72: The Worst Technology Rollout in History (Ft. Paul Ford)

Show Notes

Journalist, technologist, and Aboard co-founder Paul Ford joins Peter and Jesse with the perspective of someone running a services firm in the middle of being remade by AI. The conversation covers the collapsing cost of software, blurring roles, what machines do well and badly, what design’s value proposition becomes, and Paul’s practical advice for staying upright through indefinite change.
Paul’s company: https://aboard.com/; Paul’s LinkedIn: https://www.linkedin.com/in/ftrain/
Peter’s website: https://petermerholz.com/
Jesse’s website: https://jessejamesgarrett.com/

Transcript

Jesse: I’m Jesse James Garrett,

Peter: and I’m Peter Merholz.

Jesse: And we’re finding our way,

Peter: navigating the opportunities

Jesse: and challenges

Peter: of design and design leadership.

Jesse: On today’s show, journalist, technologist, and agency founder Paul Ford joins us to share his perspective from all of those perspectives on the tidal wave of change from AI. We’ll talk about shifting expectations, shifting identities, holding onto the human in the face of rapid change, and a whole lot more.

From the pages of Wired and The New York Times to your ears, here’s our conversation with Paul Ford.

Peter: Paul, thanks for joining us.

Paul: Oh my goodness. Hello friends. It’s good to see you.

Peter: It’s good to see you. I’ve been following your work on the internet for probably 20, 25 years.

Paul: This is it’s just ridiculous at this point. Like it just… the three of us, anyway, it’s great to see everyone. I don’t even know if we’ve even met half the time. Peter, I think maybe you and I have been in a room a few times, but

Peter: yeah.

Paul: It’s lovely to see everyone just getting our career started, getting stuff moving.

Peter: Career. I think we’re getting our career restarted.

And maybe that’s a place to begin. It appears you’ve been going through it the past few years, and I’m curious how you introduce yourself, how you talk about what it is you’re up to now.

Getting Our Careers Restarted

Paul: It’s changed from time to time, but essentially I am the president of a services company. It works like a services company. People come to us and they say, I want you to build my thing, just like the industry I’ve been in forever, except the nature of building a thing has changed really fundamentally.

You can spin up software very quickly and, I’m the president of Aboard. We are a AI, sort of, delivery org. We help you get your head wrapped around a lot of these changes. And we help you ship software very quickly. Sometimes with our custom tooling, sometimes just using Claude Code or whatever.

And so I come from that background. You’ve all been in and out of services, like I have throughout your career, and actually it’s, kind of relaxing for a minute because a lot of times I’m talking as someone who is also a former journalist and I have to be on and know everything.

In this case, you understand where I’m coming from. I don’t know what the hell’s going on. Like I used to charge a certain amount of money. I used to, I used to be like, Hey, I’m gonna do this and I’m gonna get a product manager. I’m gonna get a designer. Maybe a junior designer too, and a couple of engineers. One for back end and one for front end. And then I’m gonna give you a thing and it’s gonna break, and then we’re gonna figure out how it needs to work. And then about six months later, you’re gonna gimme a lot of money, like hundreds of thousands of dollars, and you will have your thing and it will be good.

And that was 20 years of my life. And that is not the case anymore. And I think you guys are probably dealing with this just in the exact same way. I’m curious what, where you’re at.

Peter: Before we answer that, I have one specific question. No. ’cause you had an agency, you had Postlight, Jesse and I also had an agency in ye olden times

Paul: What was it? No, you gotta, you gotta name it’s for the people.

Jesse: Name it for the people,

Peter: Oh, they know Adaptive Path.

Paul: There we go.

From Product Company to Agency (Again)

Peter: And the sense I got, and correct me if I’m wrong, ’cause I’m witnessing this from afar, you left Postlight, you started Aboard, which I got the sense was started as a product company, you had an idea for a thing, you were gonna build a piece of software to put into the world.

And that was my operating assumption until literally three days ago when I went to your website and I was like, wait a minute. They’re an agency. Why? What happened? And I’m curious what that story is, ’cause based on what you were just saying, I think it’s probably illustrative of something going on right now.

Paul: Yeah, no, fair enough. So I was the CEO with my co-founder, Rich Ziade, who was the president of Postlight, which was a product focused software development firm in New York City, very New York based. We did a lot of work for big banks. We did a lot of work for the MTA.

And so everything that you would expect about boutique and it scaled, it grew pretty quickly. And it got to about a hundred plus employees many in the US, a few globally. Especially in Beirut, which is where Rich is from. He’s from Lebanon.

And all those things selling all day long, so on and so forth. We, during the pandemic, stepped back, new leadership took over and it felt like time to move on.

The company was purchased by NTT Data Americas and became part of a big product group. And I was contractually obligated not to sell services because my power as a salesperson is so incredible.

Jesse: Yeah.

Paul: Just, Just, if you’re not careful, I will sell you a six month product engagement while you’re not even looking.

Peter: Alec Baldwin has nothing on you.

Paul: No, exactly. I’m just ready. I’m ready at all times. I love to close. A, B, C.

Jesse: A, B, C.

Paul: Exactly. But Richard and I really like working together and we didn’t go with the acquisition.

It was one of those where they were just like, cool, we’ll take it from here. And so what we sat down to do was to build, and it’s now been through many iterations, a kinda white label….

We were like, let’s take the advice everybody always gives you. Let’s build a version of what we used to do at the agency that we could license. So every project you sit down and you build Kanban and you do user authentication and integration with Microsoft a lot of just the basic data and scaffolding stuff.

And so we built that. That was version one.

And then we were thinking, there’s just a series of decisions that you make when you’re building something, some of which are good and some of which are bad. So we made a version that was a little more open to the world where people could manage data, but the theme has always been from day one that we would take this to organizations, and organizations would use it to accelerate software development.

And so that felt good. And we had leads and we had clients customers really, because you’re selling a product and they were using it internally and all those things were happening.

And then ChatGPT showed up and started writing code. And I was like, whoa, this is gonna change everything.

And we all kind of agreed. And so then the last couple years have been about iterating in reaction to the fact that the cost of software is going down. And the need for acceleration, even as a concept, is becoming commoditized because you can just deliver stuff all day long.

And so a lot of the things that I thought I was way in the future on, I’m not anymore. Nobody is.

Because what the hell is a Kanban board? Or like sales or a CRM if you can build a credible one in two hours?

And so we’ve been riding that wave and what’s happened is that you react to your inbound. And our inbound has been orgs saying, help us get organized around this. That’s great that you have some custom things that you’ve built that help us build. We love that. That’s really nice. Good for you. But we’re just trying to get this going here. Can you help us solve this? Can you help us integrate these three things? And by the way, could we also make sense of AI while we’re doing it?

So we’ve backed into kind of consulting plus custom tooling as opposed to custom tool with consulting around it.

Jesse: I think that one of the challenges that’s going on for enterprises right now is that they don’t know what they do know, and they don’t know what they don’t know.

Paul: There is a great Reddit thread now, where I think it’s in just like the Claude Reddit where an engineer is just like, CEOs keep giving us code in engineering, and I’m losing my mind.

right?

Jesse: terrible. It’s like the worst case scenario. And so it is this interesting sort of tension that teams find themselves in where the walls are collapsing. And the gap between thought and reality is shorter than ever.

Paul: Which we should be excited by, given our careers.

Jesse: We should be. Yeah. exactly.

And then I wonder about these teams, your clients, our clients, the people who are inside these organizations who are trying to get things done inside these much slower moving cultural environments,

Patrick Swayze in Ghost

Paul: I gotta tell you, I don’t know if you have this experience, but you’re dabblers, we’re all dabblers because we were early in this technology and that was what kind of got us into it.

And I just wrote a post about this for our newsletter. I feel like Patrick Swayze in Ghost: I’m telling everybody what’s happening. Like, You’re in trouble. There’s danger. And everybody’s like, wow, it’s cold in here.

I keep having that experience and what I realized is that what has happened is so weird. It’s so weird. And we’re used to weird things, but most of the world is organized….

i don’t know if you guys had this experience, I’m gonna be another web old timer, but like, there’s a lovely program. I taught it for a while and it was with people like Jeffrey Zeldman at the School of Visual Arts, the MFA in Interaction Design.

And there was a point where I went in, and it’s like me and Jeffrey and Karen McGrane and I’m like, oh, we’re the older weirdos here. We’re, like, the early ones because there were these just very put together graduate students who were like, I’m gonna learn interaction design, I’m gonna build a really good career.

And then they would go to FAANG companies like right away. They were good product people. And there’s that moment when you realize oh, this is becoming professionalized. That’s really good. That’s actually healthy. People are seeing this as a stable thing, as opposed to a wild wave to ride and figure out.

And I feel that that gulf has never been more profound in my life. Like, it’s very hard to see something so transformative. And then you take it to people and they’re like, yeah, but what about those Mets?

Like they’re just kind of like, they just would like, please, can we not discuss this anymore?

Peter: in my experience, it’s one of two things. Either, this is the only thing people want to discuss or this is the thing everyone’s trying to avoid. It’s hard to have just a casual conversation that might touch on it and then discuss other things.

Jesse: It’s got a gravity well, right.

Paul: Yes.

Peter: It is the gravity well. It reminds me Paul, I don’t know if you know Matt Jones, from the UK, one of the people behind BERG. I remember his frustration in 2004 or five, where all anyone wanted to talk about was the web. And he is like, there’s more to the world than the web.

And he was right. And I’m sure there’s more to the world than AI, but right now, like this is the actually intellectually interesting thing as well.

There’s something here that’s forcing us to reconsider stuff that we felt was maybe assumed. You outlined it.

And that’s been a theme that Jesse and I have been pursuing how for 20 years, from about 2000 to 2020, 2022, there was just this nice, steady kind of pace of growth and development and reliability.

Jesse: It felt like we were building a stable new thing

Paul: That’s right.

Jesse: For about 20 years, right. And then suddenly it wasn’t stable at all anymore.

The Tinkerers and the Professionals

Peter: Right and something’s happened. Even pre-AI with the layoffs and stuff in 23. But it’s definitely kind of timed to the AI moment.

You were talking about professionalism and you were talking about how we started as tinkerers. And I wrote on the whiteboard behind me the word humanities because the three of us went to school at around the same time, and all studied, I studied anthropology. Jesse was a journalist, and you studied English, the famous degree of futility.

But with our generic humanities backgrounds and not being afraid of technology, we were able strike out on these emergent technology of the web and the internet.

And now there’s a group of folks who… It got professionalized at places like SVA or bootcamps and other things in the last 15 or so years. And the thing I’m seeing, and I’m wondering what you’re witnessing, Paul, is that folks of our vintage are in some ways less rattled by what’s happening now because we came up in a discombobulated moment and in fact many of them are excited because it’s like, Hey, this feels like the late nineties again.

Whereas folks who arrived when it professionalized are the ones who are maybe the most off kilter because their assumptions are now shattered.

Paul: I think that there’s a few ways to hit at that.

I’ll take two angles. So angle one is, there was a great quote. They asked Linus Torvalds, who created Linux, who’s a little bit older than us, but not much. And they’re like, what do you think about AI? He is like, oh, it’s wild. All the vibe coding, it’s really interesting and it’s gonna really, it’s a real accelerator. Like he’s just vaguely into it as a technologist. He’s not shocked, he’s not an open source purist about it. He is just oh, there’s gonna be vibe coding. It is what it is.

But he is, like, but let’s be clear, nothing was as transformative as compilers, which was a really funny thing to say.

But then when you think about it, until compilers, you had to literally put in the code that the CPU runs in like hexadecimal numbers and, like, you were flipping switches and then one day you could issue a list of English-like instructions.

And that’s like at a thousand times cognitive speed up. Like, so many people, we just take it for granted. We take it for granted that you can have something that looks like Python or JavaScript and that will run, or even HTML and it will do something on the computer and you can learn that and you can teach it and so on.

And that, that actually wasn’t there for a really long time in the history of computing for all sorts of reasons. And so I really liked that framing because I was like this is another transformative thing. It is accelerated, it is different.

But we’ve been through major transformations before. That said, what’s different is it was not a, multi-trillion dollar, 30-million-person industry as it is today. And so the change is being felt really radically.

And so, we are very interdisciplinary by nature. It’s what attracted us to the medium because it was the first time that you could smosh the humanities degree with the technology, with the visual, with the written.

Like, It all started to congeal and you could kind of mess with all of it. And nobody could tell you no, you know, nobody could be like, you’re not allowed to do that. ‘Cause there’s the web, you can do whatever the hell you wanted. And so I think that is always very attractive to us.

But most people really want a little stability in their lives, like they really do. And we’re the freaks here, right? Like, we’re like, ah, well more chaos. Let me just huff it up.

Boy, that’s some good disaster. Get me more.

Peter: I dunno, I think I’ve reached a certain age where I’m fine with stability.

Paul: This is the thing…

Peter: Maybe when I was younger, chaos I could thrive on.

Paul: But think about actually, Adaptive Path is a good example. It was a transitional moment where you guys went, okay, wait a minute. We actually have to make this boring. And I don’t mean that in a bad way ’cause I’ve done it many, many times.

Like, I need to go to a bank like the one that eventually acquires you and I need to help them make sense of this because their world is interacting with this, but their systems can’t tolerate so much change.

And that’s you as little babies, you’re little babies going okay, we gotta figure this out because rhis change is so difficult for relatively nice people who I have meetings with. So we gotta actually systematize and structure this. We’re gonna make lots of diagrams with boxes in them, and it’s gonna be okay.

And we’re just not there yet with this stuff because every fricking week Claude is oh, by the way, it can make shoes now.

Jesse: You’re right. Yeah.

Paul: What would Adaptive Path be for the vibe coding era? Jesse?

Jesse: I, well, yeah. Okay. What would it, all right. Hang on. It’s such an interesting question because I think that a lot of what we did was about deliberately diving into unexplored territory. We were continually inventing methods and developing techniques around things that we had no idea what to do because no one had ever done them.

And so we had to improvise and adapt from a methodological perspective around the work, and what was actually to be delivered.

So we just recently celebrated the 25th anniversary of the company. And one of the things that I pointed out to folks is we did not even call this work design when we started doing it because we felt like we were after something bigger than what people thought design was.

And in this moment, in this time, people think they know what design is.

You Sold Empathy for Large Systems

Paul: No, I mean, let me make an observation about what I think you actually sold. I think you sold empathy. You said you had a lot of skills, but I think you actually sold empathy for large systems. It was like we’re a big system and we don’t really know what to do next. And you were like, I have the skills that you need.

But more than that, ’cause that’s tactical, I actually have empathy for the fact that you’re doing the thing you’re doing and I would like to understand it. I’m curious about it, and I would like to bring my skills to it. And I think that was very comforting to people with the change, right?

And so when I think about the organization I’m building today. Everyone is a solution engineer. That’s like a mix of product manager, design, engineer, ’cause if you can ship anything, right? So let’s call everyone a solution engineer.

But our job is to be empathetic to people who are driving through this change who want to be able to access the system without destroying everything on day one. And I think that’s what you guys were too. I think it like that’s what everybody’s really hungry for as far as I can tell. They don’t wanna just build, they want to understand.

Can the New Computer Clean Up the Old Computer’s Mess?

Peter: I’ve got two questions about your business. The first is, is there a type of client or type of project that is pretty typical that is coming to you and what are those?

Paul: So there’s what’s in-house and then there’s what’s inbound, right? And I can describe both in broad terms.

So in-house is like helping a big insurance company doing things to help them manage policies in new ways and analyze data.

So it’s a portfolio of, like, projects, like classic projects with a little bit of AI and a little bit of acceleration and a little bit of strategy and kind of putting that together.

The larger orgs are like that. They’re like, let’s just try to get this in. Let’s see what we can get. We like the advantages of it, but we’re gonna take it step by step. Security is very important. Auditing is very important.

The other in-house is, it’s a giant medical informatics play from a big not-for-profit. They want to aggregate lots of data around children’s health, and they need to do it in a HIPAA compliant way that’s very secure and safe. And then they want to be able to almost talk to the dashboards, right? Like they wanna be able to ask natural language questions about how services are being delivered across a large number of clinics. And so, that’s another one.

Those are kind of like almost classic big project engagements that are enabled and interesting because they’re AI. You know, there’s a team on them and so on.

What I’m finding more and more is that if there’s a theme, it’s that, wow, computers made a mess of everything, and if we don’t have to live with it anymore, that would be really awesome. Can the new computer clean up the old computer’s mess? And the answer is yes.

Like the huge pile of documents that was OCR’d can now be really rendered way more useful. The invoice system can be better.

The number one question that everyone asks on Earth, as far as I can tell, including babies, is what would it take to get me off of Salesforce?

That is like the absolute, like, literally people who are dying, their last words are, I wish I could have gotten off of Salesforce.

That’s, please don’t let me go…

Peter: They don’t want to have spent more time with their family. It was just off Salesforce.

just get me off Salesforce.

Paul: Why couldn’t we ever get off Salesforce?

I hear a lot from people who either tried to build up from SaaS products into a kind of a working system and then just find endless friction every day. I hear a lot from people with legacy systems who would just like to move a little bit faster. And then I hear a lot from people who are just like, I gotta get this into the org, but I don’t know how. And so those are the shapes.

And then at the same time, a lot of the things I’m seeing and living, delivering the services are really confusing because the design phase is different and the engineering phase is different.

Building the Product During the Pitch

Paul: And sometimes, and I have had experiences, a lot of times I’m building the product during the pitch, right? Because it’s just so fast.

And so then do I show that to them? When do I show them the value that they can have access to? Not because I want to get them… but literally ’cause it’s so overwhelming.

I don’t know if you’ve had this, but you create these artifacts and people don’t write back for two weeks because

Jesse: it’s

Paul: so hard to metabolize like what they just saw.

Jesse: This gets back to a very old tension between Peter and myself in our consulting practice, where I was always reaching for the farthest frontier that the client could possibly see, and then complaining that they didn’t sign on to my ideas.

And Peter’s like, well just, you know, don’t give them the 10 mile view. Give them the one mile view and get ’em to sign on to that.

Peter: Yeah. Give them the thing they can actually do.

Paul: No. We’re gonna just, I just want to pick at this scab. This is exciting.

I think, look, I’m with both of you, right? I like to see where the future’s going. And actually if there’s anything, my dynamic with my co-founder, I’m probably more the Jesse and my co-founder is more the Peter in this, but it’s always that collaboration, right?

Because I’m the guy who’s wow, whoa. I literally, I hit Claude Code in November when it got really good and I walked in the next day and I just turned to Rich and I went, death is coming.

That, those were my words. And he is like, slow down. The insurance industry doesn’t move that fast. And actually the insurance industry, if death is coming, they’re like, cool. We priced that in. It’s all great.

Jesse: Yeah.

Paul: So no, I think, it’s very tricky. Organizations cannot metabolize change.

I will say, I think there is a thing coming for all of us that no one has fully factored in yet, which is the security profile of the entire industry has to change because it’s much easier to hack systems than it used to be.

So I think if you were to really think about the right way to walk this into an org and just be like, let’s do some automated auditing and figure out where some of this stuff could be like made a little more firm and stable using AI, like let’s clean up some mess. Let’s think about security. Let’s audit your open source. Let’s make a good auditing trail.

Like, there’s a lot of system stuff that’s gonna have to happen because you can hack all day now at a nation-state level for $200 a month.

And so those changes are coming and they have really high costs. And I think probably we do have to react to them as an industry way faster than we were even expecting.

Or not. Or not maybe, hey, things work out. Security is one of those things.

Jesse: be fine.

What Is Design’s Value Proposition Now?

Jesse: You touched earlier on the blurring of lines as you’re defining this role of solution engineer for yourself within your organization, and for a lot of the folks that we hear from who listen to this podcast, they are in-house design leaders.

They have got teams of 20, 50, a hundred, 500 people who are dedicated to design, whatever that means, as a value prop inside their organizations.

And now they’re looking at engineers doing what they used to call design and product managers doing what they used to call design. And their designers are looking for purchase within this landscape. They’re looking for a place where they can get some really meaningful traction. And so it comes back around to this question of what even is the value proposition of design as a function anymore in this blurrier environment? What do these leaders and these teams have to look forward to in terms of what they bring to the table and what they can potentially offer?

Paul: I think this one is thorny and it’s not thorny in a like, oh my God, time to start that bakery kind of way. I think it’s more complicated than that.

Let’s take this step by step. I’ve been doing a lot of thinking about this, I think there’s a few realities. One is an average human being in a technical context, an engineer or product manager can do far, many more design tasks without asking a designer than they used to be able to.

And I think one of the ways we get in really big trouble is we go, well, that’s not really design. But if it has a function for the audience and the audience doesn’t really feel that they’re being duped, then you’re at a loss now.

It might be disrespectful of the discipline, it might be not at a level of quality that you would ever put out for yourself, but the reality is you are at a loss because the audience has gone, oh, cool, that’s there. And it’s there in 10 minutes. So that’s, I like that. It feels good. It looks like an app to me.

And then there’s the secondary thing, which is that some of the models are good at certain kinds of design. I’ve absolutely had moments where I was like, that would’ve been a junior designer. I’ve never seen anything that come out where I’m like, that’s a senior designer level artifact. But, absolutely things that are, like, wireframe-plus or relatively light, type heavy, here’s how the onboarding would work, right? I, I’ve seen good product.

So I think you have to get close to that and you have to know it, and I think you have to know how to steer it and understand it so that you’re delivering it with the information that you have in your discipline as opposed to getting frustrated with other people for doing it.

You have to enable them. Now if you’re in a hostile environment, there’s a lot of bad blood in our industry, right? There’s a lot of like, oh, design always takes forever. And then like product is snooty, but also can’t do anything. And engineer….

And the reality is everybody thinks they can do everybody else’s job. And then it always blows up. And this has been going on way before AI, so product is so excited that they don’t need engineers, ’cause engineers are exhausting.

But I keep seeing product managers ship the same fricking thing that just looks like a spackled together database that they’re like, I’m vibe coding, aren’t I amazing?

And I’m like, yeah, but honestly that’s like, you could have done that in Airtable. Like, I get it, I get that. It’s cool, but it’s not really a product even though you’re the product manager. Like you forgot that part. Everybody keeps forgetting their actual thing, right?

But I gotta say our design team, we’re relaunching our marketing website because of course the company changes. And we were trying to use… Figma has this product that lets you actually publish straight to the web. Now I would not recommend that but we’re doing a lot of Claude Code based automatic translation, like out of Figma straight to the site.

It is good, it’s a little too much tailwind for my liking. Like it’s a little too garbled, but it’s good, it’s credible, it’s a first pass. We’re gonna need to turn it into a proper CMS template and that’s gonna require human intervention and engineering.

But those are designers. And at one point I said to one of our designers who’s like, look, I really wanna learn this stuff. I was like, okay, I’m only gonna talk to you for 10 minutes. I won’t give you any more advice. You have to go do this and then report back. Go create a component system to solve, I might have said video or like some sort of problem related to one of our clients. And I was like, make a component system and then I want you to ship it as a React storybook, complete, and figure out a way to host it.

And two hours later he came back. He’s like, wow, I, that’s okay. I can do that. It wasn’t the greatest React storybook, it was missing, there weren’t a lot of animations. There weren’t a lot of variations, but it was pretty good product, right? It was okay.

And so designers are not out in the cold, except in that way that designers always put themselves in the cold.

You’re like, I’m a small bean, and I cannot think technical thoughts, only art thoughts. And like we all have to get over that.

Peter: So I wanna unpack a little bit of what you were just saying.

You mentioned earlier how everybody in your organization’s a solution engineer, but when you were just telling that story, you were referring to designers. So you still have a mental model of people maybe in the roles that they had been in.

Paul: Oh yeah. Let’s not, yes, I’m the president of the company, but I’m also a solution engineer who uses the tools that we’ve created that are sitting in our GitHub repository and lots of Claude Code in order to prototype.

However, I really don’t push things to production in the org. I’m mostly here to explain where I think things are going work with my partner, et cetera.

Our engineers are absolutely able to think in architectural ways that our designers probably won’t, and our designers are able to think in terms of visual systems. What I’m finding, and I think we’ll see it, the novelty of it is so wild, right? Where it’s like, Hey, use giant component system X.

If you tell it, if you say, Claude, man, I just want you to use IBM Carbon, good system, go to town. It will make you something that looks very component-y and good and solid, and then it all starts to look the same. I don’t know, is design in the future just skinning component systems? I don’t know.

An enormous amount of our work still ends up going through classic design processes. But if a designer can solve a problem by typing words into a box, they should be able to do that. They shouldn’t necessarily have to wait. It should be auditable, somebody, a product manager.

So there’s different kinds of ownership, right? Like a designer should be able to do stuff if we’re really just changing a component and I can see a future state where that could ship to production without a lot of other people involved. We’d be willing to take that risk.

Someone whose background is design probably shouldn’t establish an authentication system, and push that to production.

But they could say, it’d be cool if login worked like this. And I actually prefer email based login. That’s better for this than trying to do OAuth. And somebody might be like, yeah, that makes sense. And then everybody types like three sentences into the box. Somebody goes, Hey, it looks pretty good.

Jesse: Yeah.

Peter: It’s a canard now around roles blurring. And, your adoption of solution engineers struck me as a recognition that whatever we might have been once, we’re all doing something similar.

Paul: Now that we have returned to our services roots, what is a solution engineer? It’s somebody you can talk to who has an understanding of technology and platform who can help you figure out. What the next step is, and often that is before a contract is signed. And so a big part of this is just when an organization comes in or a person comes in and they want some help, you’re like, all right, talk to a solution engineer.

A solution engineer might be coming out of design, product management, engineering, front and backend might be an organizational leader or just a pure manager, but they’re able to do something. They can make you an artifact that looks and feels like code, and so I think that it’s working backwards from that change as opposed to like this very prescriptive role.

It’s just like when someone walks in, I want to introduce ’em to someone who can solve their problem. Hence that title.

Peter: And this touches on something I’ve been thinking about. So I do a lot of thinking on organization design, and I’m wondering how you think about shaping teams as roles blur. What is your organizing principle to say this collection of people, anywhere from two to five or however many, are the right collection of people to solve this client problem that’s coming in? Like how does that calculus operate?

Paul: It’s still pretty classic, right? First of all, I don’t believe, in general, people should not go alone in any client engagement. It’s dangerous.

Jesse: It is dangerous to go alone.

Paul: it is, yes, exactly. Take this and I hand you, Claude Code.

Like I don’t like to sell alone. I don’t like alone in a services org because you should always have a couple people thinking thoughts. But when I think of currently org structure, what I’m finding is that product manager as solution engineer kind of depends on the profile.

If they’re more technical, they can really get through to the very advanced demo, they have access to, like a good platform, they know how to get things into production. They can do an awful lot.

Validators, Not Producers

Paul: It’s really almost, and I feel sad about this. I’m not really ready to commit to it, but I feel that the roles become more about validation than about production, right? So it’s like a product manager validates that this is a real product. It’s not just like pixels slapped on top of a database, but that it really actually understands the flow of the business or the organization and their needs.

A designer thinks about experience and they think about components and they think, honestly, a lot of times designers are thinking about scalability of a design experience. They are willing to use all the tools necessary to make sure that people can have consistent, reliable experiences when they come and sit down. That is a validation pass, if you are not the one who is moving the pixels around.

Now they might mock something up in Figma, but then a robot might build that rather than sending it to an engineer. But an engineer should review that robot’s output to make sure that it is valid, good, stable output. And I’m not imagining a world actually where all we do is say the robot did a good job. Because I think there’s lots of stuff that needs to come from people, and I think there’s a lot of stuff we need to do.

But that said. I’ll give you an example. I have an automatically generated newsletter every day, but I built a lot of tooling around it. It reads five or 600 sources. It keeps sources connected to summaries and so on. I never share the output of it. It’s just there for me and a couple people, and it’s my AI coordination newsletter, and I use it to know what to post on LinkedIn and so on.

I have to go read the articles, I have to do lots of stuff. It extracts numbers, so on and so forth. I am the validation before any of that content can reach an audience on LinkedIn or in a newsletter. It has to be there because otherwise it’s really meaningless. The Pure newsletter experience, I think is interesting, but unless somebody’s using it that way, it’s not that great of a product.

It’s a very good product to feed what I’m doing, but it’s not a good product unto itself.

Jesse: So it’s interesting to think about what all of this means for the leaders of design teams inside organizations as they are trying to reshape, resculpt their teams and honestly their value propositions, right?

Paul: What are you guys hearing? Like you know, more designers than I do.

Jesse: Yeah. So in the conversations I’m having with design leaders all the time, they’re terrified.

Paul: Yeah, that’s fair.

Jesse: The thing about it is that what the machine is best at is mediocrity. Which is the thing that designers have been fighting against this entire time, right? Like designers really see themselves as holding the levees against the incoming flood of mediocrity that executives are so desperate to unleash upon customers.

And then as you point out, then the question becomes what are you bringing to the process anymore? It’s not execution against base level design standards, best practices, value propositions. It’s gotta be an elevation. Right?

Paul: Yeah, it’s not like you’re like, Hey, here’s the branding guide. Like the fantasy now is I gave you the branding guide and the robot will do all the design for you based on the branding guide. It is, it’s very scary.

Let me describe a future state, but I, here’s what I’m gonna describe a future state, but I don’t really have a way to get us to the future state.

So right now you have these big

Peter: So you’re an underpants gnome.

Paul: Yeah, exactly right. No I, the problem is I think that the bullet point two or three, I can’t remember what the underpants …

Peter: Steal underpants. Two, question mark. Three: profit.

Paul: Yeah, bullet point two I think might be like a lot of chaos in reorganization.

We have a podcast at work. You can find it on our website, aboard.com. I just interviewed a really wonderful writer and thinker. His name is Andrew Leland, and he has a degenerative eye condition, so he is losing more vision every year. And so he’s become, even though he used to edit the Believer, he is become a really good…

Peter: Oh, I’ve read his work about that. Okay.

Paul: He’s wonderful. He’s wonderful and the reason I wanted to talk to him is he’s become a big sort of nerd, right? Because he is in text editing and using Claude Code and he started to talk about how blind users and low vision users are hacking their own tools and building their own systems.

You know, like Wikipedia is very accessible, but it’s not always really scannable to someone who doesn’t have vision. And so people are making ways to read and find Wikipedia for themselves, right?

So I want you to think about a future in which those needs, which could be the needs of a disabled person or the needs of a corporate leader, it’s like they have really specific things that it could actually help them do their jobs. But literally somebody who’s good with this stuff could sit with somebody who realizes, who’s starting to lose their vision, but has other particular needs or websites they really must visit for their job or things like that, and build a solution just for them.

And a designer would be incredibly useful there. They would be thinking about things that aren’t visual actually. And at the same time, I don’t wanna be too utopian here. You need to fix the real problems first.

But what really stuck with me is that, if you think about a future where everything is commoditizable and moves fast, and people can have very customized stuff, then individual needs can be addressed in ways that they couldn’t before and maybe they’ll scale out to other people.

But how do you get there from a 400 person team inside of like a giant electrical utility company? I don’t know if there’s an organic step-by-step path where everybody gets to keep doing what they’re doing.

The Cycles of Disaster

Jesse: It’s interesting to see where the personal meets the professional within this, because the web, as you pointed out, kind of made its way in enterprises because people inside realize the value of these tools and they advocated for them and they set up web servers under their desks and all of that stuff.

And in this age that you describe, which I imagine as an age of infinitely disposable software, that you can create something incredibly complex, incredibly bespoke, that would’ve cost you $500,000 for a team to build for you, that is just for you to plan your vacation and you’re gonna use it for just that purpose, and then you’re never gonna look at it again.

But the cost of it is almost nothing to actually create. So with that in mind, then, what you see inside enterprises is every team has its own tooling. Every lead designer has their own way of orchestrating and architecting the solutions that enable them to deliver what they deliver.

Paul: Yeah. So now we have to build unified systems for bringing all of those systems together. And then everybody’s gonna hate the unified system and go and vibe code their own solutions. Except it’ll be for Claude Opus 35. And it’ll do it just ’cause they whistle at it.

And that’ll generate a whole lot of data and be insecure, like the cycles of disaster that define our ridiculous industry are also getting accelerated.

I can tell you from history, right? Like I’m in New York, I’m very publishing adjacent. I worked on a lot of big content platforms and so what happens is everybody has one unified content platform for the entire magazine conglomerate, right?

And everybody hates it. And then they fire the CTO and the new CTO comes in and is it’s cool, you can use WordPress. And then five years goes by and they fire that CTO and the new one comes in and goes, how the hell can everybody be running their own WordPress? This is a disaster. We’re gonna bring everything, we’re gonna build our own custom CMS, we’re gonna bring everybody into it.

And it is a beautiful cycle. It’s like a, it’s like a, it’s almost mythic, right? And I think you’ll just get that, except it’ll be like literally on a four-month interval, which will be tough for the CTOs ’cause it’s hard to source that many.

But here we are.

Peter: Jesse mentioned design leaders being terrified. I think I’m seeing a broader swath in the conversations I’m having.

I have one design leader, I mentioned this person in one of my newsletters, who’s actually primarily frustrated because he is so far ahead of the rest of his own organization that he doesn’t know how to bring them along with using this tooling.

Paul: It is the ghost problem. He can’t, yeah, I get it.

Peter: Well, yeah. he’s seeing opportunities for this tooling to improve how the business operates, and he can’t get anyone else to care about it. They just are like, no, I’m just gonna keep doing what I’ve always been doing. So there’s frustration.

I’m seeing empowerment from some leaders. Jesse, you and I have been talking about Z from CloudFlare, and she’s an atypical design leader, but she’s a design operations leader at CloudFlare, and she’s been writing on her Substack, and I’ve had some conversations with her about it, about how she’s gotten the 30 some designers at CloudFlare outta Figma and just into an LLM that has the design system, and they’re just like making live prototypes that they hand over to engineering to be wired in and shipped.

And she’s like, this is awesome. We’re designing faster, better, closer. There’s fewer levels, fewer stages of interpretation where things can go wrong.

You hear this from some other design leaders where now they can have their own repos and clean up three years of design debt that was never gonna be prioritized by engineering. Engineering’s like, please have at it. Stop bugging us. We are more than happy to let you just solve those problems.

Paul: Yeah, right now, but like a year from now, they’ll be like hold on a minute. Wait a minute. I’ll take it. I’ll take it. It’s okay.

Peter: Well maybe.

And then, I just had an email from a head of design at a large bank, and he was somewhere in the middle, which is just like uncertainty. What do, he was asking me, what are you seeing? We don’t know what’s going on.

And so it wasn’t fear so much as just ah…? And then yes, there are the anxious and fearful folks.

And so I’m seeing a spectrum and I keep drawing normal curves because the normal curve is helping me categorize what I’m seeing, which is that, the farther you are along on your AI maturity, right? There’s folks who are here and they’re on the more confident end of the spectrum and they’re doing fine.

And then there’s people in the middle who are like, I don’t know what, and then there’s people on the kind of the other side. We just had Jorge Arango on, and he had this line: don’t think that this is a nothing burger. This is a burger, right?

He, we might not know what kind of burger it is, but it’s a burger.

But there’s folks back here are the, either the nothing burger, or, to overextend the metaphor, they’re looking at the menu and they’re just like, all the burgers are like about to fall upon them. They just can’t make sense of the variety of burgers that are out there.

And so there’s this spectrum and I don’t quite know what to make of it. Except that it feels like the folks who are feeling best are those who aren’t afraid. They’re approaching it with curiosity as opposed to anxiety.

Paul: I will tell you how I approach this.

So first of all, I think that, as a technology landing, this is the worst rollout in history, right? It never made sense. It shows up out of nowhere. The leaders of the companies are, they have many qualities, but they’re not great communicators about what’s coming. They don’t know the capabilities of their own platforms.

And it comes on a level of hype that is essentially trying to match Bitcoin hype, but with a weird database that talks to you.

Like, a lot of folk narratives have emerged, you know, it uses more water than anything. And then people come out and they’re like no, it doesn’t use that much water. But I’m like, let ’em have it. Let ’em just have the fear. Let ’em have the sense that it is ecologically bad. Let ’em have it. Don’t fight everybody tooth and nail on the fact that some of these things are very destructive and they’re very confusing and they’re not good for society and they’re messing stuff up.

So that’s a, just go in with empathy and know, and I felt this, I’ve been writing about it in public and my God, you take some heat when you say, I think this is real, because people are very committed to a narrative of the nothingburger.

Peter: Well, in particular, people with background similar to ours, but who are not as technical, who stayed in humanities practices and who have seen more directly, I think their livelihood threatened by these tools. And so they’ve just been like, they’ve noped all the way out.

Paul: That’s right. I gotta tell you though, I just had a class come in, it was a journalism class from the New School. It’s a very progressive school in New York City, and they came in, and I actually thought they were gonna be grad students.

They’re all undergrad and they were, they’re babies. They’re just lit–, they’re so young and they’re just full of ideas. And I didn’t have to say much. They just asked questions.

And at one point I was like, look, what do you guys do with this stuff? And almost universally, I would say, they’re all, yeah, average age is like 19, 20. And they were like, oh, I don’t use it. I’m in college for my brain. And it was just this very weird moment where you realize this has been there since high school for them. It’s very normal.

Some of them said they use it for work. One person was a vibe coder and so they were using it those ways, but their relationship to it, and I think what happens is, everything gets to such a fever pitch so quickly that we forget that humans are higher primates capable of reasoning and organizing their lives, even with new technologies.

And I’m looking at them and I’m like, ’cause you go online and it’s just like, it’s either people saying like, everyone who even thinks the letters AI should be shot and thrown in a river. And then other people who are like, the Lord came to me last night in the form of a prompt.

And you’re just sort of like, you’re caught between it and you just get this assumption, especially ’cause the news is so bad and like…

Peter: Yeah, this couldn’t have happened at a worse time, at least in the United States from a geopolitical standpoint.

Paul: And also from a climate change point of view, like the timing’s terrible and everyone who is in the news running the shop is an absolute reprobate right now.

And so you assume that that is the world. Most humans really just want to have a nice dinner with their friends on Friday and actually don’t want to talk to the robot all day.

And, but then the thing is also we are talking to the robot all day, I think we get very skewed. I think a lot of people don’t take this anywhere near as seriously as we do.

Now that said, you go back to that design team of 500 people. Is this existential? It could be, and we have to be honest about that.

I don’t know, but when I talk about a future state where everybody can get their thing, and that’s very utopian and positive, but if we talk about what would happen to go from step to step, it’s a lot of pain.

So I think there’s probably balance in here. I don’t think we get to this moment where we’re like, ah, that was fun, but it’s over now. Let’s get back. Let’s open up Aldus Freehand,

Jesse: Right.

Paul: and get back to work.

Jesse: I think that touches on an interesting thing in here. You and I both come out of a writing tradition, a journalistic tradition. And for me that writerly approach has always been a part of the way that I’ve approached design in terms of thinking of it, in terms of storytelling and so forth.

But what I’ve found in decades now working with other designers, is that most of them don’t think that way. Most of them are, they’re much more nonlinear, they’re much more visual, they’re much more abstract, honestly, in the way that they think about things. And they are now being asked to turn their ideas into words in a way that they haven’t had to before.

And I find myself wondering what that means for the skillset. So let’s leave aside the blurring of the roles between design, product and engineering, but just simply the skillset of describing what you intend to create in language. And what does that mean for the shape of teams, the way that people’s value is even judged, you know?

Paul: There’s a couple different ways to look at it though. This is a very assistive technology in this way. I have a friend who, he is a developer and he is, he’s a black guy.

He works for Microsoft. And he’s like, oh man, this really helps me. I actually can make my prose sound a lot more white with these tools. It’s really useful for me when I’m like, doing a brief. And I was like, oh yeah, that actually the logic of it was impeccable.

He’s a very good, very clear writer. He writes a newsletter. But I got it. He was just like, this is, it’s a safety net for him, right? He was just like, I’m in this corporate environment. It’s a way for me to be in charge. And I really respected that. I respected that use of it and it is often in my head.

And so I think we’re just kinda, everybody is, am I gonna be marooned because no one will help me? Am I going to have to figure all of this out myself?

And I think the way these technologies have landed, like I was just saying, right? Like they, came down from nowhere with no instruction manual and the sort of eager beaver types like us are like, okay, I know what to do.

There’s no big O’Reilly manual yet. I’m in paradise. But now build a system for them, right?

It’s really cheap to build a system now. So if we want to help designers, what are designers good at? They’re good at thinking in terms of visual or, but also just systems. Systems and ways that data interacts and usability and how to represent data and ideas in ways that are tractable and sometimes touchable.

And so what tooling could we be building for them? Given that everyone can build a lot more tooling than they used to be able to, to help them express themselves. Could we give them a nice interface with some guidance so that they could do more of this? Because the cost of that is really low.

And so there’s this assumption that’s driving me a little bit bananas. It’s not you, it’s me too. Like everybody’s caught in this, which is just, no, it didn’t pass the test.

So you’re going to the island and then you’re gonna have to fight for beans, right? It’s just everybody is, oh my God, yeah. I just need a can opener so I can survive. And it’s just yeah. You may not be in Figma all day, which I mean, God, I don’t even know what Figma is anymore. Have you looked at that thing lately? It’s like an operating system. It has a CMS built in. I don’t know.

But, we could be building tools for designers to explore and figure out where their talents are gonna go inside of these orgs. And I bet you’d see value in a week, ’cause it’s so fast to build the tools, but we’re not there yet.

Like I say that in a very abstract way. I’d love to work on that. I think it’s a really interesting puzzle, but it’s like nobody’s quite buying that yet. Like nobody at the big bank wants to build a custom design exploration LLM tool in order to maybe help designers contextualize themselves in the world of AI that’s like $200,000. They would really not want to spend, they’d rather buy a product or hire a vendor, right?

Unless you’re just going to do the Oracle thing and fire 20,000 people, which I don’t think is the right move at this time. You might want to think about that. You could build frameworks here and we can get ’em done real fast.

Throw ’em away if we don’t like ’em.

You Can’t Yell It Away

Peter: So, a separate thread that Jesse and I have been pursuing on our podcast is something we call Liminal, and it’s addressing… what we were talking about is now is this liminal moment, between a way the world worked that we all understood, and a new way the world is about to work that is very fuzzy and unclear, and that we’re caught in this.

And, something that I’m witnessing in your life, career, whatever, is that you’re navigating liminality quite actively, and I’m wondering what you found that helps you through that. Like how are you staying somewhat true and upright when you’re probably being buffeted by these winds of change constantly.

What’s worked for you?

Paul: Well, I mean, I think it’s very real, right? The way I describe Anthropic and OpenAI is, I’m looking right now in our office, we have a view and I can see the Hudson River, and I’m like, it’s just like two aircraft carriers just came up the Hudson one day, and it’s like, what am I gonna do?

I can’t, what am I gonna yell at them? It is a very different industry than it was a year ago. So I think acknowledge that you are in a liminal state, acknowledge the frustration and the change, but I think you also just have to acknowledge that this new power exists and you don’t control it.

Social media gives you an illusion of control. You can yell really loud. So that was the last, big web revolution with social media. And it let you feel that you had a voice and it made you feel that if you didn’t use that voice in a very specific way, you would be betraying something, right?

So we’ve been going on like 10, 15 years of, you gotta do something, you gotta be in there, and now this thing has happened and it keeps happening, and you’re not gonna put it back in the box. You can’t yell it away. It won’t behave, it won’t do what you want.

And I think just accepting that, and, you say that and people are like no, it won’t. And I’m like, okay then it’s down to like government regulation, terrorism. What are you gonna do? And they’re like it just has to stop. And I’m like, I can’t do that for you.

And so just like accepting that we’re in this moment where we’re a little bit powerless because the next reaction people have is no, you’re not, you need to do something. And it’s no, I am powerless to go back to the way that the technology industry used to be. I cannot go back. It won’t work.

So that’s A.

So then B is, what are your limits here, right? My limit is, it never is gonna write a word for me. I don’t want it to. If AI’s gonna write a word, it can say at the top, this was written by AI, but it’s never gonna be me, so I’m not gonna let it. That’s just, sort of like, the writing.

But the coding, I don’t feel as strongly about the code. And so figuring out what the parameters are and so on. And then acknowledging that this is equivalent to sitting down to a Thanksgiving dinner mentally, where you sit down to dinner and it turns out that you’re actually having Thanksgiving every day, which is a lot.

It’s nice once or twice, but then you’re like, I’m good. And then the turkey is made out of nougat, like, it’s so much coming into your brain. You just, you’re like, ah, oh, you know, you thought you just want a sandwich sometimes and you can’t even get it because Claude, Anthropic just released 25 new products in the last 40 minutes.

And it just keeps accelerating. It keeps accelerating. Everybody’s in it and everybody feels like they’re on an island.

And the design team feels abandoned because everybody’s so excited. And the CEO is vibe coding and the engineering team is like, what the hell am I gonna do with all this mess? I don’t know.

It’s like the classic thing. You gotta talk to your neighbor, you literally, you have to go and talk. Nobody will do it. Nobody will be like, Hey, how you doing? Nice to see you. That’s too much for us right now. Cognitively, we just can’t and can’t handle it. But that’s it.

So I, I talk to everybody about it. We host events. When people hate it, I listen. And I’m trying to make this a part of my social reality as opposed to just this thing that’s being thrust upon me. And that helps me.

Mostly I love it. You know, Simon Willison, the developer, he’s as far into this as you can go.

They talk about the idea of Deep Blue, which is a funny pun if you’re a big nerd, but like the idea that, you get this sort of sense of sadness watching your career get devalued.

And I think that’s fair. You should acknowledge it.

But also and this is the last bit I’ll say on this, but like what the hell was our goal? Our goal was to give this to everybody.

This is why the web was cool. Everybody could publish, everybody gets to participate. And it is weird and upsetting that we built this enormous multi-trillion dollar mega thing that has made a lot of people able to have private jets and gotten a lot of other people good jobs in houses.

That part, a lot of that’s really good. But the flip side is that everybody’s gonna get to make a thing. It’s like back to Xerox PARC ideas from the seventies and in all the anxiety, we keep slipping away from that. I want to give everybody all the software they ever could want and see what happens.

I think that’s amazing.

Jesse: Fantastic. Paul Ford, thank you so much for being with us.

Paul: Anytime my friends.

Jesse: If people wanna catch up with you on the internet, how can they do that?

Paul: Oh God. The worst thing is, it’s just LinkedIn these days. But go to Aboard.com our website. You can easily find me ford@ftrain.com. Paul dot ford@aboard.com. I’m on Bluesky, but not as much as I used to be. Anyway, let’s not even do that right now. Let’s, anyway, I’d love to hear from everybody. Everybody love it. I love it. All the thoughts.

Peter: All the thoughts to Paul. Thank you so much.

Paul: Blessings, my friends. Thank you.

Jesse: For more Finding Our Way, visit findingourway.design for past episodes and transcripts, or follow the show on LinkedIn. Visit petermerholz.com to find Peter’s newsletter, The Merholz Agenda, as well as Design Org Dimensions featuring his latest thinking and the actual tools he uses with clients.

If you’re looking for help with AI transformation or you just need a private advisor to help you solve your hardest leadership problems, visit my website at jessejamesgarrett.com to book your free one hour consultation.

If you’ve found value in something you’ve heard today, we hope you’ll pass this episode along to someone else who can use it. Thanks for everything you do for others, and thanks so much for listening.