Product Thinking

Essays on shipping AI-first products

Notes on discovery, prompt engineering, retention, analytics and decision-making — pulled from real products I've shipped.

9 / 9

AIJul 2026

Windsurf vs Cursor: an AI PM's take

Both Windsurf and Cursor are strong AI-native IDEs, but they optimise for different product jobs. Here's how I'd choose between them as an AI PM shipping real product.

7 min readRead →

PlaybookMay 2025

How I build MVPs in 6 weeks

A repeatable operating cadence for going from a fuzzy problem to a working, validated product in six weeks — without heroics.

6 min readRead →

AIApr 2025

AI Product Development in 2025

Building AI-native products is less about the model and more about the loops around it — evals, prompts-as-spec, and human-in-the-loop.

8 min readRead →

DiscoveryFeb 2025

Customer discovery without a research team

You don't need a UXR org to do great discovery. You need 5 interviews a week and the discipline to write the notes.

5 min readRead →

StrategyJan 2025

Choosing a North Star metric

A good North Star is a leading indicator of retained value, not a lagging vanity number the CEO likes on a slide.

4 min readRead →

AIDec 2024

Prompt engineering as product design

The system prompt is your product's constitution. Treat it like the most important UX artefact you own.

6 min readRead →

GrowthNov 2024

Retention loops that actually compound

Acquisition without retention is a leaky bucket with a fancier faucet. Design the loop before you turn on ads.

5 min readRead →

AnalyticsOct 2024

Product analytics on a startup budget

You don't need Amplitude on day one. You need three events, a spreadsheet, and a weekly cohort review.

4 min readRead →

LeadershipSep 2024

Decision-making under ambiguity

Most PM decisions aren't right or wrong — they're reversible or expensive. Treat them accordingly.

6 min readRead →

Back Home