0:00
/
0:00
Transcript

The Strategy Stack - Tech Breakdown (With Guest: Ryan Ong)

Episode 5: Fast Code, Slow Innovation: The Enterprise AI Paradox

Strategy Stack: Tech Breakdown

👥 Featuring:

  • Guest: Ryan – AI engineer, PhD in Knowledge Graphs & LLMs, founder of Weekend AI and Substack writer.

  • Host: Alex (The Strategy Stack)

Get more from Alex Pawlowski in the Substack app
Available for iOS and Android

Key Discussion Topics

  • Enterprise AI Trends

    • Ryan shares consulting experience with UK-based clients deploying GenAI for internal tools and chatbots.

    • Highlights tension between software engineering speed and lack of innovation in traditional orgs.

    • Challenge of evaluating LLM performance in dynamic, non-deterministic workflows.

  • Tooling & Automation

    • Multi-agent systems and RAG (retrieval augmented generation) in production.

    • Launch of Us, a Substack-based chatbot using RAG to answer questions from newsletter content.

    • Importance of "human in the loop" design to keep AI outputs reliable and verifiable.

  • Tech Adoption & Global Shifts

    • Comparison of GenAI adoption: India/China (high), US/Europe (relatively low).

    • Ryan discusses relocating to UAE or Singapore for better tech ecosystem support.

    • UK seen as increasingly less innovation-friendly.

  • AI Productization & Platform Gating

    • Discussion on how AI agents may compress traditional business workflows.

    • Questions around platform securitization and creator compensation in a world of AI scraping.

    • Brave browser and Model Context Protocol (MCP) as emerging solutions for access control.

  • Future of AI Interaction

    • Excitement over Google’s VEO-3 (video + audio generation) and real-time language translation.

    • Debate around multi-modality, sensory inputs, and brain-AI interfacing (Neuralink, vision, etc.).

    • Limitations of today’s models and the need for spatial or context-aware AI co-pilots.


💡 Takeaways

From Alex:

  • Fuzziness is inherent to generative AI, but human-in-the-loop workflows and tighter platform access controls will define the next era.

  • AI needs world understanding before making judgments—text is not enough.

  • Sovereignty, privacy, and compensation will shape the evolution of AI agents and data-sharing platforms.

From Ryan:

  • Users don’t care about backend complexity, they care about fast, intuitive value delivery—even from "naive" systems.

  • Simple tech can outperform advanced systems, if user outcomes are prioritized.

  • Adoption depends on role and region: AI as Google replacement vs. life advisor vs. operating system.

Discussion about this video