The video explains how YC built superintelligence within their company by deeply integrating AI as a foundational layer, creating a transparent, trust-based infrastructure that empowers all employees to automate complex workflows and continuously improve through shared AI tools and skills. It emphasizes the importance of decentralization, openness, and cultural change to transform AI from a developer-controlled tool into a democratized, augmentative system that enhances collective intelligence and productivity across organizations.
The video discusses how to build superintelligence within a company by deeply integrating AI as a foundational layer rather than just a copilot tool. Pete Kumman, a general partner at YC and creator of Optimizely, shares insights from YC’s journey in adopting AI internally. YC has developed an extensive AI infrastructure that empowers teams across the organization by enabling them to encode workflows in natural language, reducing reliance on software engineers for every task. This approach started with finance tools and evolved into a general agent loop with a shared tool registry, allowing non-technical employees to query databases and automate complex workflows efficiently.
A key advantage YC had was running its own software on a unified PostgreSQL database containing all critical organizational data. This centralization allowed AI agents to access comprehensive context, dramatically increasing the volume and complexity of questions employees could ask without bottlenecks. The system’s openness and transparency, where agent conversations are viewable by all employees, fostered a high-trust, egalitarian culture that accelerated learning and adoption. This environment contrasts with many legacy organizations that remain siloed and cautious, highlighting the importance of trust and openness in building AI-native companies.
The infrastructure YC built includes a tool registry with over 350 tools tailored to various teams and workflows, enabling agents to perform diverse tasks from managing office hours to booking financial entries. They also developed “skills,” which are abstractions over tools that can self-improve through autonomous loops analyzing past interactions. An example is a skill that helps founders craft concise two-sentence company descriptions, which has evolved to outperform humans by learning from partner feedback and meeting transcripts. This iterative improvement exemplifies how organizations can embed collective knowledge into AI, creating a shared organizational brain that enhances communication and decision-making.
The conversation also touches on the broader implications of AI adoption, emphasizing the shift from deterministic software controlled by developers to AI-native systems where users have more control. The speakers argue for decentralization and personal agency in AI, warning against centralized control by a few corporations. They envision a future where individuals and organizations run customizable AI agents that integrate seamlessly with their data and workflows, enabling just-in-time software that adapts dynamically to user needs. This democratization of AI tools is likened to the personal computer revolution, empowering users rather than confining them to rigid, corporate-controlled systems.
Finally, the video highlights cultural and organizational changes necessary to realize AI’s full potential. Companies must embrace transparency, trust, and investment in AI infrastructure to raise the baseline capabilities of all employees, enabling faster onboarding and more effective collaboration. The speakers caution against viewing AI as a replacement for humans, instead framing it as a powerful augmentation tool that eliminates drudgery and unlocks creativity. They encourage leaders and builders to choose openness and empowerment, shaping AI’s role in organizations to foster collective intelligence and supercharged productivity.
