AI PM Accelerator
PM Interview Toolkit
Product Learning Library
Fundamentals of Product Management
PM Communication Toolkit
Being a great AI PM is hard. This newsletter makes it easy.
One clear idea per email. That’s it.
Why the man who invented "vibe coding" hand-wrote his last project — and the six moments where judgement, not the model, decides whether your product wins.
RAG changes what your AI knows. Fine-tuning changes how it thinks. Prompting changes what you ask.
Prompt injection is the biggest open security hole in AI. Here's why your AI can't tell your instructions apart from an attacker's — and what to do about it.
Airbnb deleted the guest fee at checkout and folded it into a single host-only fee near 15.5%. How the change quietly earns Airbnb more on every booking.
OpenAI's Sora hit No. 1 on the App Store and landed a $1B Disney deal, then shut down six months later. The story of why a viral hit got too expensive to survive.
Perplexity owns neither the models nor the content. Here's how that became a $22B strategy and why the open web is now suing it.
You can't human-review every AI output at scale. So you have one model grade another against a rubric. How LLM-as-a-judge works, and when to trust it.
Why do some AI products get smarter the more they are used, while others stall? It's because of the data flywheel: the loop that separates a feature from a moat.
Uber built an AI "PRD Evaluator" that assembles company-wide context to pressure-test product docs before human review. Here's how the system works.
Bigger context windows didn't fix AI quality but made it easier to overload the model without noticing. Here's what context rot is, why it happens, and the TRIM framework to fix it.
I've spent the last 6 months diving deep into building with Claude Code. And now I've created a FREE course that teaches you Claude Code inside Claude Code.
Most AI training data is no longer written by humans. Synthetic data, generated by AI models, is faster, cheaper, and often better. Here's how it works.
