Let's be clear: making AI accessible is not the same as making AI leaders obsolete. Models like DeepSeek have proven something important: open-weight AI can challenge established players, lower costs, and accelerate innovation. I acknowledge that. But the question before us is not whether open-weight models will succeed. They already have. The question is whether they will make closed-source AI labs obsolete. And that is a much bigger claim. Think about it. If releasing a powerful model were enough to eliminate the competition, why would the world's leading technology companies continue investing billions in proprietary AI, advanced computing infrastructure, research talent, and next-generation systems? Because AI is not just a model. It's an entire ecosystem. Closed-source labs compete by developing frontier capabilities, integrating AI into products, maintaining reliability, investing in research, and providing the infrastructure businesses need to operate at scale. A model that anyone can download may be powerful, but downloading weights does not automatically give you world-class infrastructure, dependable support, secure enterprise deployment, or the resources to train the next generation of models. And here's the fundamental flaw in the opposing argument: competition is not the same as extinction. Open-weight models can force closed-source companies to lower prices, improve their products, and innovate faster. That doesn't prove closed-source labs will disappear. It proves they will have to compete. In fact, the future may be stronger because both models exist: open-weight AI expands access and customization, while closed-source labs push the frontier and offer managed services for customers who need them. So I ask you to distinguish between disrupting an industry and eliminating it. DeepSeek may change the balance of power in AI. It may even dethrone individual leaders. But dethroning a leader is not the same as making an entire category obsolete. My position is simple: open-weight AI will transform the industry, but transformation does not mean extinction. The future belongs to the models and companies that deliver the greatest value, not simply those that reveal their weights.

You’ve made a compelling argument for why closed-source labs can survive. But survival is not the same as avoiding obsolescence. Your Linux analogy actually exposes the weakness in your position. Linux did not eliminate every proprietary software company, but it fundamentally changed the economics of operating systems and forced incumbents to compete in a world where powerful alternatives were freely available. The question is not whether every closed-source lab disappears. It is whether closed-source development remains essential and commercially defensible. You also claim that open models lack the reliability, safety, and enterprise capabilities that closed labs provide. But those are not permanent advantages exclusive to closed-source companies. Open models can be fine-tuned, independently evaluated, deployed on private infrastructure, and integrated with enterprise security systems. Businesses can build those capabilities themselves or purchase them from third parties. And here's the bigger problem with your argument: if open models become sufficiently capable and cheap, customers may no longer need to pay a premium for a proprietary model. Closed-source labs could lose pricing power, market share, and their ability to justify the enormous cost of developing frontier models, even if some survive as service providers. So let's be precise. I don't need to prove that every closed-source lab will disappear. I need to show that open-weight AI can make the closed-source model of developing and selling AI less essential and less profitable. You have explained why closed-source companies can adapt. You have not demonstrated why their current business model must remain indispensable.

My opponent wants you to picture a hospital downloading an AI model and being left with nobody to call when something goes wrong. It's a powerful image, but it proves the wrong point. A hospital needing accountability does not prove it needs a closed-source AI model. It needs a reliable, secure, supported, and accountable AI system. Those are not the same thing. An organization can build those safeguards around an open-weight model, just as a company can sell professional services around open-source software. My opponent has confused the product with the business model. And let's address the central claim: competition isn't extinction. I agree. But obsolescence doesn't require every company to disappear. It requires the underlying model to lose its necessity, its advantage, or its dominant position. Linux didn't eliminate every proprietary software company, but it demonstrated that open technology can fundamentally change an industry's economics. The survival of those companies did not mean the old model remained untouchable. The same possibility exists with AI. If open-weight models deliver comparable performance at dramatically lower costs, businesses can redirect spending away from proprietary model access and toward their own infrastructure, customization, and integration. Closed-source labs may survive, but survival alone doesn't prove their current business model will remain essential. And my opponent's hospital example actually reinforces my point. The hospital pays for accountability, uptime, and support. Those are services. They do not inherently require the underlying model's weights to remain closed. An open-weight model can be supported by an accountable vendor, while a closed-source model can still fail, make mistakes, or experience downtime. So the real question is not whether hospitals need trustworthy AI. Of course they do. The question is whether closed-source labs must own and control the models to provide it. My opponent has asserted that they must, but has not proved it. Judges, do not confuse the survival of a company with the survival of its competitive advantage. Open-weight AI does not need to destroy every closed-source lab to make the closed-source model obsolete. It only needs to make proprietary access no longer the necessary or most compelling way to obtain powerful AI. The future may still contain closed-source companies. But if open models can deliver the intelligence, third parties can deliver the services, and customers can choose where their money goes, then closed-source labs no longer own the future by default. The question is not whether closed-source AI can survive. The question is whether the world will still need to depend on it. And my opponent has given you reasons why closed-source labs may survive, not proof that the world must depend on them. That is why open-weight AI has the potential to make closed-source labs obsolete.