things

s21e24 · 21 August 2026

What if they’re all right?; I didn’t make this; I write because

This issue explores how AI tools are reshaping work and professional roles while raising questions about what we actually lose when entire job categories are eliminated. The author argues that designers, developers, and product managers each bring irreplaceable individual perspective and taste to their work—and that AI's apparent ability to make any single role "redundant" misses the point that these roles represent fundamentally different ways of thinking and problem-solving, not interchangeable functions. Alongside this, the author documents experiments in creating animated text visualizations of social media posts and reflects on the distinction between "making" something versus thoughtfully defining what should be made—suggesting that the real skill may lie in articulate vision rather than execution.

1,631 words · 8 min read · 11 links · canonical link · JSON

0.0 Context Setting

Friday, 21 August 2026 in Portland Oregon where the timer continues to tick down to Skynet Sentience day. A busy day today: the 9am show, and then a couple calls, along with making sure that I get some proposals and overviews out to people.

0.1 Events

That’s a wrap on How People Work, Live! ... with Marcin Wichary. You can go watch the recording now, or you can look at the playlist, or even (deep breath) subscribe to my channel.

Next up!

Margo Stern, the author of Good Work: Managing Yourself and Managing Managing, a two-sided book that shares all the tips, tricks, hacks and cheat codes to working in tech. Drawing not only on her experience managing and being an IC at places like Google, Twitter, and Lyft, she shares the wisdom from interviews with PMs, Designers and Engineers who have been there too. Margo is also a Staff Content Designer at Lyft, and she lives in San Francisco, CA.

... and continuing the weekly clinic, you can register for How People Work... Does Happy Hour! this Tuesday 25 August at 1pm Pacific.

1.0 Some Things That Caught My Attention

1.1 What if they’re all right?

Stop me if you’ve heard this one before. A designer, a developer, and a product manager meet up after work and go to a bar after they’ve expended their token budget for the day.

“I got so much done!” says the designer, “I don’t even need a developer or a product manager anymore.”

The developer goes next:

“I also got so much done! I don’t know why I need to work with a designer or a product manager anymore.”

Then the product manager:

“I too got so much done! It’s fantastic, I don’t have to talk to any designers or developers anymore.”

The next day, they go back to their respective companies and find out they’ve each been made redundant.

At the designer’s company, all the designers and product managers have been fired, leaving just developers.

At the developer’s company, all the developers and designers have been fired, leaving just product managers.

And at the product manager’s company, all the product managers and designers have been fired, leaving just developers.

Meanwhile, the stock price of each of their companies does something non-deterministic in response to leadership trying to wiggle the “do something” dial with AI. Nobody really knows what’s going on, but at least the line is going up for now.

This might seem simplistic. I think the observation goes to not really knowing what it is that people -- individual people are doing in their jobs, what they see their goals and objectives are, and so on. And also a failure to recognize that people are unique! Job descriptions, siloization etc are all results of industrialization and attempts to manage production. The interesting deal with “AI” to the extent that it can be used by individuals to make software is that the tools are and can be very personal and are themselves aimed at producing software.

For example, can you tell me what that designer was doing in their use of agentic tooling to create “software” or some kind of artifact? How they went about it? The gestalt of knowledge and experience that they worked with in order to instruct a sort of non-conscious familiar? Do you know what that designer considered important and spent more time on, decided to focus on? And then similarly, for the developer and the product manager. Which of them was also a generalist, which of them had slightly more domain knowledge, which of them had the taste? Which is why I think eliminating entire categories of people, of jobs, is stupendously short-sighted (if not cruel, etc) and, well, stupid.

So. Maybe they were all right, in their individual sense. Think of the tool as malleable -- you’ll have to go with me, I’m afraid, in accepting that the tool is useful in the first place, and we agree to stipulate that it meets someone’s definition of valuable, and that there are always degrees of externalities. But right now, the wrong lessons, amongst others, are being drawn and generalized in terms of “do we need x anymore”.

1.2 I didn’t make this

I don’t have an AI policy, but Hank Green appears to have smashed open the Overton window in having one. So I guess I’m writing one and I’ll publish it soon. Feels a bit like a “soon everyone is going to have a code of conduct” moment. The concept -- a personal AI policy, mind -- has come up at least twice with different people in recent conversations.

I didn’t make a thing earlier in the week. What I did was a couple of experiments in form. They’re videos, and for some reason I philosophically do not embed videos, so there’s links:

Anyway. I had a hunch here that it would be interesting to see what it would look like to render Bluesky text posts (not reposts, not replies, not links or other media) as animated text in an interface that’s a bit like Reels or TikTok. There’s a whole bunch going on here in terms of “people don’t read” but at the same time, people do read captioned videos on vertical video platforms. What’s going on there? What is the form like?

So at around 3:30am I had this idea and sat down and wrote a product requirements document. I like to think that I’m thoughtful and good at describing what I want and what it’s for, so these documents start with a goal. In this case, it was explicit that this was a prototype to demonstrate/validate an idea about a social media form. I said what the product shouldn’t do and what it should do, and thought about what was important.

I will not -- I don’t -- say that I “made” whatever came out. I wrote down a spec and handed it to something, I don’t have any pretence of making software, but what I absolutely am comfortable with is that I came up with a good product definition that could equally have been handled to a developer or a designer or a team. That’s what I’m good at, and I think that involved knowing what was important, what to focus on, and what to leave behind. It didn’t need to be production ready, it needed to do a certain job of validation with live data.

1.3 I write because

A dear friend wrote to me asking broadly about how I write. There are clear stretches of time when I don’t write this newsletter, or I will admit, more accurately that I am unable to write this newsletter. But they noticed that there also appeared to be clear stretches where I write anyway

like, it’s clear that sometimes you’re writing almost because some part of you is reluctant to. How? Why? What do you say to yourself?

So I replied, not just because this was a dear friend, but because it felt pretty important. This is what I said:

To my chagrin, yes, there are long stretches of time where I do not write.

So how do I do it when I don’t want to? There are a number of ways.

  1. I start writing during a session with my ADHD coach (Mondays, one hour, totally worth it). This does not always work: I start them during the session but as soon as the session is over, I can run out of steam and never finish.
  2. I literally force myself to, and it’s not fun. Literally here means “I really must” and it feels like torture.
  3. I do a sort of ADHD pomodor technique where I use a sand timer for actually FIFTEEN SECONDS because I have been told and have experience that invariably, if I can start, I can probably keep going. (Finishing, though, is another matter if I get interrupted or there is a reason to stop). (There is one particularly famous (to me) writer who has written multiple books to whom I look up to who I think told me that they also have a pomodoro timer with a resolution of under one minute. That’s inspiring.)
  4. Another one is to remember what my wife told me once:

You don’t write to feel better, you feel better because you write.

  1. I don’t really have a list of prompts as such, and I am very bad at writing things down to remember to write about them. Often just the act of jotting down a thing to write about short-circuits the desire to write and makes me feel like I’ve already done it.

The thing is, I don’t write when I’m really really fucking depressed. I did have a practice when I was more carefree and younger. I am older now, with more responsibilities.

But lastly, it happens when I take away the pressure and the responsibilities and I just write and I don’t care. I really, really, don’t care what anyone else thinks, I don’t care if anyone thinks it’s useful or valuable because I remind myself I am only writing for me and it is an accident that you all can read it. I am honestly at times baffled that you are reading it, to be honest. Sometimes taking the pressure away means taking some ativan.

But, my advice is if you’re finding it hard to write when you used to find it easy? You might be burnt out.

How are you doing? Thank you for everyone sending notes, I love getting notes, and I always reply to them.

Best,

Dan