The speaker, Kim Seo-joon of Hashed, discusses a transformative shift in entrepreneurship and society driven by AI and agentic systems, terming it the emergence of an "agent-native" era.

His "click" moment occurred after observing a non-developer, Jae-hong (Across), successfully create the GPTO service by reverse-engineering LLM web search and ranking algorithms. This demonstrated that individuals without traditional development backgrounds could leverage AI for highly technical tasks. Kim Seo-joon personally experienced this paradigm shift with Claude Opus (4.5), where he, a long-retired developer, rapidly built an Ethereum valuation site and an Abu Dhabi travel guide. This demonstrated that complex applications, which previously required teams and months, could be prototyped by an individual in hours, leading to his realization that "VCs are doomed" under the traditional model.

This new reality fundamentally alters the venture capital landscape. Hashed's strategy now focuses on providing three core values to AI-native founders:

1. AI-native Mentorship: Providing guidance from individuals who deeply understand "vibe coding" and agent ecosystems, speaking the same language as these builders. This necessitates all Hashed team members to engage in "vibe coding."
2. Network and Trust Shortcuts: In a world where "what's visible is open source" and technical solutions can be rapidly replicated, the value shifts to connecting founders with global stakeholders and providing strategic business development.
3. Community Building: Creating a peer group for top-tier AI-native builders who experience intense excitement and anxiety, fostering a collaborative environment for idea exchange and mutual support.

The characteristics of these AI-native talents often diverge from traditional profiles. They might come from specialized technical high schools rather than elite universities, as the traditional functions of universities (selection, education, community) are being unbundled by platforms like GitHub, LLMs (replacing MOOCs), and intentional, self-organized meetups. Their core value lies in "intention" ($I$) and "judgment" ($J$), as execution is increasingly offloaded to AI agents. They operate with a "no is not an option" mindset, embracing rapid experimentation across domains.

The "agent economy" envisions a future where individuals and organizations leverage AI agents as extensions of themselves. A 20-person VC firm, for instance, might aim for the performance of a 1,000-person enterprise by having agents handle repetitive tasks, freeing humans for higher-order problem-solving and judgment. This model suggests that human activity will increasingly converge on "intention" ($I$), with agents handling the "real work" or "execution" ($E$).
Let $A$ be the set of AI agents, and $H$ be the set of humans.
In the traditional model, human output $O_H = f(I_H, E_H)$, where $I_H$ is human intention and $E_H$ is human execution.
In the agent-native model, human output becomes $O_H = f(I_H, E_A)$, where $E_A$ is agent execution. The human role shifts to maximizing "judgment density" ($D_J$), "accuracy of judgment" ($A_J$), and "responsibility" ($R$).

Blockchain plays a crucial role in the agent economy by providing the necessary infrastructure for trust and identity. Just as corporations have legal personhood, agents will require a system for identity, reputation, and verifiable transaction history. Standards like ERC-8004 are being developed to "register" agents, record their actions, and track their performance. This transparent, immutable ledger, superior to centralized databases, is essential for enabling a secure and trustworthy economy where agents can perform transactions and interact. Even major tech companies are collaborating on these blockchain-based standards because the inherent trustlessness of blockchain is deemed more reliable than any single entity's centralized server.

Regarding future business opportunities, Kim Seo-joon envisions an abundant society where AI and robotics automate production to such an extent that basic needs are met for everyone, shifting the societal focus beyond mere survival. In this future, the primary drivers of value will be "taste" ($T$) and "inspiration" ($I_{sp}$). He illustrates this with observations from Abu Dhabi, where individuals, free from economic constraints, invest in highly specific "tastes" (e.g., collectible trading cards). Businesses will gravitate towards the "inspiration" axis, away from the "utility" axis, as new technologies rapidly commoditize. While utilities like electricity and communication began as inspirational marvels, they ultimately became commodity services. The enduring value will be in content and narrative creation, which humans uniquely provide. He notes his firm's venture into a fund that invests solely in trading card games (TCGs) as an example of investing in "IP" and "inspiration," viewing it as a precursor to digital-native assets and communities that provide value beyond pure utility.