Weeknote #35 [W26.26] - Product Management with LLMs, and half-yearly reading and watching recap

Product Management with LLMs

I’ve been working as a Product Manager at work for the last two months.

I think this is probably the best time to be a PM because of LLMs.

The most fun part of the role right now is the amount of autonomy I have relative to when I was working as a product designer. I can take up a project and just get it done end to end in a way that I couldn’t before.

Last week, I took up a project for rolling out an NPS survey to our users. This is something I had been wanting to do for the last six months. But as a designer, it is difficult to get things like these prioritized — there’s always more urgent issues that needs taking care of. And it would also need developer bandwidth, which is always in short supply.

But now that I am a PM, and also have Claude, I can just decide to take something like this up and run with it.

The reason I took up the project was mainly so that I could also make an analytics dashboard to go with it. I like looking at data. I spent my childhood making Excel sheets out of everything I could think of. I started a coin collecting hobby just so that I could maintain an Excel inventory of all the coins I had, with tables and charts for which year and mint each one of them was from. I like looking at tables and charts. And I like presenting data in a way that is clear and useful. So this was also just an excuse to (i)indulge myself, (ii)find a distraction from otherwise not so interesting work.

Anyway, what LLMs made possible here were faster and cheaper iterations. Because of which I worked on it in a way I have not done before in a work environment. If this had gone through a normal workflow, we would first have had to write a PRD, figure out all the edge cases, make designs end to end, and so on. But since this was unlikely to break anything critical, I skipped all that. I started with rolling out the survey form directly, which took not more than a handful of prompts on Claude. Then I made a very simple dashboard design with Claude Design (which is a tool I have changed my mind over a couple of times already — from feeling it was completely useless, to a cute gimmick, to now feeling that it is almost essential; maybe I’ll write more on that in a later note), and then kept adding things based on feedback from the team about how presenting the information would be most useful.

It takes less time now to build something and deploy it to production now than it would previously take to write a PRD for it. That, to me, is a very exciting. For I have always hated PRDs and thought of them as a waste of time. A functioning prototype is, in most cases, more useful than a PRD. (I don’t think this applies to everything, of course. For anything critical, where shipping the wrong thing could break things, I think we should still need to be slow and considered.) And there are many side projects which used to remain stuck at the bottom of the pile because they were never urgent enough. These can now be taken up, shipped, and iterated with very quickly.

Half-yearly reading and watching recap

I read a lot last week because I realized that June was coming to an end and that I was way behind my yearly reading target. I think I ended up finishing three books just last week.

I read a lot on the weekdays whenever I got a chance, plus I also went to Cubbon Reads again on Saturday (where I pleasantly ran into Santrupti, Jatan, and Devanshi).

Since half the year is now over, this seems like a good point reflect on the reading and watching goals I had set for myself at the beginning of the year.

For films, my target for the year was 250. I am currently at 128, which means I am slightly ahead of where I need to be.

For books, my target was 80. Which means I should have been at 40 by now. Instead, I am only at 28.

This is funny to me because at the beginning of the year I had confidently assumed that 80 books would be the easier target, and that 250 films would be the more difficult one. It has turned out to be exactly the other way around.

The reason that I am ahead of the film target is mostly the film club. Curating for it has meant watching more films than I otherwise would have. It has also meant finally ticking several things off my watchlist.

Reading, on the other hand, suffered a bit in March and April when I was moving houses and hosting people for dinner all the time.

The extra film watching has been both fun and a bit like homework. I have watched things I would probably never have chosen on my own, and I am glad for that. One Cut of the Dead is one example. I have watched it three times already this year. It is so much fun, so clever, and one of those films that makes you fall in love with films all over again.

But there have also been films that I had to drag myself through. Holy Motors, Cecil B. Demented, Angst are a few that come to mind.