PM → AI-aware product thinking
For: A product manager who needs to ship AI features that work, scope what is realistic, and avoid the common "we built an AI feature and it embarrassed us" outcomes.
You will be able to: You can scope AI features realistically, choose the right architecture (prompting / RAG / fine-tuning / agents), evaluate whether they are working, and avoid the predictable failure modes.
Steps in order
- 01
How large language models work
The mental model that fixes most prompting confusion — prediction, training, inference, why hallucinations happen, why prompt phrasing matters so much. For the operator who wants to understand the mechanism.
— The mental model. Without this, scoping is guesswork.
10 min·/learn/ai/foundations - 02
What AI is good at, and what it still gets wrong
A blunt capability map. The categories of work where AI is reliable, the categories where it bluffs, and the in-between where it works if you verify.
— The capability map. Knowing the no-go zones is half of feature scoping.
7 min·/learn/ai/get-started - 03
Fine-tuning vs RAG vs prompting: which one fits your problem
The three ways to make a model behave better for your case — cost, persistence, updateability, when to use each, and when to mix them. With the decision matrix and the math for "is fine-tuning worth it."
— The architecture decision. Most "should we fine-tune?" debates resolve here.
9 min·/learn/ai/foundations - 04
How AI agents work (and where they break)
The minimum that makes something an agent (LLM + tools + loop). What agents are good for, the six predictable failure modes, the autonomy spectrum, multi-agent vs single, and what to log in production.
— When to build an agent vs a workflow. The failure-mode section is your scoping checklist.
10 min·/learn/ai/foundations - 05
Context windows explained: what they limit, what they do not
How much text a model can consider at once, what counts against the window, why quality degrades long before you hit the limit (the "lost in the middle" effect), and how to budget context in production.
— Why "just put it all in context" stops working at scale.
8 min·/learn/ai/foundations - 06
How to evaluate an LLM feature is working (without fooling yourself)
Why "looks good" is not evaluation. Building a small eval set (20 cases beats 200), the four grading methods (programmatic, reference, LLM-as-judge, human), what to measure, and how to spot production drift.
— The mechanism that prevents your AI feature from silently degrading after launch.
9 min·/learn/ai/foundations - 07
AI for product managers: PRDs, research, roadmaps
Drafting PRDs from rough ideas, synthesizing user interviews, building roadmap narratives. Where AI helps and where it actively misleads.
— The applied playbook. Brings the foundations together for the PM job.
7 min·/learn/ai/use-cases-by-role
When you finish this path
You can scope AI features realistically, choose the right architecture (prompting / RAG / fine-tuning / agents), evaluate whether they are working, and avoid the predictable failure modes. For the next step, browse other paths or the full library.