🤯 How Generative AI Changes The Game

In the past 25 years there have been multiple waves of innovation—Internet 1.0, the rise of cloud computing, the mobile revolution. Each has changed the way we build and deliver products. Yet none of these previous transformations fully prepares us for the radical shift that Generative AI brings to the table. We cover:

• How generative AI is not like previous technology shifts

• How AI impacts the different types of product work (Feature, Growth, Scaling, PMF)

• How AI changes the core responsibilities of a product manager

• How AI changes customer expectations and behaviors

🧠 Base: LLMs, Foundation Models, Base Capabilities

The gap between success and failure in AI isn't just technical - it's about deeply understanding what's possible, what's coming, and what's better bought than built. With each new capability or architectural improvement in AI, your foundational knowledge will help you adapt and leverage these changes faster than competitors who never took the time to learn the basics. We cover:

• How Gen AI models work at their core

• What are their critical limitations across knowledge, reasoning, tools, memory, and more

• How to think about Gen AI as a core building block to upgrade and enhance from

🛠️ Upgrade: Knowledge, Reasoning, Tools, Memory

The base capabilities of Gen AI models have critical limitations. In order to know how to craft the AI product experience we want, we need to understand all the ways to upgrade/enhance the base models. We cover:

• Knowledge Augmentation: Naive RAG, Enhance RAG, Modular RAG and their components (Embeddings, Vector Databases, and more)

• Reasoning: Type 1 vs Type 2 , Chain of Thought (CoT), Chain of Thought - Self Consistency (CoT-SC), Auto Chain of Thought, Tree of Thoughts (ToT), Graph of Thoughts (GoT)

• Memory: Conversation Buffer Memory, Summary Memory, Entity Memory, Long Term Memory, Memory Decay

• Tools: Types of Tools, How to Integrate Tools, Tool Orchestration

📈 Improve: Learning, Human Feedback, Fine-tuning, Evaluation

One of the most powerful parts of AI products is their potential to learn and evolve over time. We cover:

• Evaluation: Methodologies on measuring the success of AI models and features

• Human Feedback: Methodologies on providing our models with human feedback to learn and improve

• Fine-Tuning: How to systematically improve models for specific and differentiated use cases

• Learning: RLHF and how models can learn autonomously while still incorporating human preferences

🎯 Lead: Agents, Goals, Decision Making, Autonomous Behavior

AI agents represent a fundamental shift in how software can operate autonomously while still aligning with human intentions and business objectives. As product leaders, understanding how to design and deploy agent-based systems is crucial for creating truly transformative products. We cover:

• How to define and structure agent goals that balance autonomy with control

• The key components of effective AI agents: planning, execution, and reflection capabilities

• Strategies for designing agent decision-making frameworks that maintain user trust

• Methods for implementing appropriate safeguards and oversight mechanisms

• How to evaluate and iterate on agent performance from a product perspective

🤝 Delegate: Multi-agent Coordination & Workflow Management

The future of AI products lies not just in single agents, but in coordinated systems of specialized agents working together to accomplish complex tasks. Understanding how to orchestrate these systems is crucial for product leaders building sophisticated AI solutions. We cover:

• Frameworks for breaking down complex tasks into multi-agent workflows

• How to design effective communication protocols between agents

• Strategies for managing resource allocation and priority across agent systems

• Methods for handling conflicts and edge cases in multi-agent scenarios

• Best practices for monitoring and maintaining multi-agent systems

• How to balance automation with human oversight in complex workflows