as new techniques emerged. This adaptability is critical in a fast-moving field like AI, where static systems quickly become obsolete. Transparency and Communication Open-source systems inherently promote transparency, fostering freer interactions among users and collaborators. While there are still some open questions about DeepSeek’s operations, their willingness to share their model represents a step forward in accountability. Transparency also builds trust, a critical element in a rapidly evolving technological landscape. Not The First Open Tech Rodeo DeepSeek’s success isn’t just a one-off.  It’s part of a pattern showing the superior power of Open Computing – although VCs and startups often forget this important pattern. History is full of examples where open ecosystems have reshaped entire industries: Linux  revolutionized operating systems by proving that open-source software could compete with (and outperform) proprietary giants like Microsoft. The Android operating system  is based on the open-source Linux kernel. Its open-source nature has allowed manufacturers and developers worldwide to build on it, making it the most widely used mobile operating system globally The rise of the internet itself  was built on open protocols like HTTP and TCP/IP, which enabled collaboration and interoperability on a global scale. More recently, Stability AI ’s release of Stable Diffusion , an open-source generative image model, has enabled widespread customization and deployment of advanced AI tools. This has led to innovations in industries like gaming, e-commerce, and design, with developers creating countless applications, from AI-powered art tools to video game asset generators.  DeepSeek is simply the latest chapter in this story. Honestly, it’s sometimes confusing why history keeps repeating its blitzscaling, growth-at-all-costs model when more efficient and cost-effective open computing models repeatedly prove investors wrong.. The Broader Implications of DeepSeek’s Work DeepSeek’s achievements hold lessons for technology builders today. They’ve shown that success doesn’t require proprietary control or massive capital investment. Instead, it requires: A focus on efficiency  over resource accumulation. A nimble and educated team  capable of leveraging open-source tools. A willingness to challenge traditional assumptions  about innovation. At Sentinel Global, we firmly believe the future of technology lies in the principles of open computing. It’s not just about building better tools more efficiently; it’s about creating an environment where innovation is accessible to everyone. Founders, investors, and policymakers should prioritize open ecosystems, contribute to open-source projects, and rethink funding models to support collaborative innovation.

TL;DR: We consider the development of power metrics over time and introduce the concept of “botpower.” We discuss its effect on coders and business-building decisions, how the question of build vs. buy will become a relic or the past (the answer: build), and frame our key views around the integration of AI in future enterprises. There are over 2 million cars standing in front of red lights with their engines going. Then we have over 2 million times approximately 100 horsepower being generated as they are idling there, so that we have something like 200 million horses jumping up and down and going nowhere. Now, we have to count that in our economy when we begin to get down to what is the efficiency of the economy. – R. Buckminster Fuller I have always found it interesting how certain measurements transcend the epochs. The mile derives from a Roman soldier’s thousand paces - mille passum  - as measured by his every other step.  That is to say, the total distance of his left foot hitting the ground 1,000 times.  Note this would be at a quick trot, not a leisurely walk.  A furlong (meaning furrow length) was the distance a team of oxen could plough without resting. This might seem quaint today except when you walk north-south on a city block which is… one furlong long.   And then there’s horsepower.  I’d like to quote this exactly from  Britannica : “Horsepower is the rate at which work is done. In the British Imperial System, one horsepower equals 33,000 foot-pounds of work per minute—that is, the power necessary to lift a total mass of 33,000 pounds one foot in one minute. This value was adopted by the Scottish engineer James Watt in the late 18th century, after experiments with strong dray horses, and is actually about 50 percent more than the rate that an average horse can sustain for a working day.” Let's underscore that.  As we think about the output of our engines and motors, not only are we measuring them to horses, but to ones that are significantly stronger than average.  Similarly, the mile was a measurement of a presumably very fit and healthy soldier trotting at a fast clip.  The poor oxen outlining a city would have to plow without resting once.  It seems like so many of our measurements are based on a desire to exceed the average. Which brings us to manpower.   The term (modernized to “workforce”) reflects the total number of human workers available or required for a task, project, or within an organization.  But there is a technical definition of one manpower approximately equaling 75 watts, which is about one-tenth of a horsepower.  That is to say, one strapping dray horse = ten men, and one average horse = five men. There is obviously a point where equations don’t make sense.  We cannot say that all of Amazon’s work can be done by 300,000 average horses.  Yet as we consider the emerging world driven by AI-juiced productivity, how should we think about the reallocation of human productivity, especially in the context of technology?  Could an equation help a cost-benefit analysis?  And do we have some sort of expectation, like with miles and furlongs and drays, that our metric should not just be the replacement of highly repetitive tasks, but ones that require very specialized strengths – say 1,000 Stanford computer science grads? Ultimately, what we are trying to define - from the trek of horsepower to manpower - is the next stop on the journey: botpower .  So let’s more closely consider those three questions around 1) human reallocation; 2) a technical definition; and 3) our expectations, to see what we can better understand. Reallocation of Human Productivity It is widely understood that AI is already automating many tasks that were once done by humans, and this trend is only going to continue in the years to come. A study by the McKinsey Global Institute found that up to 800 million jobs could be transformed due to automation by 2030, representing 15% of all jobs.  This includes fields like data entry clerks, telemarketers, cashiers, tax preparers and translators.   This is perhaps undercutting it.  No offense to tax preparers, for example (my mother-in-law is one), but one tax preparer doesn’t create 500 new tax accounting software programs as part of her job.   So let’s consider coders instead.   One study, conducted by researchers at the University of California, Berkeley, found that approximately 40% of the code committed to GitHub Copilot, a popular AI-powered coding assistant, was written by AI. This suggests that AI is already playing a significant role in code generation, and, profitable, with Github copilot surpassing $100M in ARR . Another study, conducted by researchers at the University of Oxford, found that AI could potentially generate up to 80% of the code that is currently written by humans. This means that AI could have an even greater impact on code generation in the future. Hopefully the implications of this are clear.  Botpower can enable enhanced developer productivity, reduce development costs and accelerate software development cycles. Here are some examples how: AI-powered tools can automate repetitive tasks, such as code generation, syntax checking, and error detection, freeing up developers to focus on more complex and creative aspects of programming. This could lead to a significant boost in developer productivity. AI could help reduce software development costs by automating tasks and minimizing the need for highly skilled programmers. This could make software development more accessible to businesses of all sizes. AI-powered tools could accelerate software development cycles by automating tasks and providing real-time feedback. This could lead to faster delivery of software products and services. And on the human capital side, the blitzscaling method of talent acquisition goes extinct, and democratization of coding proliferates.  AI could make software development more accessible to individuals with no prior programming experience. This could democratize software development, allowing more people to create their own software applications. Of course, AI-generated code may not always be as high-quality as human-written code. Developers need to carefully review and test AI-generated code to ensure its quality and maintainability. It may introduce new vulnerabilities and security risks, and developers will need to adapt their skills and practices to work effectively with AI tools. The Question For Enterprises All in all, what we are facing is a massive transformation of our business building landscape.  As a result, for many firms of the future, the question of “build versus buy” may easily become a relic of the past, as the default answer will always be “build.”   Why?  Rather than utilizing manpower, they can increasingly utilize botpower.  The repetitive stress testing, the lack of downtime, the continuous learning, the (near) total system control and the ability to utilize not just internal data but public code will enable companies to be able to build on top of existing systems and/or more efficiently produce new ones.  Without the need for employees to actually know how to code, enterprises can utilize their existing workforce to create and execute complex programs without the theoretical need for third party software providers.  SaaS becomes an internal function of an institution. How should an enterprise model these considerations into their financial and business plans?  Personally, I am eager to determine a precise equation to quantify botpower, inspired by Watt’s determination of horsepower.  Yet maybe the cost-benefit analysis will just be pretty crude and simple: an analysis of the cost of building software internally and the number of people not hired to do it.  I remember a former AI-related portfolio company of mine not too long ago that priced their product based on how many data scientists a business would not need to employ if they used their solutions.   Expectation of Greatness At the start of this blog, we noted all the many ways that our units of measurements are based on an inflated metric of strength.  As we proceed with our exploration of botpower here at Sentinel Global, we are mindful not to make too many assumptions of either the abilities or deficiencies of human productivity and the achievements or alarm generated by AI.  Humility and realism must be key.  We instead are focused on these primary points as we consider our investment views in AI Skills Development: While AI may displace certain jobs, it also has the potential to create new roles and redefine existing ones. The focus should not solely be on job displacement but on identifying opportunities for innovation and entrepreneurship.  Collaboration:  Rather than complete displacement, the integration of AI in coding is likely to lead to a collaborative relationship between AI systems and human coders. Successful collaboration will require coders to understand AI tools, leverage their capabilities, and contribute uniquely human skills, such as creative problem-solving, critical thinking, and domain expertise.  Quality Assurance: AI algorithms may inadvertently introduce biases, errors, or security vulnerabilities in the code. Ensuring ethical coding practices and maintaining code quality will be critical. Implementing rigorous testing procedures, conducting regular code reviews, and incorporating ethical guidelines into AI development processes are vital steps to mitigate potential issues.  My colleagues and I at Sentinel Global  are only at the beginning of our botpower exploration. Together, we look forward to considering more together with you.  Share your thoughts with us here. hello@sentinelglobal.xyz

TL;DR: The world of computing and business is not currently bound by the binary distinction of open and closed systems. However, in the future, open computing technologies will hold the key for enterprise success with their inherent advantages of cost-effectiveness, flexibility, and innovation. We believe these technologies will be embraced by institutions with far-reaching impacts on a global scale. Open minds unite, Code flows in open spaces, Future's freedom bright. – ChatGPT (when prompted to write a haiku about the future of open computing) In our blogs here at Sentinel Global, we frequently refer to "open computing”. This is our informal tagline for the development and sharing of technologies among multiple parties or networks. These technologies enable three important things concurrently: Alleviate the burden of customized, centralized R&D (i.e. decentralized innovation) Enhance interoperability across enterprises and tech stacks Enable easy access, verification, development, and security across multiple participants We believe that the bulk of future enterprise value will be built on top of such open computing systems and networks. Why?   Here are our seven reasons. In an era of increasing business complexity and rising enterprise costs, companies are compelled to pursue greater efficiencies  not only in software but also in their infrastructure, networks, and systems. As firms expand into new markets, these optimizations become even more imperative. The continued need for increased scale and lower costs remains paramount. Talent is flocking to open systems . An early and leading indicator of value creation in emerging platforms is developer engagement.  In 2012, there were 2.8 million million developers on GitHub, the open-source platform.  In 2022, there were 94 million. Basic supply and demand constraints . Startups lacking the resources or inclination to establish extensive internal cybersecurity infrastructure will seek open computing solutions . These solutions provide an alternative for those unable to construct and uphold their own private networks.  The proliferation of   different programming languages matters as well. Today, 20-30 languages are “commonly used” while The Online Historical Encyclopedia of Programming Languages ( OHLO ) lists over 8,945 distinct programming languages.  Leveraging established and open infrastructure minimizes the need for developers to master all of the languages needed to complete their specialized tasks. Obviously this gets accelerated with AI , which can be a co-pilot to developing code itself.  By building on public data, that code can offer more code. It is no longer an arms race of amassing headcount to build complex proprietary software. AI can effectively augment privately developed code with exponential public solutions.   The increasing value of open computing is supported by historical evidence.  Open collaboration has consistently led to successful innovation, resulting in improved efficiency and security, from nautical charts  to open source software . There is no reason to believe this iteration of history will be any different. Lastly, and perhaps most obvious is that accessible information begets more information . Retweets, the comments section, memes, and other forms of shared media have highlighted the power of permissionless innovation  on public information.   Looking Ahead The world of computing and business is not currently bound by the dichotomy of open and closed systems. For the next decade, it will thrive on a workable coexistence of both, where their strengths complement each other to create a robust and versatile infrastructure.  That said, while proprietary software solutions offer advantages in bespoke customization, security, and legacy compatibility, open computing technologies hold the key to the future of computing and business . Its inherent strengths – cost-effectiveness, flexibility, and innovation – should ultimately propel it ahead in institutional technology architecture. As legacy infrastructure gradually transitions to digital systems, the compatibility concerns surrounding open computing should diminish . Additionally, the increasing adoption of open standards and open-source components by major technology vendors is further paving the way for seamless integration with legacy systems. Sentinel Global stands at the forefront of this transition . If you are a startup building upon open computing technologies or an institution seeking to learn more, we would love to hear from you.  Reach out to us at hello@sentinelglobal.xyz

TL;DR: We introduce our fund, Sentinel Global, and share why we are so focused on technologies that are open to all. These are the innovations that will be creating businesses that are better, cheaper, stronger and safer. An invention has to make sense in the world it finishes in, not in the world it started.” Timothy O’Reilly, who coined the terms “open source” and “Web 2.0” Introducing Sentinel Global Sentinel Global is a venture capital fund that centers around transformative technologies and business models.  We believe these technologies will reshape industries by boosting efficiency, safety, and security. Decentralized networks, open AI applications, streaming data infrastructure, and supply chain optimization are just a few of the areas we are keenly focused on.  They will help form the foundation for enhancing technology interoperability, optimizing processes, and maximizing efficiency. The ultimate result? Businesses that are not only stronger and safer, but also more efficient and cost-effective. Rather than closed-off proprietary systems, centralized tech monopolies, and incompatible devices and networks, there is a growing movement, especially among developers and entrepreneurs, to pursue a greater connectedness in the innovations they are building.  We are on the edges of a new industrial era, driven by technologies that are open to all, and Sentinel is super excited to be at the front gate. Our Era of Openness Why do we believe that technologies built with an open ethos will ultimately prevail? Well, let’s take artificial intelligence. AI was actually established as a field in 1957, with a number of hype and trough cycles since then.  So what made the explosion of interest in AI in 2023 seem so different?  For the first time in history, multiple high performance models were distributed and made open to the general public , allowing them to build their own AI applications that could rival the performance of closed source models . Rather than being developed only by the government or secretive tech and security firms, AI became available to the masses. Open source models like Mixtral and Llama 2 gave savvy developers the ability to fine-tune models based on their own proprietary data sets and expanded the original bounds of these foundational models. Just take a peek at hundreds of models created on the Hugging Face Open LLM Leaderboard  and you can see the innovation, creativity, and excitement generated by making these models open to everyone. Here’s another example: digital assets.  Since Bitcoin’s inception in 2008, we have witnessed the development of a globally decentralized, open, and secure internet where value is exchanged peer-to-peer.  Everyone can see each transaction and all can vouch for its validity.  Blockchain technology has further  spurred a head-spinning (and sometimes head-scratching) number of additional innovations around decentralized finance (DeFI), digital collectibles (NFTs), and tokenization of real world assets, among many others.  Value can be transferred online in the same way that information is exchanged online - without intermediaries and centralized parties. The sum of that value, just in terms of the market cap of all cryptocurrencies, is $2.6 trillion as of March 18, 2024. As we delve into subsequent posts and engage in debates, we can question the validity of various use cases in AI, blockchain, and other emerging technologies. However, at our core, we firmly believe that the most value will be unlocked from businesses, products, and applications that embrace an open ethos. Why This Matters As our world becomes increasingly digital, the need for effortless interoperability and global scalability has become more critical than ever before.  The complexities of diverse technologies, global connectivity, user expectations, scalability requirements, data accessibility, and the pursuit of a competitive edge in the digital landscape will all impact innovation and cross-industry collaboration. That said, It can be incredibly challenging for legacy institutions in particular to adopt new technologies.  It will require the buildout of institutional grade investment products and service providers to craft solutions that break down technological, infrastructural, and security obstacles for implementation. It will require builders and adopters to come together, and recognize efficiencies, partnerships, and a common purpose for the future. This tactical adoption is why Sentinel Global exists.  It is our mission not just to identify and invest in the next generation of transformational technologies, but manifest them into institutional adoption.   The Sentinel Formula How does Sentinel achieve this mission? Enterprise Adoption:   We connect innovative companies with traditional institutions, facilitating partnerships that bridge the gap between cutting-edge technologies and established frameworks. We help large enterprises adopt open computing infrastructure, paving the way for scalable and efficient technology solutions. We create thought leadership to drive the adoption of open computing, shaping the industry’s conversation. Global Connectivity:  We scale startups, securing regulatory support globally, with a particular focus on markets beyond the U.S.  We cultivate a valuable network of institutions and strategic partners, especially in emerging markets where opportunities abound. We drive programmatic value-add by curating communities that foster collaboration, innovation, and mutual support. Investment Integrity:  We are experienced practitioners who understand the fundamentals of open-source tech. We focus on investment discipline and rigor, corporate governance and controls, and a company’s management team formation.  We are the sober people at the hype party, never launching into an investment without thoughtful analysis, healthy skepticism, and a belief that expertise matters. Achieving our mission might sound extremely challenging but we couldn’t be more excited about it.  As the opening quote noted, “an invention has to make sense in the world it finishes in, not in the world it started.”  That’s what we are dedicated to building at Sentinel Global. We intend to foster a seamless investment, thought leadership, and convening platform to connect the world’s builders with the world’s adopters so that our future world will make sense through the technologies our founders are developing today. We are just getting started.  It’s opening day, and time to play ball.  Join the fun and send us a message at hello@sentineglobal.xyz .

TL;DR: In this Sentinel "Deep Dive," David Renne and Andy Reed share how - after extensive research and conversations with industry leaders - they’ve come to see how supply chain data management is a never-ending game of whack-a-mole: fix one issue, and another pops up. AI offers a path forward, but without standardized, interoperable data, even the best systems struggle. The real winners will be the adopters and innovators who break the cycle and make seamless data integration a reality. "Every success story is a tale of constant adaptation, revision, and change." - Sir Richard Branson Introduction In our recent trips to several supply chain technology conferences ( see our takeaways here ), we reconnected with various industry leaders to hear their findings and further understand their pain points. With decades of experience in the supply chain planning space, their insights as both users of legacy systems and early adopters of new technology have been invaluable to our research. As we shared in prior blogs ( here  and here ), one of the biggest challenges in supply chain technology investment is adoption. And underpinning that macro challenge is a litany of micro challenges that have largely held the industry at bay over the past decade. One of those micro challenges, and a key area of focus for us, is the widespread issue of master data management (MDM) across supply chains . MDM refers to the standardization of data - both internal and external - ensuring interoperability across supply chains and their systems - regardless of suppliers, distributors, carriers, or other stakeholders. Without high-quality MDM, operations can grind to a halt, visibility becomes obscured, and supply chains lose their agility. While it may seem intuitive, data-driven decision-making only works when the data is accurate, up-to-date, and accessible .  Relying on incomplete, outdated, or siloed data is a recipe for inefficiency, resulting in imprecise and under-optimized outcomes.  In this Sentinel Deep Dive, we’ll explore some of the common pitfalls in tackling the industry-wide MDM problem and highlight potential solutions for the future. A big thank you to the adopters we met at each conference for sharing their expertise and experience with us. If you, like them, have insights, challenges, or visions for the future of supply chain technology, please reach out!   The Scope of the Master Data Problem The lack of quality data management can cost businesses worldwide up to 25% of their revenue  through lost time and lost sales. This is no different in supply chains – and it is often exacerbated when data inconsistencies ripple from suppliers to customers and back again.  The compounding nature of different data across supply chains creates a whack-a-mole problem . For instance, if you are a large multinational shipper with hundreds of suppliers and want to impose MDM standards across all of them, you must drive change across a vast network of companies. This requires both upstream and downstream counterparties to comply with the standards that you, as the shipper, prefer – standards that may not necessarily align with what is best for each individual business. Now, suspend disbelief for a moment and assume this standardization is achievable in today’s world. What happens when one of these suppliers adjusts its standards to accommodate another large multinational shipper? This is where the whack-a-mole syndrome takes hold, and the ability to manage every possible node of your supply chain becomes untenable. This is a point that harkens back to our earlier piece  – a key barrier to adoption is clean, accessible, and usable data. We believe the lack of MDM standards across the industry have both created opportunity and simultaneously sunk some promising startups and talented founders in recent years. It is worth recognizing previous attempts to standardize data. The first that comes to mind is  EDI standards . EDI, or electronic data interchange, was developed in the 1960s but not broadly adopted until the 1990s. Despite its long history, it remains the gold standard for most trading partners around the world. However, our conversations indicate that EDI implementation and use remain cost prohibitive. Other solutions, like  ISO 8000 , have attempted to force the issue (and ISO 8000, specifically, has necessitated it). However, we still see major gaps in communicable data between supply chain participants. Again, it bears repeating that only incremental strides have been made over the past 60 years.  However, a modern, adaptable, and, perhaps most importantly, affordable solution remains elusive for the industry as a whole. Data is an acknowledged issue here, but the unspoken, more specific concerns are archaic systems and data pipelines. It costs too much to rebuild system architecture, leading many organizations to opt for bolt-on solutions rather than addressing core infrastructure. Key Barriers in Data Standardization and Integration Imagine, if you will: the year is 1950. Goods shipped internationally are loaded into sacks, barrels, and crates – each differing in size and weight, making loading and unloading 1) difficult and 2) time consuming. Enter Malcolm McLean – the father of modern containerization. In 1956, McLean introduced a standardized size and shape of international shipping containers. Effectively, he personally ushered in a new standard for global trade and reduced the time to load and unload international shipments onto and off of ships from a matter of weeks to a matter of hours. Today, nearly 80% of goods worldwide are moved over earth’s oceans in shipping containers born out of McLean’s vision. You may ask - what does this have to do with MDM? The story illustrates that despite the seemingly infinite nature of global supply chains, consensus building and standardization is possible - and McLean accomplished it prior to the advent of the internet and modern communication technology. Fast forward to today, and despite those revolutions in communication standards, international supply chains still operate in siloes, with global standards for how data and people interact seemingly far away.  Why haven’t we achieved a similar breakthrough in supply chain data management? The advent of technology like EDI connections and APIs, though expensive for smaller supply chain participants, are steps toward global unification. However, larger resistance to change and unification remain, as few suppliers are incentivized to adhere to one customer’s requirements or suggestions.  This leaves us with fragmented, disparate systems that are unable to properly communicate and unlock value.  McLean’s revolution worked because there was only one viable standard. Today, we have 30 competing systems, each offering distinct benefits.  How do we choose a solution that provides the necessary features while ensuring seamless, secure, and private data sharing? This question remains unresolved, but we believe the winners in this space will be those who successfully address this challenge – and we may already be seeing some early answers.. Emerging Solutions and Their Challenges As we’ve seen across different industries ( and have written about previously ), open source technologies have revolutionized the way employees, companies, and consumers work, live, and collaborate. As we see it, ubiquitous open-source supply chain planning software could provide gateways to better collaboration and, over time, a potential solution to our MDM conundrum. Additionally, startups in the space can begin to explore freemium models to achieve peak virality, incentivizing participants to coalesce around standardized data to maximize ROI.  Another thought: we believe one effective solution to wrangling the master data problem could be grassroots efforts at universities and large organizations to better inform supply chain talent. This way, use cases are better understood and innovation can more scalably be achieved. After all, supply chain technology is unique, and its best programmers, data engineers, and developers have a very thorough understanding of the industry, its participants, and their needs.  Other industries have democratized access and improved data quality by standardizing data pipeline rails, getting insights closer to the point of creation, using common data stores that play nicely with other internal systems, and allowing for the quick ingestion and transformation of that data to be usable by all factions of the business. When we think ten years ahead, we believe history will favor the founders and adopters who lead the way toward a more democratized supply chain, potentially following the pathway paved by many software and AI startups today. But even within an organization, achieving consistent data flow across teams is a significant challenge, particularly as organizations scale. Teams often rely on different systems (e.g. ERP systems like SAP, Oracle, Acumatica, Plex Systems, and Blue Yonder, to name a few), typically chosen based on leaders’ prior experience and preferences. This fragmentation creates additional barriers to effective MDM within the organization itself. Efforts to integrate disparate systems often result in the same "whack-a-mole" syndrome observed at the broader supply chain level. Even when an organization commits to standardization, the process can take years to materialize due to resistance from teams unwilling to abandon their established workflows and alter how they operate.  Additionally, many large organizations opt for custom-built systems over third-party solutions like Kinaxis. For instance, in our experience with enterprises such as Flipkart, most teams prioritize building proprietary tools tailored to their unique requirements, citing the need for customization and data privacy. This approach frequently exacerbates the problem of data inconsistency, as suppliers are left scrambling to accommodate the distinct data requirements of multiple clients , further destabilizing standardization efforts. Another significant hurdle lies in the skill levels of staff at all levels of supply chain operations. Many advanced tools and platforms demand a steep learning curve, yet employees often lack the necessary training or expertise to navigate these systems effectively. To address this gap, organizations frequently resort to simpler data standards that can be managed by less-experienced staff, inadvertently deprioritizing long-term standardization needs.  Small and medium-sized supply chain businesses, in particular, often operate in a reactive, "firefighting" mode, focusing on short-term survival rather than investing in long-term improvements to data quality and standards. This tendency further undermines efforts to create cohesive, standardized supply chain ecosystems. While some employees have memorized the standard systems they use daily, introducing new tools may seem like an effective strategy.  But as noted, change is often slow and stubborn, and the implementation of new solutions must not feel overbearing on the user.  Path Forward: Strategic Recommendations Shaping the future of interoperable commerce requires leadership.  To that end, it is paramount that leaders in the industry – including ourselves and other VCs –, find and fund companies, founders, and ideas that will shape the future of interoperable supply chains. Companies with sufficient scale in the space should also exert their influence to help the industry coalesce around data standards and systems, leading to a future where the MDM conundrum becomes a relic of a bygone era.  One potential path toward standardization is promotion of open-source or freemium models in the supply chain planning and supply chain technology space, as previously mentioned. Owners and managers can drive adoption by natively embedding these technologies into operations, and building through communally determined standards for data presentation and storage.. Another potential solution is the rise of vertically integrated supply chain AI companies that can clean and transform disparate data into standardized formats. However, much like the containerization revolution ushered in by McLean, this innovation will only be possible once a universal standard is established across the industry. Without a common framework, even the most advanced AI systems risk being undermined by the same fragmentation that plagues current systems.  Different supply chain processes will naturally require tailored standards. For example, the data requirements for procurement differ significantly from those for freight tracking. To address this complexity, we foresee these vertical AI companies developing specialized models for each standard, enabling a more targeted and efficient approach to data standardization. The potential impact of such a solution is immense. By integrating advanced AI capabilities with clearly defined standards, the industry could achieve not only greater interoperability but also enhanced agility. Conclusion The challenges posed by MDM underscore the high stakes nature of supply chain operations. Without effective solutions, minor inefficiencies cascade into significant disruptions, and bottlenecks beget other bottlenecks, stalling entire supply chains and eroding business performance. To address these challenges, collaboration, innovation, and a willingness to embrace new paradigms are paramount. By prioritizing standardization, fostering creative solutions, and leveraging advanced technologies like AI, the industry can overcome its current limitations.  Now is the time for stakeholders across the supply chain to work together and pave the way for a more resilient, efficient, and interoperable future. Now is the time for startups to find their footing and develop bold, innovative solutions that can change the face of the industry at large. Another big thank you to the adopters who were kind enough to provide the inspiration for this post. Please reach out to us if you’d like to speak more about supply chains, interoperability, and data pipelines or if you want to share any interesting ideas you may have yourself. david@sentinelglobal.xyz and andrew@sentinelglobal.xyz

How Programmable Digital Dollars are Transforming Finance Hi Readers! At Sentinel, we're kicking off a series of deep dives -- larger, longer thematic explorations that tap into the full depth of our team's experience and insights. We’re drawing on our network with institutions and our relationships with founders around the world to explore the deeper forces shaping industries and innovation. In this case, Sylvester Wee and Josh Shaked discuss stablecoins. TL;DR:  Programmable Digital Dollars (PDDs), or stablecoins, are revolutionizing finance by enabling instant, low-cost, global transactions. They address inefficiencies in traditional systems, like slow cross-border payments and high fees, and unlock transformative applications, from micropayments to decentralized finance. Challenges such as regulatory uncertainty and scalability persist, but innovators tackling these hurdles are paving the way for mass adoption. This begins to realize the "Finternet," a seamless, borderless financial internet, as the future of commerce. Introduction In just three years, programmable digital dollars (PDDs) have surged from nearly zero to over $180 billion in circulation, with transaction volumes that surpass even Visa. However, we believe mainstream finance continues to overlook their explosive potential as digital infrastructure. At Sentinel Global, we see what others gloss over: PDDs aren’t just digital assets.  They have the power to reshape the financial landscape by making money as accessible and programmable as data on the internet. At Sentinel Global, our foundation in financial institutions and venture capital gives us unmatched strategic insight. With over 50 years in financial services, we engage with more than 20 bank innovation teams every quarter, meet over 200 fintech innovators each year, and have backed fintech pioneers that went public and transformed the industry. Our networks and research decode how programmable dollars are disrupting finance—and why it matters In this article, we’ll take you through the market landscape, break down the real-world applications of stablecoins, and share why we’re bullish about the future of what we like to call the Finternet. A Financial System in Flux: Why Programmable Dollars Matter We believe the  global financial system as we know it is slow, expensive, and geographically fragmented. Credit card networks cost U.S. businesses $130 billion annually in fees . ACH transfers take 1–3 days to process and are regionally limited. Meanwhile, national systems like Brazil’s PIX and India’s UPI operate only within borders, limiting global utility. Consumers and businesses are looking for a financial infrastructure that can keep pace with a globally connected world. This demand is fueling the growth of PDDs. Unlike traditional currencies, PDDs live on open, programmable ledgers, allowing them to move with the same speed and ease as data. These PDDs, which can also be referred to as stablecoins, offer a solution to the inefficiencies plaguing finance. In a conversation we had with Itay Tuchman, former head Global Head of FX at Citi and a digital finance pioneer, he describes the shift as inevitable: “Today’s users want a seamless, interest-bearing asset that can be stored, sent, and received without the high fees or delays of legacy systems.” At Sentinel, we see PDDs not just as digital cash but as a new category of financial infrastructure. By enabling instant, low-cost transactions globally, stablecoins are fulfilling a need that traditional systems can’t. The potential for a “ Finternet ”—a financial internet where money moves as freely and securely as email—is no longer a vision; it’s a reality in the making. As Agustin Carstens of the Bank for International Settlements (BIS) puts it, “The Finternet aims to bring the same progress to finance that we’ve seen with data: any asset, any amount, any time, anywhere, at near-instant speed. ” Growing Traction: What the Establishment Tells Us About Digital Dollars Mainstream finance is starting to recognize the impact of PDDs beyond simple theory.  With $180 billion in stablecoins  circulating globally and over $2.6 trillion in transactions  in just the first half of 2024, PDDs have outpaced traditional payment methods. The widespread adoption of stablecoins is clear evidence that programmable dollars are creating a more agile, efficient financial infrastructure. For enterprises, PDDs offer more than just a new way to transact—they provide tactical advantages that traditional systems can’t match. Stablecoins are already enabling faster, more affordable cross-border payments, which is essential for global businesses and migrant remittances alike. According to the World Bank , typical cross-border fees average around 6%, with delays stretching days. In contrast, stablecoins reduce costs and process times to near-instant. Our research indicates that PDDs are also transformative for day-to-day enterprise needs. They provide transparency, reduce fraud, and enhance supply chain management. The programmable nature of stablecoins means that they can power advanced applications like atomic escrow, where funds are securely held until transaction terms are met. We’ve seen this impact firsthand with partners exploring PDDs to manage capital flows, optimize liquidity, and improve cash management across borders. Real-World Applications: How Digital Dollars are Being Used Today The growing adoption of PDDs highlights their versatility. Here are some of the practical applications we’re seeing take root in the market: Global USD Yield Accounts : In regions with volatile currencies, stablecoins allow users to hold interest-bearing accounts in a stable currency. This provides stability and growth potential, offering a lifeline to economies facing inflation and currency devaluation. Store of Value : For individuals in countries with unstable currencies, such as Argentina or Turkey, stablecoins offer a way to hold value in dollars, providing a reliable asset that isn’t subject to the same volatility as local currencies. Aid Distribution : Stablecoins simplify and speed up aid distribution, reducing corruption and administrative costs by enabling direct, transparent payments. This has been particularly impactful in crisis situations where traditional banking channels are inefficient. Trading and Hedging : For financial professionals, stablecoins facilitate efficient trading and hedging on decentralized platforms, providing a stable unit of account and liquidity in an otherwise volatile market. Case Studies Highlighting Early Institutional Adopters of Digital Dollar Solutions As digital dollar frameworks evolve, prominent financial and payment institutions have begun exploring and adopting PDD solutions. Key cases include: PayPal : In 2023, PayPal launched its own stablecoin, PYUSD , enabling users to make payments using a digital dollar tied to the U.S. dollar. This move underscores PayPal’s commitment to fostering secure and seamless digital transactions within its ecosystem, signaling a major endorsement of stablecoins within mainstream payment platforms. Visa : Visa introduced the VTAP  (Visa Tokenized Access Platform) , a platform that enables banks to issue fiat-backed tokens on blockchain networks. This system is designed to simplify digital dollar transactions across financial institutions and networks. Banks like BBVA  have already piloted this platform , positioning Visa as a pioneer in bridging fiat currency with blockchain infrastructure. Stripe : Stripe integrated USD Coin  (USDC)  stablecoin support on its platform, allowing businesses to access and transact with stablecoins worldwide. This strategic expansion includes a partnership with Coinbase  to promote broader stablecoin adoption globally. Stripe’s recent $1.1 billion acquisition of Bridge  further enhances its blockchain capabilities, emphasizing its commitment to facilitating digital dollar use cases in global e-commerce. SWIFT : The global financial messaging network SWIFT  has announced plans to initiate digital asset transaction trials  by 2025 , a significant step toward enabling seamless digital currency transactions between traditional financial institutions. This move signals an increasing interest among established financial players in supporting digital dollar and other digital asset transactions. Additional institutions, such as JPMorgan Chase, UBS, Ant Financial, the Monetary Authority of Singapore (MAS), Siam Commercial Bank, and State Street, are exploring or piloting similar initiatives, indicating widespread institutional interest in integrating digital currencies into the global financial infrastructure.  Challenges in the First Iteration of Stablecoins: Strategic Hurdles and Problem Solving As stablecoins push the boundaries of digital finance, early-stage challenges expose areas for strategic investment and innovation. Addressing these issues is key to transforming stablecoins from niche digital assets into the backbone of a digital economy. Fragmented Ecosystem: Opportunity for Interoperability Leaders The stablecoin landscape is an open field, with multiple players—USDC, USDT, PYUSD—each charting different paths for regulatory compliance and technical frameworks. This fragmented approach creates interoperability challenges that prevent a seamless exchange between these assets and the broader digital finance ecosystem.  SOLVE THE PROBLEM : Strategic investment in interoperability solutions and standardized frameworks represents a high-growth opportunity. Companies that can bridge this divide will secure a vital foothold in the stablecoin market, establishing themselves as the linchpins of a unified digital currency network. Limited Integration with Traditional Payment Systems: Bridging the Divide Stablecoins currently operate on the outskirts of mainstream financial systems, constraining their potential for mass adoption. Lacking direct integration with legacy banking systems, stablecoins face usability challenges for consumers and businesses alike.  SOLVE THE PROBLEM : Advanced API layers, partnerships with established financial players, and novel regulatory-compliant infrastructure are required. The intersection of digital currency with fiat systems is a space ripe for innovators who can build secure, scalable on- and off-ramps. The potential ROI is considerable for companies able to create seamless, compliant connectivity between fiat and stablecoin ecosystems. Watch for Josh Shaked’s upcoming report on Stablecoin Settlement, which will dive deeper into these integration opportunities. Regulatory Scrutiny and Uncertainty: Clear Pathways Needed for Scalable Growth Stablecoins sit at the intersection of innovation and regulatory oversight, with shifting guidelines and regulatory bodies—like the SEC and ECB—pushing for clarity. The regulatory uncertainty has created hurdles for operators and suppressed broader adoption,  SOLVE THE PROBLEM :  This is a window for well-positioned companies to shape and set the standard for compliance. Startups and established players that proactively engage with regulators can help define these frameworks, opening the door to exponential growth and trust-building. Clear and consistent regulatory pathways will catalyze market expansion. Market Volatility and Trust: Rebuilding Confidence Amidst Skepticism Stablecoins promise stability, but recent high-profile failures, such as FTX and TerraUSD/Luna, have rattled user confidence and illustrated vulnerabilities in the current ecosystem.  SOLVE THE PROBLEM : Building robust systems that withstand market volatility and establishing transparent reserve backing are essential steps in restoring trust. The companies that can prioritize security, transparency, and robust governance will stand out and capture a trust premium, making them attractive candidates for investment. As trust in stablecoins stabilizes, these early leaders will reap the rewards of a loyal user base and increased market share. Adoption and Infrastructural Scalability: Scaling for Mass Market Utility Although stablecoins have gained momentum, scaling up for widespread adoption will demand investment in both user education and underlying blockchain infrastructure. SOLVE THE PROBLEM : Drive awareness and education in tandem with technological advancements. Scalability on blockchain networks, particularly regarding transaction speed and throughput, is critical to support high-volume use cases. Backing projects that focus on infrastructural improvements will pay dividends as stablecoins move from niche applications to mainstream, high-frequency financial tools. Unlocking the Future: New Applications for Programmable Digital Dollars By overcoming key hurdles in interoperability, regulatory clarity, and scalability, the stablecoin ecosystem opens the door to a new wave of financial applications that have the potential to reshape global commerce, digital interaction, and personal finance.   Micropayments and In-App Transactions Imagine a world where every interaction holds tangible value. Stablecoins make it possible to send a few cents instantly and affordably, unlocking microtransactions for digital interactions. A “like” on social media could translate to a 5-cent payment, and users might instantly earn a few dollars for survey responses or for sharing their insights. This micro-economy enables greater engagement and allows individuals to monetize everyday interactions, giving rise to new revenue streams and user experiences across platforms. Integrated Financial Systems In a world where stablecoins are fully interoperable with mainstream payment systems, users can move funds seamlessly across digital wallets, bank accounts, and global payment networks. Stablecoins become an everyday payment option, accepted at point-of-sale terminals, online checkouts, and financial apps alike. This level of integration democratizes access to fast, secure, low-fee transactions across borders, enabling individuals and businesses alike to manage finances more flexibly and inclusively. Imagine the potential for a truly borderless financial system where stablecoins serve as the backbone of global commerce. Programmable Marketplaces with Atomic Escrow Contracts With programmable stablecoins and atomic escrow contracts, marketplaces can be transformed into secure, automated ecosystems where transactions are transparent, self-executing, and fraud-resistant. Picture a smart-contract-powered digital marketplace where assets are bought, sold, and transferred securely, with funds held in escrow until all conditions are met. This programmable layer of trust offers enhanced security and efficiency, fundamentally changing how goods, services, and even data are exchanged. Real-Time Streaming Payments The concept of streaming payments becomes a reality with stablecoins, allowing users to receive earnings in real time. Instead of waiting for bi-weekly or monthly payroll cycles, imagine being paid by the hour, minute, or even second, enabling individuals to access earned wages whenever they need them. This model, enabled by stablecoin infrastructure, could redefine personal finance and wage accessibility, providing greater financial stability and liquidity for workers, contractors, and freelancers. Decentralized Finance (DeFi) Accessible to All A world where stablecoins are seamlessly integrated into DeFi platforms opens up new opportunities for individuals to access savings accounts, loans, and investment products directly from a digital wallet. This ecosystem extends financial tools traditionally limited to institutions or affluent individuals to anyone with internet access, creating an open, decentralized, and more inclusive financial landscape. Cross-Border Remittances and Aid Distribution Stablecoins simplify the process of cross-border remittances, allowing families and communities to send and receive funds with minimal fees and instant settlement. Imagine aid organizations using stablecoins to provide direct, transparent support to recipients globally, bypassing traditional intermediaries and ensuring that funds reach those who need them most. Stablecoins could become a powerful tool for financial inclusion, especially in underbanked regions. The Road Ahead: Sentinel’s Vision for the Finternet Era The Finternet is set to transform how we think about and interact with stores of value. At Sentinel, we bridge the gap between innovation and adoption, guiding enterprises, investors, and individuals toward a future where financial transactions flow as freely and securely as data over the internet. The programmable future of finance is here, with PDDs at its core. By staying ahead of market trends, forging strategic networks, and conducting evidence-based research, Sentinel is leading this financial revolution with our partners and portfolio companies, empowering the world to embrace a digital-first economy. We invite you to join us in building this new future. Reach out at hello@sentinelglobal.xyz  to explore collaboration opportunities.

TL;DR: In this Deep Dive, David Renne and Andrew Reed take a hard look at why supply chain tech hasn’t yet reached its potential. They examine the critical gaps between investment and real-world application, the challenges of integrating AI and automation, and the crucial need for practical, user-focused solutions that inspire trust and deliver clear return on investment (ROI) Introduction Supply chains – the systems behind everything from your morning coffee to global trade – are at a very interesting juncture.  On the one hand, they are on the brink of transformational innovation but on the other, they are mired in outdated practices, fragmented data, and tools that don’t always deliver on their promises. Despite billions invested in the latest technologies, true adoption remains limited, often hindered by legacy workflows and a workforce accustomed to traditional methods. This paper takes a hard look at why supply chain tech hasn’t yet reached its potential. We  examine the critical gaps between investment and real-world application, the challenges of integrating AI and automation, and the crucial need for practical, user-focused solutions that inspire trust and deliver clear return on investment (ROI). Sentinel Global is uniquely equipped to address this topic. Our team members have decades of experience in the supply chain ecosystem, having made significant investments across the sector.  To stay ahead of industry trends, we conduct regular surveys with founders and operators across various stages, gaining valuable insights into adoption patterns. By focusing on tactical learning curves, we have honed our ability to guide successful adoption strategies. Additionally, we maintain a relentless commitment to staying informed about emerging developments and groundbreaking technologies, ensuring our approach remains innovative and impactful. While we’ve seen steady advancements in supply chain technology, we believe substantial opportunities remain as the industry works to reduce cyclicality and build resilience against Black Swan events. Supply Chain in the Early 2020s The COVID-19 pandemic exposed critical weaknesses in outdated and fragile supply chain processes, as country-wide shutdowns and severe labor shortages brought entire industries to a halt.  The obstruction of the Suez Canal   -   a key waterway enabling global trade since the late 1800s - meaningfully disrupted ocean trade in early 2021. More recently, geopolitical conflicts, like the   Red Sea crisis   and an ongoing strife between Russia and Ukraine , have suffocated important trade routes.  Venture money poured into the supply chain technology space hoping to solve the problems, with ~$42 billion invested into the ecosystem in 2021 alone, up 75% from the year prior, per PitchBook Data (link to 2023 Supply Chain Tech report ). Investors placed bets on companies and ideas to transform outdated processes and reduce, if not eliminate, the cyclical gyrations that have long been synonymous with the industry. Based on a 2024 survey by   HERE Technologies , less than 25% of companies across the US, UK, and Germany felt they had made significant progress toward supply chain digitization, despite billions invested in this space and substantial customer spending. In our view, this suggests a major gap between purchase and adoption – a gap that needs to be addressed in the coming era of supply chain technology.  While there was a short period of high deal activity, investment pace has slowed as interest rates have risen and the industry corrected itself from the peak. Despite the influx of capital, the core issues of supply chain resiliency, durability, and interoperability remain unsolved. In our view, one of the key challenges to a tech-driven supply chain revolution is achieving tactical, user-led adoption. This, in turn, limits the progression toward full interoperability in supply chains. The question is: how do we convert begrudging users into supply chain technology evangelists? More importantly, how can we support the industry’s sharpened focus on the end-user experience? User Base and Workforce Complications Given the complexities inherent in transportation and logistics, the industry tends to rightly value years of experience and knowledge acquisition. Trucking, rail, ocean shipping, and air freight networks are complex, fragile organisms that need to be nurtured, which means experience matters.  Let's take the $1 trillion US trucking industry as our first proof point.  Based on the United States Bureau of Labor Statistics data,   52% of the trucking industry   is aged 45 or older compared with   ~30% for software developers . Indeed, the practical realities of training a more-tenured workforce to adapt to new technologies, while certainly not impossible, has proven challenging. In our view, this has created key hurdles for the industry, adding critical re-education requirements alongside the need to demonstrate the value of these emerging technologies. While high level statistics are directionally useful, we like to supplement these views with perspectives from industry practitioners. Conversations with operators reflect the pains felt by entrenched workforces. As one of our contacts in the freight tech space noted, “… our customers have no use for partial solutions. A customer of ours is using some of what we offer to put it into their own spreadsheets to support workflows, through the use of macros and formulas. No one has yet cracked the code of replacing dated working styles, and until that happens, there will be insurmountable headwinds to adoption.”  Think about this for a minute. While industries everywhere are starting to utilize autonomous agents to create workflow solutions with minimal human intervention, the US freight industry is stuck in the 1900s. Transportation analysts across the industry, regardless of their position in the value chain, trust only what they have been doing and seeing for 50+ years. They were trained in spreadsheets, putting the onus on technology providers to prove tangible value.  Due to data silos and data continuity issues – exacerbated by transportation analysts and data scientists moving  data out of technology platforms and into brittle spreadsheets – there remains a major trust gap between data consumers and data providers. For many transportation analysts, efficiency (or inertia) often leads them back to familiar, if outdated, tools that allow them to maximize productivity within their time constraints. We believe many of the challenges facing freight technology adoption are borne out of a highly frictional relationship with its core user base. Clear winners in the ecosystem will only emerge when technology can immediately demonstrate ROI to front-line workers, both at a personal level and to the organization as a whole. All of this is not to ignore one of the bigger elephants in the room: AI.  We would expect headcount in the industry to trend lower over time but mass layoffs are unlikely. Bots and “service as a software” will unfold slowly, at a pace directly tied to the accuracy and reliability of these models themselves. As this happens, revenue per employee will rise and businesses will be more profitable. Budgets could come from software spend, could be net new, or could come from existing labor buckets.  For a workforce reticent to adopting new technology, we are interested to watch dynamics between new advancements and the people they could replace. For decision makers, the key consideration, as ever, is ROI. The rise of “service as a software” creates a delicate balance for operators in the space in the coming years. As we consider this dynamic, we are reminded of this chart from Coatue : Inundated with Tech, not Solutions As noted earlier the fundraising opportunities for supply chain technology were abundant during the early 2020s. The result is that supply chain workers and participants have more tools than ever at their fingertips to mitigate the risks of port disruptions, natural disasters, and pandemics. Shippers, flush with cash  during COVID-19, chose vendors with little diligence in an effort to spend up to elevated budget thresholds. Resultantly, employees of large shippers were left with an abundance of new tools with little training or understanding of how to implement or properly leverage them.  While software products geared toward solving all issues in the supply chain have struggled, we companies focused on solving a single, specific challenge have achieved notable success. This targeted approach allows them to go deep in pursuit of effective solutions.  Take Gnosis Freight , an ocean freight visibility and execution platform, which raised money from Vista   in 2024. Serving one of the most global and interconnected portions of the supply chain, Gnosis has developed a low-code, customized visibility and  container management solution for the ocean industry. Gnosis built deep expertise in the ocean industry and allowed customers to approach the platform in unique ways – allowing for custom solutioning and eliminating the need for additional tools throughout the lifecycle of an ocean shipment “…from booking until returned empty.” The importance of simple but resilient solutions cannot be overstated.  One of the key themes we’ll be watching over the coming years in supply chain will be companies that don’t try to be all things to all people, but be a simple, effective solution for real issues that users experience on a day-to-day basis.  Data Complementarity and Trust Additional key considerations in building the supply chain of the future include data complementarity, interoperability, and breaking down silos.  Consistent with our core values at Sentinel, we believe that open computing and interoperable commerce are critical for the supply chains. The industry, while global, has struggled to conform to consistent data standards (e.g. EDI) and what matters most to one customer may not matter at all to another customer. This is not unique to cross-industry relationships, as customers in the same  industry may hold different data needs as well.  Our work in the space suggests shareability of data remains an issue as well. Take, for instance, a customer in an industry shipping high-value goods. This customer may not want to share data about shipments and orders outside of just a few people in their organization. While frontier technology implementations like zero knowledge proofs could help alleviate this issue, the technology stacks in most supply chain enterprises are not ready to apply such solutions. Contrast this with a customer in a different industry who needs  this data to flow to all participants in their supply chain. To us, it is clear that customers’ data needs and requirements have placed additional constraints on supply chain technology companies’ resources and platform scalability. Lastly, as we noted before, users and consumers must have confidence in the data they rely on for decision-making. Building this trust is essential, as upcoming advancements depend heavily on data integrity and reliability. Founders who can prove to the market that users trust their data, and provide that data in near real time, will be positioned for long-term success. Building the Future Despite some of the complexities we’ve surfaced (and more likely, because of them), we remain excited about venture scale opportunities in supply chain technology. Operators in this industry continue to innovate, building upon the foundation of those who laid the groundwork for technological ingenuity. Supply chains remain high-touch in nature, and the human component cannot be ignored. For this reason, customer- and, more importantly, user-obsessed founders are well-positioned to drive the next decade-plus of supply chain disruption. Here are few key themes we see emerging in the coming years: Breaking down data silos to allow Interoperability of supply chains and the expansion of collaboration between nodes LLM systems / GenAI tools for data/document ingestion and classifications to standardize data AI agents for workflow autonomization built on top of existing systems to decrease stubbornly manual activities (e.g. end to end AI procurement tools) Enhanced computer vision technology that will provide immense safety and productivity gains Increased automation in warehouses from a fully connected suite of IoT devices to packing and picking automation Focus on the developer experience and opening the landscape to non technical adopters (e.g. Low-/no-code applications to bridge the talent gap and improve ROI for companies and their employees) Development of tools that allow shippers, carriers, and 3PLs to take direct (or AI-assisted) action inside of tech platforms, instead of acting on insights through legacy manual processes Sub-sector-specific supply chain technology or technologies Risk mitigation tools such as next-generation technology for aiding in the detection of supply chain fraud Further standardization of supply chain processes In a subsequent blog post, we will be sharing stories that bring many of these examples to life. Conclusion Supply chains are far too important for founders, investors, and industry participants to ignore. We see a bright future for supply chains, both from an investment perspective and for positive global impact. Previous supply chain revolutions have spurred underdeveloped areas around the world, and we anticipate that continued innovation will bring even greater advancements. The next phase of growth depends upon the areas of development we’ve outlined, and we are lucky enough to be in a position to help accelerate this growth. The total addressable market for these problems is vast – finding demand is not the challenge. The real challenge is developing solutions that move the industry closer to a future of more resilient, more adaptable, more adoptable, and more efficient operations across the globe. At Sentinel, we look to support founders who recognize the very tactical realities and issues that supply chain participants face. We are invigorated by the notion that we will partner with those founders who can supercharge the expert, motivated workforce across our global, interconnected supply chain. If you’re a founder, an industry adopter, or a leader in the space, our team would love to hear from you.  Reach out and let’s chat:  Jeremy Kranz David Renne Andrew Reed

TL;DR: Rumi Morales, our Operating Partner and long-time executive in large financial firms, shares the mistaken assumptions that startups make about legacy institutions, the real reasons that corporations hold back from adopting new technologies, and how builders should work with enterprises as a result. It is not the strongest of the species that survive, nor the most intelligent, but the one most responsive to change. – Charles Darwin At Sentinel, we pride ourselves as the bridge for builders and adopters.  The most innovative emerging tech startups and the most established traditional institutions often have a challenging time discerning each other’s capabilities, creating clear expectations, executing on mutually beneficial timeframes, and accepting two drastically different corporate cultures.  What sets the Sentinel team apart is our long tenure as both executives in massive institutions and our embedded presence in startup ecosystems.   I’m happy to be Exhibit A in this regard.  In my whole working life, I’ve spent exactly half of that time at Goldman Sachs  and the CME Group , and the other half in startups and venture capital firms, like Outlier Ventures  and Digital Currency Group , supporting wildly creative technologies.   Here’s one sacrilegious takeaway I’ve come to realize through this experience.   Large established companies are not completely different from startups when it comes to innovation and ambition.   I reject the myth that all corporate dinosaurs will be blown away by a meteor of cool new companies.   What I’m going to attempt to convey in this blog are three main points: The mistaken assumptions that startups have of institutions The real bottlenecks holding back corporate adoption of new technologies How builders should approach institutions when seeking commercial relationships Mistaken Assumptions about Adopters Back when I was running the venture arm of the CME Group, I would come across startups that assumed we desperately needed their help.  Instead, what I realized is that they weren’t aware of the innovations that were happening behind our doors.  I understand it can be challenging for a startup to know this - corporations usually don’t publicly disclose their POCs and technological achievements.  Big firms don’t need the press recognition and brand awareness that a startup might.   So here’s my attempt to shed some light on the assumptions a builder should NOT make about an adopter.  I talked about this in an interview with the Chicago Tribune  almost nine years ago, and  these three points are still worth iterating.  Enterprises need you to solve their problem - WRONG!   Many times, what you think could be their problem simply isn’t.  Or they solved it internally and just haven’t announced it.  The key takeaway is to confirm ahead of time that they have Problem Z but don’t have the internal capacity to address it yet.  Corporations don’t innovate - WRONG!  Check the age of most traditional institutions.  I mentioned the CME Group, which has been around since 1848 through the Chicago Board of Trade.  That’s 176 years.  Do you know how much innovation it takes to keep growing for 176 years ?  Where will your startup be in 176 years? They don’t have the technology talent - WRONG!  Fun fact: 25% of Goldman Sachs’s workforce are developers .  That’s 10,000 people all around the world.  Given as well that the acceptance rate of getting a job at GS is roughly 4% , one would have to humbly consider that those developers could be really good.  So What Does Hold Institutions Back? I’m not trying to pretend that corporations don’t have real challenges with evolution.  In the Dow Jones Index‘s lifetime, there have been 58 changes  to the 30 companies.  Citigroup, Hewlett Packard, AIG, General Electric, and other firms that were behemoths in my early working career have all gone DJIA bye-bye and I know I’m old, but not that old.  So what are the bottlenecks that make it hard for corporations to adopt new technologies?  Here’s a highlight reel. Budgets:  These are often set well in advance of any first meeting a startup will have with a corporation.  If your project is not budgeted for, it will be tremendously difficult to implement Internal Prioritization: Google famously had a 70/20/10 rule where 70% of an employee’s time was focused on core business, 20% on core-adjacent projects, and 10% on unrelated projects. And that’s Google! Many more traditional firms don’t even have that generous of an allocation.  Anyways, for most innovative tech startups, you will fall into that 10% category.  And note that when business conditions get tough, 70/20/10 becomes 98/2/0. Regulations:  it’s both a question about the regulatory clarity around a new technology industry (you) and the existing regulatory restrictions for large institutions (them). Security:  New technologies can introduce new security risks, which institutions will deliberate over and over and over again.  Safety in Stability: “Nobody gets fired for buying IBM” goes the old expression.  It takes a very charismatic, skilled and experienced founder to persuade an executive to take a chance on something quite new. Legacy Systems: They’re like your family at Thanksgiving.  You deal with them, you love them in your weird way, you want to fix them but not totally change them. Similarly, corporations are not going to quickly discard their legacy systems without trying to work within their constraints first. Internal Politics : you can call this corporate bureaucracy too.  Either way, expect it.  Heavily. What do all seven bullet points have in common?  They all lead to a seemingly long amount of time to execute on new innovations. But I used the word “seemingly.”  Bill Gates once famously said, "Most people overestimate what they can do in one year and underestimate what they can do in ten years.”  And I will again draw from my own experience.  Ten years ago, there were only three people out of 2,640 employees at the CME Group who even remotely thought about Bitcoin on a semi-regular basis.  I was one of them.  Now, the CME has nearly doubled its average daily Bitcoin trading volume , from $1.6 billion a day in 2023 to over $3 billion of daily trading volume in 2024. Knowing All This, Let’s Get to Work So with some of the assumptions dispelled and the bottlenecks clarified, how do startups go about approaching and working with enterprises?   Here are a few recommendations. Research their real problems that they can’t solve, especially because of a lack of prioritization, time, and resources - not necessarily a lack of technology expertise. Understand their corporate structure and governance, especially the key decision-makers and budgeting process. Respect their innovation, ambition, and legacy. Survival is hard and they’ve achieved it. Expect to work within their existing tech infrastructure.  Build your tolerance. Have every question around regulation, privacy, safety, security, auditability, accountability, funding, and customer service capabilities ready to be answered. Be honest about your internal financial health.  Working with large enterprises can take a lot of time, always far longer than a startup wants, and the mismatch between getting a revenue-generating project approved by a corporation and a startup running out of runway is very real.  Adjust your expectations and if you don’t have the luxury to  eat the elephant one bite at a time , don’t eat the elephant. This is all my honest advice based on many years in the market.  Yet I’m always willing to be educated and hear your experiences too.  Feel free to connect with me on LinkedIn  or reach out to us at hello@sentinelglobal.xyz

TL;DR:In a recent post, we introduced the concept of “botpower” in the evolution of productivity metrics. Unlike horsepower and manpower, botpower doesn’t measure physical strength but evaluates AI-driven technological capacity. In this blog post, we discuss if we can quantify an equation for botpower, and the roles of standardization, valuation, and corporate leadership in its calculation. One accurate measurement is worth a thousand expert opinions. - Grace Hopper Shifting from the Age of Horsepower to Manpower to Botpower In a prior blog post, I wrote about our concept of botpower.  Read all about it here  and think about it like this: throughout history, people have calibrated value in terms of units of work performed. In the industrial age, “horsepower” represented the might of machines that replaced manual labor. The subsequent shift was towards “manpower” or “workforce,” measured the capability of human workers. I believe we are now in the era of "botpower," pointing to the strength of machine learning and artificial intelligence. This shift makes sense since it's typically the top units of measurement that drive economic and technological progress. Laying the Foundations of a Botpower Equation GIven a general grasp of botpower, how can we more precisely measure AI's impact at work? It will involve calculating and assessing AI's importance and influence on various tasks, merging technology with a lot of assumptions. While any approach will be subjective, including ours, it's crucial to begin somewhere as an actual equation's benefits can be highly impactful. There are reasons for this. Grace Hopper's quote in the beginning of this blog stresses the importance of one precise measurement over many expert opinions. Building business models and financial forecasts are foundational to growing any company – and they require concrete inputs. In our first Botpower post, we mentioned that for numerous companies, the ongoing "build versus buy" debate will fade away. The predominant option will be "build," due to reduced costs and improved integration efficiencies. What should that dollar amount be? Botpower Equation: Deciphering the AI Effort In coming up with an equation, perhaps we should start with a basic scenario.  Let’s consider a situation where an AI system translates volumes of text. The equation could contain several variables: AI Efficiency (E): How quickly and accurately the AI performs the translation task. Task Complexity (C): The intricacy of the content being translated. Scale (S): The number of documents the AI translates. Operational Overhead (O): The cost of maintaining and running the AI system. Developmental Cost (D): The resources spent in building and training the AI. A straightforward formulation could then be: Botpower = E (C S) / (O + D) Easy peasy!  My work here is done! Okay, maybe not, but it’s a start.  We next have questions to explore around standardization, valuation, and the costs of bot-human trade-off, in order to enhance any possible equation.  Can We Standardize Botpower? The Feasibility Debate The concept of botpower standardization raises some interesting questions. Since AI systems are highly specialized and diverse, creating a universal metric that can capture their overall capabilities is not straightforward. Factors like application purpose, training data quality, and integration challenges could introduce significant measurement disparities.  However, considering the rapid progress made in AI development and implementation, standardization is crucial to ensure responsible use of these systems. To address this issue, various organizations have taken on initiatives to establish guidelines and metrics for measuring AI performance. For example, the IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems  has developed a series of ethical principles that aim to guide the design, deployment. Establishing a common framework could help in comparing the performance of AI systems across different applications, facilitating better decision-making for deployment. Though achieving this standardization is no easy feat, the benefits in terms of scalability, interoperability, and innovation could be substantial, driving forward the advancement of AI in many sectors.  Can We Price Botpower? Valuation in the AI Age As we explore the practicality of a botpower equation, here’s another question: can AI's worth be accurately valued? This conundrum boils down to the classic build-versus-buy decision, which ironically is morphing because of AI itself.  In the past, organizations would have to weigh the costs of developing AI capabilities in-house against the convenience and possibly lower costs of purchasing or licensing them from the market.  However, today open computing  is changing all of that. Open computing is the development and sharing of technologies among multiple parties or networks. Through this shared tech, we can alleviate the burden of customized, centralized R&D enhance interoperability across enterprises and tech stacks, and enable easy access, verification, development, and security across multiple participants.   So bringing back our simple equation: Botpower = AI Efficiency (Task Complexity Project Scale) / (Operational Overhead Of Maintaining AI + Development Costs in Building and Training AI) We see that open computing will drive down the denominator significantly as operational overhead and development costs collapse.  However, we are still in the early days of open computing architectures being widely used by enterprises.   Furthermore, as the routine operational, construction, and training tasks may diminish due to AI, there arises a need to consider the necessary oversight, analysis, and internal political management based on its outcomes. It is imperative to evaluate the human costs within a botpower equation, leading to point 3.  Can We Manage Botpower? A Test for Corporate Leadership Botpower will force us to confront  the meaning of work, and the inevitable existential questions in an era where machines take on roles that were once the domain of human prowess and intellect.  How do business management and HR balance the increased efficiency and potential economic growth against the displacement of workers?  Some will argue this is why it's crucial to have proactive workforce development strategies and re-skill employees. This way, when some jobs get automated, employees can undertake new challenges, potentially in more creative, strategic, or complex problem-solving capacities that machines can't easily do. But here are some sober facts: According to a survey  conducted by the Digital Data Design Institute at Harvard’s Digital Reskilling Lab and the BCG Henderson Institute, the average half-life of skills is now less than five years, and in some tech fields it’s as low as two and a half years. For millions of workers, upskilling alone won’t be enough. The World Economic Forum estimates  that upskilling the 1.37 million workers in the US at risk against AI, will require a total investment of US$34 billion.  This translates to approximately US$24,800 per individual. Expanding this effort globally by a factor of 100 reveals the truly monumental scale of resources needed. And obviously the costs vary across industries and roles, leading to a lack of uniformly applicable management practices across enterprises.  Given these time and expense pressures, corporate leaders need to build their strategies now to manage through this evolution. Final Thoughts: Embracing Botpower Quantification Botpower represents a technological leap that should be welcomed and grasped within the wider scope of human experience. There’s no doubt about the challenges however.   Quantifying Botpower and creating a universal metric would be a way to help enterprises manage, forecast, and lead through this change.  Like measuring horsepower and manpower in the past, these abstract measurements eventually found concrete applications and contributed to building the world we know today.  With our quest to measure botpower, we're at a pivotal point of assessing tech innovation, economic growth, and human capital management in one unit.  While we haven’t (and may never) land on a perfect equation, there is no denying that business leaders need solid ground in this age of AI.  Rather than having opinions about what AI means for a company, we need to have the facts.  Attempting to find that right quantification is a way to achieve that. Our exploration into Botpower remains an ongoing journey which requires the input of many minds.  We welcome you to share your thoughts with us at hello@sentinelglobal.xyz