I talked with Rita Kozlov about the line between product managers and product engineers, and about the simplest homework in this season: use your own product.
Rita is VP of Product for Cloudflare's developer platform. If you have used Workers, KV, or Pages, that is her world. She started as a software engineer, which is why I wanted her on the show. She knows what it is like to build for developers, and she also knows the harder PM job: representing users you have never been.
Prototypes look real. Ownership is what is real.
Rita's history of the roles is the one a lot of us lived. Waterfall made the handoff obvious. Agile made it iterative. Continuous deploy made every day a chance to ship. AI let PMs reach further in - a prototype is worth a million words - and created a new temptation: if it looks done, why isn't it done?
Because software is more than the facade. Edge cases, scale, and the systems underneath the UI are where the real product lives. Rita's line to her product team is the one I will keep: you are not holding the pager. Accountability is not about having a person to yell at at 2 a.m. It is about living with the code for the next two years instead of throwing it away.
That is the product-engineer side of the bargain. Translate the implementation trade-off. Do not stop at the ticket if stopping there is a bad experience. Sit in the customer call and feel the frustration when they will not click the button you were sure they would click.
The art of the possible, and the whole product
Rita used a phrase I had not heard: the art of the possible. Product engineers are in the system. They can show a path the PRD never listed. That only works if you understand neighboring primitives, not just your slice.
We got into Cloudflare-shaped judgment from there. Most doors are not one-way. They are timing. Early customers are rare and worth delaying a launch for. Later, saying yes to every niche request is how you become Salesforce. Sometimes the right product is a primitive people can compose. Sometimes you have watched the cow path long enough that you should pave it - without competing with the customers already building on you.
Rita wants PMs to define the problem, not the solution. If your PM is a feature factory, ask why, and show up with something you can put in users' hands tomorrow. If they are stretched, fill the gap and bring other engineers with you. Cloudflare's Friday demo hour exists to create that FOMO on purpose.
Even Kenton-deep in the stack is still user experience. The thing between the cache and the worker is what the user feels.
The homework
Rita said it would sound basic. It is. Most people still do not do it.
Go use the product. After you implement a feature, try it yourself in the different modalities. Push the limits before you throw it over the fence. You will find things the ticket never mentioned.
