Podcast Summary: Work For Humans
Episode: Designing AI Tools That Think With You | Dmitri Glazkov
Date: October 21, 2025
Host: Dart Lindsley
Guest: Dmitri Glazkov, Strategy Lead at Google Labs
Episode Overview
This episode explores how artificial intelligence tools, particularly Google Labs’ Breadboard and Opal, can be designed to augment creativity, harness tacit knowledge, and reshape the experience of work for individuals and organizations. Host Dart Lindsley and guest Dmitri Glazkov discuss the shift from traditional automation to tools for creative assembly, the difference between “dandelion” and “elephant” organizational strategies, extracting and sharing tacit knowledge, and how AI tools might capture the unique “variation” each person brings to their work.
Key Discussion Points & Insights
1. Origins and Philosophy of Breadboard and Opal
- Google Labs' Mission: A small, experimental team tasked with exploring new frontiers for Google, focusing on flexibility and fun.
“Our job… is to find new interesting opportunities for the company to pursue… looking around the corner is what we’re about.” — Glazkov [04:41] - Breadboard Concept: Inspired by electronics prototyping boards, but here, "components" are units of cognition or prompts, assembled to create multi-step cognitive workflows. “With a breadboard, you assemble an electronic device. With Opal, you assemble a tiny brain.” — Glazkov [09:38]
- Opal’s Design Ethos: Launched with minimal fanfare to attract “the right” users and remain flexible for rapid iteration, resisting early pressure to lock into feature requests for automation.
- Distinguishing from Traditional Automation: Unlike process automation, these tools are intended to be fun, expressive, and creative—empowering users, not just automating away their tasks.
2. Modular AI: Prompts as Cognitive Building Blocks
- Unit of Composition: One prompt to an LLM is the atomic “block.”
“We ended up with something very simple… The unit of composition is one call to an LLM, and then stuff grows from there.” — Glazkov [15:12] - Metacognition by Composition: By chaining together prompts, users can create systems that “think about thinking”—mirroring human strategies for examining problems from multiple angles.
- Practical Applications: From basic tasks (like assembling biographical summaries) to complex, multi-lens analyses in fields like international law (as with Anthea Roberts’ use-to find consensus between models), the tools encourage iterative, exploratory work.
3. Tacit Knowledge and the "Avatar" of Self
- Capturing Individual Expertise: The most transformative potential lies in encoding the tacit, personal knowledge of individuals—what makes our thinking unique—into repeatable workflows for others, or even as digital avatars. “There will be a moment where you go, you know what? I actually want to share this with the world because I think the world will be a better place if this wisdom is shared out there.” — Glazkov [21:08]
- LLMs and the “Average”: LLMs are fundamentally “average machines”—the value added by a user is their specificity, variation, and unique context. “The value I bring to tasks with AI is my variation from the average.” — Lindsley [24:29]
- Future of Personal AI: Lindsley posits that future professionals will assemble their own “training data” as they move through education and work, raising questions of ownership and privacy.
4. Lensical Thinking
- Definition: The ability to consciously use different conceptual lenses (frameworks, perspectives) to interrogate a problem, rather than defaulting to a single viewpoint.
- Embedding in Tools: Tools like Opal allow explicit switching between such “lenses” for richer, more thorough analysis. “When you’re facing a challenge… have a bag of lenses… that allow you to just switch them around and say, if I think about it this way, is this that or is this this?” — Glazkov [28:16]
5. Organizational Metaphors: Dandelions vs. Elephants
- Biological Analogy:
- Dandelion (R-selected): Rapid replication, low cost per experiment, high variation—analogous to startups and experimental teams.
- Elephant (K-selected): Slow, high-investment, knowledge passed down deliberately—analogous to stable, large organizations.
- AI’s Moment: AI enables “dandelion” experimentation due to low costs of iteration. “With AI, because the medium is so cheap, will we ever get to the point where there’s elephants?” — Glazkov [36:22]
- Strategic Aperture: The capacity for organizational change narrows as products or businesses mature and accumulate value; early-stage “dandelions” have agility, but growth brings rigidity. “One thing I’m trying to avoid is… getting locked into a very particular shape… That’s not what we want to do.” — Glazkov [07:59]
6. Building Creative, Adaptable Teams ("Dandelion Farms")
- Pirates vs. Sailors:
- Pirates: Seek opportunities, bend rules as necessary, thrive in uncertain environments.
- Sailors: Seek order, expertise in maintaining stable systems.
- Ideal Team Mix: Experienced “pirates” who know the rules well enough to judiciously break them. “First, you don't know what the rules are. Second, you know what the rules are. Third, you know which rules to break.” — Glazkov [44:10]
- Cultural Requirements: Balance of agency (autonomy) and belonging. Mental flexibility (“holding their ideas lightly”). “Lensical thinking is one of the key elements of the organization.” — Glazkov [48:12]
7. Embodied Strategy and Pace Layers
- Embodied Strategy: An organization’s true strategy is often implicit, reflected in unspoken processes, accumulated habits, and cultural DNA, not just formal plans. “Strategy picks you in some ways. The different bits of culture you have embedded… define what the strategy will be.” — Glazkov [54:40]
- Pace Layers (from Stewart Brand): Different organizational functions change at different rates—variation can be advantageous at the edges (e.g., product design), but stability is critical at the core (e.g., payroll).
Notable Quotes & Memorable Moments
- On the “Tiny Brain” Metaphor:
“With a breadboard, you assemble an electronic device. With Opal, you assemble a tiny brain.” — Glazkov [09:38] - On the Purpose of Work:
“I think it’s purpose. I’d like to wake up every morning and have a sense… of something I’m doing that directly contributes… to the arc of humanity.” — Glazkov [60:18] - On Variance as Value:
“The value I bring to tasks with AI is my variation from the average.” — Lindsley [24:44] - On the Limits of LLMs:
“Humans think and reason… no matter how much we want to believe, LLMs don’t think or reason. They perform some process… we can call thinking or reasoning… but it’s not really that process.” — Glazkov [14:34] - On Building Dandelion Organizations:
“If you are to start a new organization that is… a dandelion farm, you have to find this fairly rare breed of people who… have the mindset of being a pirate but have the wisdom of having been in the Navy.” — Glazkov [47:23]
Timestamps for Important Segments
- [04:41] — Glazkov explains Google Labs, Breadboard, and Opal’s origin.
- [09:38] — The tiny brain analogy: conceptualizing Opal as cognitive assembly.
- [15:12] — Prompts as atomic units; metacognition by composition.
- [21:08] — The vision for capturing and sharing tacit knowledge with AI.
- [24:29] — Lindsley highlights the individual’s “variance” as key AI input.
- [28:16] — Definition and value of lensical thinking.
- [34:06] — Elephant/dandelion (K/R selection) strategy explained.
- [36:22] — AI as an enabler for dandelion strategies.
- [44:10] — Pirates vs. sailors: behaviors and metaphors for innovators.
- [54:40] — Concept of embodied strategy in organizational change.
- [60:18] — Glazkov’s personal purpose in work.
- [65:43] — Glazkov on adapting work environments (“more whiteboards!”).
Further Resources & Where to Find More
- Breadboard (open-source): GitHub — Breadboard
- Opal: opal.withgoogle.com (currently US-only, limited cohort)
- Dmitri Glazkov’s Blog: glazkov.com
Character & Tone
The conversation is rich with playful analogies (dandelions, elephants, pirates, and sailors), candid reflection on the design and impact of AI, and an infectious excitement for building expressive, user-empowering tools. Both host and guest display a bias for curiosity, experimentation, and humility in navigating fast-moving technological terrain.
Summary
This episode acts as both a deep dive into the mechanics and philosophy behind cutting-edge AI tool design at Google Labs, and a broader meditation on how organizations and individuals might evolve—balancing structured expertise with continuous experimentation, leveraging both human uniqueness and scalable cognitive machinery. It is essential listening for anyone interested in the intersection of work design, organizational psychology, and artificial intelligence.
