Pragmatic AI with Matt Stauffer

Matt Stauffer:
Welcome back to Pragmatic AI, where we talk about using AI in the real world. What works, how to use it well, and when it causes more harm than good. Practical tools, real trade-offs for builders and business leaders. My guest today is my friend Jordan Keller, writer, producer, content director, ⁓ burgeoning programmer, you know, just constant curious person. So Jordan, obviously you have a pretty complicated role, set of roles and set of interests. I did kind of when I was pitching and I said, Jordan, my friend, the writer, right? So that was kind of like the easiest one, but obviously you're much more the writer in your current present. So can you just kind of say hi and talk about like, what do you do? What have you done? And what's the world that you're in that will be the context for our conversations around AI?

Jordan Keller:
Sure. glad you noted the word curious there and I was reflecting on how I kind of wanted to pitch myself here because there's a lot of ways I could do it. I think that is the defining word of what I do is I can do a lot of things. It's sort of the jack of all trades, master of none, maybe a few. ⁓ But it's curiosity and following ideas wherever they go, really obsessing over the how does this work and why does it work the way that it does.

I apply that to a lot of different areas. ⁓ I was a writer, when we first connected we bonded over being English majors who quickly face the reality of what it's like to bring that degree into the world and have people be like, I don't really see any money here. It's true. ⁓ But I think ⁓ I'm probably getting ahead of myself here, but one of the things that I've appreciated about you and I working together,

Matt Stauffer:
Yep. Yep.

Jordan Keller:
The thing that made me want to work with Tighten kind of the first thing, was the values page and the way you articulated what your company was about and how you think about things. You articulated the why and the how in this very clear and human and nuanced and concise way, common language. Like it just hit me so clearly that this was a group of people that were thinking about things the way that I wanted to refine my own thoughts and get sort of the machinery

a little bit tighter. ⁓ I've completely gone off the rails with this question here now, like, yeah, so I, so I, I ⁓ professionally, I, was, I was at Groupon for the bulk of my career. ⁓

Matt Stauffer:
Well, that's okay.

Jordan Keller:
which was an interesting experience watching the entire life cycle of that company from sort of the Halcyon days of it being this like startup darling, the fastest growing company ever, whatever. ⁓ The main thing I took away from that was working as part of an editorial team where we valued language. That was the defining part of that brand was we had all of the writers who graduated into the recession with no ⁓ way forward wound up here. So I was just surrounded by all these really cool creative people at actors, comedians, musicians.

Matt Stauffer:
Mm-hmm.

Mm-hmm.

Jordan Keller:
writers, screenplay writers, and ⁓ everyone was working on stuff outside of work. That was a prerequisite to the job. As they would interview you and say, what is your, what's the thing you do outside of work? I was a songwriter. I was a filmmaker ⁓ and dabbling in all sorts of creative stuff. So ⁓ I got to kind of bring that natural creativity that I've always had that's in there, ⁓ that, you know, is a blessing and a curse throughout life. ⁓ And ⁓

Matt Stauffer:
Mm-hmm.

Jordan Keller:
Marry that with ⁓ operationalization of it. We had a, god, was like a 200 person editorial team, I think, at the High Watermark, which is insane. I know, it's insane to even imagine now that we had fact checkers. We had people that we paid to just check the facts. That was a whole team. ⁓ I think about this sometimes in the layers of quality that we had, and it ⁓ was really run like a newspaper for a while there. ⁓

Matt Stauffer:
Wow.

Mm-hmm.

Yeah.

Jordan Keller:
formative experience, I transitioned into more of like a video and content production side. So I was writing and producing and directing ⁓ brand and marketing content videos. And then in the last kind of phase of that, of my time there, I transitioned into functionally like an associate content director role, ⁓ worked on a lot of like big global brand campaigns, marketing stuff. But through the whole way, the process has been ⁓

Matt Stauffer:
Mm-hmm.

Jordan Keller:
shaping ideas and ⁓ I always try to bring it back to like the thing that again full circle here the thing that connected me to Tighten was like the human aspect of it ⁓ and I know that gets thrown around a lot and I'm sure we're gonna to talk about that today but like the ⁓

If that's not there, I think the idea or the work isn't long for this world. You have to be hitting people in the heart and solving real problems that they're facing and the words and languages and sounds that are used to kind of make that connection. I just obsess over that stuff. I think it's so fascinating. Endlessly fascinating. Endlessly frustrating, but endlessly fascinating.

Matt Stauffer:
Yeah,

So and you had mentioned Tighten a couple times. So what you're doing right now with Tighten is you're helping us with some writing and editing and some kind of content planning. ⁓ But as you mentioned, so you you write music, you produce, you know, you act, you write ⁓ content, you do marketing. So you're in like the creator world, but not just like I'm an I, know, not a content creator, not a, know, not just a creative, because you've also done this in the corporate world.

Jordan Keller:
No, unfortunately probably. ⁓

Matt Stauffer:
So

writing, marketing, ads, all that kind of stuff. And so that, ⁓ I told you before we started, I'm excited to have you be my first person on the podcast. It's not either a programmer or someone who's running a programming company in some way, or form. Obviously you understand tech. You're working with Tighten. You are inherently a technologist just because of being a nerd. But that's not your day-to-day job, right? ⁓ So I wanted to start with kind of just talking about what is your personal relationship

Jordan Keller:
Mm-hmm, have to be.

Matt Stauffer:
to AI and then I wanna kind of go from there into what you understand your industry's relationship to be. And I wanna kind of place a little kind of like seed of that conversation where at times I've said, you know, we're dealing with this and you're like, oh, that's what we dealt with in writing two years ago. So I kind of want us to move in that direction. But let's start with personally, what do you, know, were you being prompted to use it at previous job? Were you being prompted to use it at current jobs? Are you excited to use it? Do you hate it? Kind of what's your current relationship with AI when it comes to

Jordan Keller:
Yeah.

Matt Stauffer:
actually doing your the work of your career?

Jordan Keller:
Yeah, so I'm going to connect this to 2020. When did Elon buy Twitter?

Whenever that happened, was the beginning. That was, that's the beginning of this story from, from my experience because I was a lifelong Twitter lurker. Again, I'm not a content creator. I have a, I, this is my first time I'm putting my face and name on a thing really that's like this raw and open ever. So, ⁓ but I was on Twitter. I curated what I felt like was my life on Twitter. could watch, I could look at all the content I wanted in the order I wanted. I could do my sports stuff. I could do my music stuff.

Matt Stauffer:
huh.

Okay.

Jordan Keller:
I could do my film stuff, could do my politics stuff, I could do, you know, it was all in one place. And then as soon as that happened and ⁓ the algorithm started changing, and I felt it, like I could tell that the experience was different. This was now three years ago. It's quaint now because it's gone just completely in the opposite direction. Not opposite, further direction. ⁓

That shook me in a weird way that I'm still dealing with and AI was the next wave of that. So I'm sure everyone's heard of the term enshittification Like I was following that really quickly, the launch of blue sky and all of that and thinking about decentralization of the internet versus ⁓ further entrenchment of centralizing infrastructure and allowing algorithms in the way that our media is consumed to be

controlled by a very small number of people. ⁓ And that, I don't know, that's probably set me up and put me in a mindset for how the first wave of AI hit, which to me, it just felt like an extension of that. ⁓ Initially,

⁓ The way I experienced AI in my personal life was that it was the tail end of my time at Groupon. We had a private equity company come in and as soon as that happens, everyone's kind of like, okay, you know, we're not long for this world. And I was very real with the team that I managed. Like, hey guys, like we all know what's going on. Like prepare, grab your work. That was a big thing. Like back up all your work because it's not going to be yours. You're not going to have access to a lot of hard lessons that I think a lot of people in the tech world, especially like the Zerpi tech

Matt Stauffer:
Yep. Yep.

Jordan Keller:
company world ⁓ have been learning the hard way. So it wasn't a surprise when the layoffs happened. And that would have been summer of 2023, I think. ⁓ And this was concurrent with the launch of ChatGPT. So in that six-month period where we're kind of waiting for doomsday,

you've got people who are saying, we can just automate, we can use AI to produce the content now, we can use AI to write the emails now, we can use AI to write the marketing stuff now, and this was early days of chat GPT and it just couldn't. And this is the thing I'm kind of curious to pick your brain about we'll see if we can talk about this as much as you want, but as a programmer, probably can look at code and instantly be like, this person knows what they're doing.

Matt Stauffer:
Yeah. Yeah.

Jordan Keller:
It's lean, it's packing as much meaning into the least amount of units as possible. I think writing works the same way. think like... ⁓

Matt Stauffer:
Pretty well, yeah. Yeah.

Yeah, for sure.

Mm-hmm.

Jordan Keller:
Conversation is one thing. I'm a rambler, unfortunately. But when it comes to writing, I really do try to get things like as tight and concise as possible because I know people's attention. Like, you know, they don't want to sit there and think, listen to every stream of conscious thought. ⁓ But just being able to look at the product that was coming out and knowing as a writer and as someone who has devoted his life for mostly worse, but in some instances better to like language and ideas and articulating

Matt Stauffer:
Uh-huh.

Jordan Keller:
them elegantly and concisely and meaningfully and to just not see it and feel like I'm losing my mind that someone is telling me this is good work, this is passable, it's good enough. Well is that what you want to be? Do want to be good enough? Do you care to like provide something of value?

as a brand, a company, whatever you are, as a person, or is it just, it's why I hate that word content. I hate having to attach that to my job title, because content is just anything. I like to get a little more specific with it and say this, here is what it is. It's a video, it's an album, it's a song, it's an essay. ⁓ So, once the, ⁓

Matt Stauffer:
Mm-hmm.

Mm-hmm.

Jordan Keller:
I took some time away and just kind of decompressed and that was the beginning of what has been, I think Jeffrey Way said this in the episode that I just listened to, that you just did, ⁓ released yesterday, that programmers are the canary in the coal mine. And this isn't the first time that a programmer, and especially working with you guys at Tighten, has said something that all of my writer friends and I said two years ago, three years ago now, like, I feel like we're the canary in the coal mine. I feel like this is coming. Everyone doesn't realize that this is coming.

It's trite to say it now and it's not to be like I was there first. But it is true that just the inherent fact that LLMs work off of language and work off of syntax and rearranging syntax and knowing how to kind of put those together, that it makes sense that kind of the first two ways that are really getting hit by this are writers and programmers. Those are different ends of the same spectrum. ⁓

Matt Stauffer:
Yeah.

Jordan Keller:
Yeah, now I forget even the question, but that's that's sort of that was that was like my experience of of the initial experience of kind of the AI wave hitting was like, well, I have to learn this. I don't have a choice. like, you know, this is a common refrain that every guest has said, like you don't. like you can you can rage at the skies all you want. And certainly on some days I still do that. like.

Matt Stauffer:
Yeah.

Jordan Keller:
You gotta figure out how to make it work for you is where we're at. It's just the nature of the beast.

Matt Stauffer:
Yeah, so

the question was, what is your career relationship to AI? And I think that, ⁓ you know, I mentioned that I want to talk about that specifically with relationship to your industry. You mentioned that, you know, quite a few previous guests, all of whom have been programmers right now have said it feels inevitable. ⁓ Imagine that people listening don't have a pragmatic, practical understanding of the ways that the world of being a professional writer have been impacted by AI.

So let's kind of step it down from philosophy and get into practically 20 years ago, being a writer, working on the internet to today, being a writer, working on the internet. What is different because of AI?

Jordan Keller:
Mm-hmm.

There's man there's a lot of ways to get in here. think my initial...

I was trying to kind of articulate this to myself in preparation for this. Writing is the residue of our thoughts. Like, the way we think about things, the way we experience and feel things. We process that through our own unique brain. And it's all of the things that you've experienced and all of the unique riverbed of thoughts and feelings that allow you to flow through the world the way that you do.

Matt Stauffer:
Mm-hmm.

Jordan Keller:
The idea, and as someone who values that experience and that curiosity more than most things in life, the idea of being constantly tempted to offload that or automate it or give that process to someone else is the defining...

anxiety I have around all of this is that it's always there in the room. You can always just say, don't know, writing is hard. Like writing a good sentence is hard in the same way that like solving tricky problems in programming is hard. And it's always an elegant, it's often an elegant solution, a simple idea, but sometimes it takes a long time to come around the circle and get to it. The idea of just being able to, ⁓

throw some slurry of thought into an LLM and be like, you know, sort this out for me. Write me 10 sentences and I'll get one that's close enough and then I'll spruce. Like the idea that that temptation is there for someone who...

values this and and you know for whatever it's worth is like kind of defined his life around this i can't imagine what it's like for writers who are breaking in i don't even know if there are writers breaking in now like i don't need like in the same way there aren't probably junior programming roles that that are going away like who's hiring a junior copywriter to come in and write ad copy just to get the chops and kind of learn the company and learn the ethos and

Matt Stauffer:
Mm-hmm.

Jordan Keller:
⁓ learn the voice and all that stuff when you can just ⁓ throw it into an LLM l and and get out something passable

Matt Stauffer:
Well, you mentioned

the temptation. So you said there's the temptation to write this way. And as a writer, you have your philosophy and your desire that I want writing to be this way. And there's a temptation. once again, let's hone in really practical. Is that temptation coming from you? Is it coming from employers? Because the thing is, I want to poke at, like, I've heard stories about people losing their jobs because why hire 10 writers when you can have one editor who reviews the output of chat? But I don't know if that's actually a common thing.

Jordan Keller:
Mm-hmm.

Yeah.

Yeah.

Matt Stauffer:
One of the things we're dealing with in programming is that there's a balance between what our clients are expecting, what our boss is expecting, and then what we're expecting. And all of those put an impact into it. So I want to know from you, what are the pressures coming from? Are they your own internal pressures that it can be easier if you use the AI? Is it your own internal pressures that you feel like you're supposed to be faster because of AI? Do you have clients and employers telling you, I need 10x the output? are they changing your role from writing to editing? What is the practical impact on being a writer

Jordan Keller:
Yeah.

Matt Stauffer:
because of AI?

Jordan Keller:
Yeah, think the big one, at least from my experience, is sort of the ambient atmospheric pressure of the vibe that everyone has. ⁓ The general kind of, the consensus of people ⁓ who, I assume it's the same way in programming, who don't actually know what good writing looks like or like well-crafted writing. Yeah. And people saying, you can do it, just because you do something faster doesn't mean that it's better. ⁓

Matt Stauffer:
Mm-hmm.

Yes, 100%. Yeah.

Jordan Keller:
or that it's at the same quality. So yeah, think there's in general...

⁓ a lack of ⁓ there's always like that, like you articulated that ⁓ balancing of people's different values and needs and wants. The desert of ⁓ values alignment that we're in right now is just it's getting drier with AI where.

If you value good writing, you have a team of programmers who can also write well and articulate complicated concepts in a way that even me as a programmer, I can read an article written by Nico or someone in your team and be like, okay, I kind of get this. And then I've had this experience, six months from now I'll be like, that's the thing Nico was talking, know? Like, okay, all right, delegated types, I get it now. And...

Matt Stauffer:
Yep.

Jordan Keller:
The pressure to produce more, ⁓ more, more, more, faster, faster, I ⁓ really, the biggest impact it's probably had on me.

is ⁓ reconciling the external pressure of just the situation we're in with my own personal value of I really do want to do like meaningful good work. I want to do I want to be of use. I want to be of use to someone who is hopefully doing a mission that is helping contribute to society. And I just don't know that like that's possible in the climate we're in right now. I think everyone's just trying to survive. So the way you survive is

In some cases, you do what you have to do. You find the assignments that you can automate certain parts of the process, especially like the research sides, ⁓ some of the outlining part of it. And in the same way that ⁓ the recent rise of skills, from what I've understood with ⁓ OpenClaw and everything, delegating some of that work upfront so that you can spend a little more time ⁓

at the page with the cursor blinking and to not maybe have to do as much manual work. Like go and find me 10 articles that are talking about this concept and give me a three bullet point summary of them so I can know which of the three I want to read, actually sit down and read. That's a way that, you know, before the research phase might have taken, you know, four times as long. Now you can at least get a little bit, can chunk that down a little bit better now. So.

Matt Stauffer:
Yeah,

I mean my wife and I talk about because she's among other things she's a writer and we talk about You know, she was very early in chat GPT and was very People around her would be using it for writing prompts and be using it to even do drafts and she's like, it's not creative It's not good. She's like I was trying to tell a story about mermaids and every single time no matter what it would just come back and be retelling the Little Mermaid it just didn't have any understanding outside of that. It couldn't help it right in But but she has found it to be particularly helpful for example as a better Google and I'm like, yeah AI is a great, you know

Jordan Keller:
It couldn't help it. ⁓

Matt Stauffer:
great better Google or a summer summarization tool research tool. So there are certainly ways where if we find ourselves in a world where it is something you feel like you have to use, there's ways to use it. And we talked about this with the programmers, but what is it good at? What is it not good at? And what way does it help me in a way that I want help? And what way does it help me in a way I don't want the help? But if you were to have put your head in the sand and said, I'm not going to learn AI because there are people who doing that, right? I know if they're listening to this podcast, but there are people I know personally.

who are ⁓ saying this is something I don't want anything to do with for whatever reason or other, do you think that today your job prospects would be impacted? Or would have been impacted already?

Jordan Keller:
Yes, ⁓ but also they would have been impacted anyway. You know, like, I'm not sure. Again, I come back to this feeling that ⁓ writing and programming being good, you know, perfectly engineered to be completely screwed up by all of this, you know, like upended by all of this, just by the sheer nature of the structure of.

Matt Stauffer:
Okay.

Jordan Keller:
piece of code versus a piece of writing. Like it's a thing that an LLM can do passively well. ⁓ If I were to have, I mean, I,

I've moved through the, we've talked about this, like I've felt a little bit like an emotional doula to the programmer folks over the last three to six months with everything that's shifted in that world because being in on the Tighten Slack and watching everyone kind of process this in real time, again, it all felt very familiar. was like me and all my friends kind of worked through this in a similar way, lot of similar sentiments being passed on. Screw it, I'm not doing this. I'm not touching this.

And then justifying it to yourself why well yeah, it is it's bad for the environment and it's stealing people's copywritten work and ⁓ You know you can you can come at this with your own you can see it however you want to see it You know and and feel about it however you want to feel it but the whole point of this podcast in the article that we've been working on together is like We have to take our feelings out of it at a certain point and just look at what is this actually? How is this impacting?

Matt Stauffer:
Yes, and that's my question I keep poking at you at and we're 23 minutes in and I'm still trying to get you to give me a very specific answer. If you don't use AI, have you harmed your... because one of the things you've said here is you've said it's inevitable, right? You can't not do it. That's one of the things that we're writing in this article we're talking about is in programming we've gotten to the point where it feels like it's inevitable. You have to, because of outside pressures, specifically...

Jordan Keller:
I'm still trying to answer yes.

Matt Stauffer:
expectations of clients, expectations of the pace that we're working at. Like the feeling of inevitability is not a feeling, it is actually like I can point to the things. So I wanna ask as a writer, if you had said, screw the inevitability for whatever reasons, which we can talk about in a second, can you say that a writer today is less employable if they are not engaging with AI? They were more likely to lose their job, they're less likely to get a new job.

Or are we at a place where there is this vibe, but it might not actually impact somebody's, and I don't know the answer. I need you to tell me. Is a writer less capable of having writing as a career if they're not engaging with AI in their work?

Jordan Keller:
Yeah. Yeah.

I think it depends on what sector you're going into or what sort of area you're in. ⁓ I can't imagine truly like creative and production related writing jobs. I imagine all of them are looking at the AI discourse and just, you know.

Matt Stauffer:
Okay.

Jordan Keller:
rolling their eyes or laughing. Ultimately, it's people in a room, like writing a comedy show or something. You have to be writer. You have to have it or not have it. If you're looking at being a writer for brand marketing side, content marketing side, I think you kind of have no choice. I don't know what... I can imagine a future where that...

Matt Stauffer:
Mm-hmm.

Jordan Keller:
just goes in a sense goes away or is just you know someone does it as part of their other more important job. ⁓ you know Jamie on the team is really great. Let's just give it to her. She'll come up with a funny email. She'll come up with a funny camp like.

Matt Stauffer:
So what's the difference there? ⁓ Because one of the things you mentioned earlier is you talked about the fact that people with the taste to see good writing, the taste to see good programming, we all know that some people who are paying money for our services have that taste and some don't. And the ones who have the taste are the more likely to stick with ⁓ artisanal code, artisanal writing, because they say even if it's A, B, or C, promise of better work.

Jordan Keller:
Mm-hmm.

Yeah.

Matt Stauffer:
faster work, cheaper work from AI, say, but we know that the good stuff comes from this human kind of, or human and AI, paired together, whatever. But taste is at least one of the delineators. Do you feel like there is less writing work for people who don't have taste? Because those people in the past were forced to pay writers to get writing, but now they can get tasteless writing for LLMs. Is that where the majority of the shift is happening? Or is that a false dichotomy that I'm making?

Jordan Keller:
Yes, yes, and those are so hard to find. Like, 100%, I think, 90%, sure, yes. And when you find those people, like when I found you and the Tighten folks, was like, oh, thank God, like people who, know.

there's still like, renewed my faith in all this a little bit, cause it was like, ⁓ you guys do great work and you're very lean and professional, but also like you value ideas and articulating ideas and benevolent ambition. I think that was one of the phrases that stuck out to me in your values page back, like, you know, be open with your knowledge and share it, ⁓ connect with people. ⁓ In my experience,

I don't even know the ratio, man like 98 to two, where 90, and this is probably the way it's always been with tracking down leads and building out your client list and everything. I've applied to a lot of jobs over the last three years, and I've worked with a lot of clients over the last three years. And I can count on two hands, the ones that really wanted to do

We're really focused on the taste and doing something that was really of quality because it in and of itself is inherently good to do that. And not necessarily because it's the thing that's gonna be the most effective or viral God. People are still chasing that in this day and age, it's insane to me, but like, yeah. Yeah.

Matt Stauffer:
to the bottom line or whatever. Yeah. Yeah.

Well,

you mentioned, you know, like, one of the things that we're talking about often is the writing technical blog posts. it's our blog posts take weeks to write because we're really, really, really intent on them being incredibly carefully narratively structured to make sure that a newcomer who doesn't understand it can. So like I have at a time in the past been able to crap out an entry level introductory blog post in an hour.

Usually it's something I know so well and I've been thinking about teaching for so long it's super easy. But generally our blog posts don't take an hour, they take weeks. And there's always technical content out in the world that is not considering those things. It's like it has the content on the page, therefore hit publish, right? And the consideration of what a good blog post and a blog post and the difference between the two is always something that's been missing. But often there's been a barrier that there's so much work to even get a blog post out there that

By the time you're doing all that work, you might as well actually do the work to do it well. And now that barrier's gone, right? You can literally just hook up a, what are the things called? I don't know why it's escaping my brain. But the thing that basically allows your AI to talk to your blog, and then you tell your AI, hey, go write a blog post about this. And it just does it, right?

Jordan Keller:
Yeah, yep

Matt Stauffer:
⁓

And I have a friend who did that recently as a proof of concept. He's like, hey, this entire blog post is written by my AI agent. I added an MCP server. I added an MCP server to my blog, and then I told it to write a blog post about how to write an MCP server to your blog. And the thing is, it wasn't a good blog post. Love my friend who did that, and he knows that, right? He's fully aware of that. But the thing is, there's a lot of people who don't know that that's what's going on. And I'm like, there is a proliferation of not effective communication, not effective ⁓

Jordan Keller:
Yeah.

Right.

Matt Stauffer:
curation of the experience of consuming a piece of writing or whatever else and actually enjoying it, learning from it, being able to remember it later. Like there isn't, I don't want to be some pretentious blowhard, but like there is an art and a science combined together to go from I have something I want to teach someone to I've taught it effectively to them. And it can look similar to here are some words on a page that talk about this thing with a

Jordan Keller:
Yeah, better in the right

shape. Yeah.

Matt Stauffer:
Right, exactly. There's

a paragraph break every four sentences and the sentences all make sense in English, but it doesn't have that same curation. ⁓ So I'm very curious as you think, just like as we're thinking of programmers about how do we not get left behind? How do we take advantage? I told my whole team you need to become experts on LLMs if not just to say here's why I don't ever use it or here's why I use it and the things I use it for but the things I don't use it for. ⁓ There's got to be, we got to learn it so that we can understand it.

Jordan Keller:
Mm-hmm.

Matt Stauffer:
What has that process been like for you as a writer, as a creator of content? What is your, cause you said, look, I just feel like I got it. I got to learn it, right? So what has the learning process been like? What have you tried it on that's worked or you tried it on that hasn't worked? What's that learning process been like?

Jordan Keller:
Mm-hmm.

Mm-hmm.

funny because I was probably vibe coding before vibe coding was even a word because it was just by accident because I would look at Tighten's code and see these weird classes of things. I'm like, what the hell is this? That led me to Tailwind. And I was like, oh, this is interesting. This was three years ago now. I'm like, okay, well, this is really cool. And then I started playing with Tailwind. And then I was like, well, what is that thing? That's Laravel. Oh, what's Laravel? So I was...

Matt Stauffer:
Uh-huh.

huh.

Uh-huh.

Yeah, yeah.

Jordan Keller:
My main experience of AI and LLMs was actually like trying in a very inefficient way, and we can talk more about this some other time, to learn how programming works. I was like, if I had a teacher, it would have been a lot more efficient. But it was just banging on things and then accidentally finding a way in. So while that was happening, I'm also doing all of my ⁓ freelance and contract work, work with Tighten and ⁓ working with my clients.

Matt Stauffer:
Mm-hmm.

Jordan Keller:
And the two started to kind of overlap in a sense where I found myself wanting to define what are my first principles as a writer? What are the things that I go back to over and over again when I'm presented with an assignment with a challenge with a problem to solve?

Matt Stauffer:
Mm-hmm.

Jordan Keller:
and getting those, again, inspired kind of by Tighten's approach to this, get those as tight and repeatable and I can, now this is my skill utility belt. If I were to throw this in, I've already thought like maybe there's an open-clawed process here where you can fly these in and be like, okay, we're gonna write this in the Lazarus-Igri dramatic framework way because it's a story, you know, and this is more of the brand marketing side, so leave that out and pull this. ⁓

Matt Stauffer:
Mm-hmm.

Jordan Keller:
A way that it has ⁓ influenced my process is it's got me thinking about first principles and like the architecture of good writing and the pooling all of the many books and experiences that I've had and trying to really get that down into like a ⁓ guidebook for myself so that I'm not reinventing the wheel every time I'm faced with a new assignment. It's not like.

that early stage writer panic where, you know, your creative director gives you assignment. You're like, great, I got it. And then you get back to your desk and you're like, what did I just agree to do? I don't know how to do a marketing campaign that touches all these channels.

Matt Stauffer:
Well, so I'm so sorry to keep interrupting you, but you keep going places where I'm like, I gotta touch on that. So if you were to sit down to have a writing task for a client today, do you use AI in the initial process? Do you use AI to write any of it? Do you use AI to edit any of it? Do you ask for AI's feedback? What are the practical touch points where AI interacts with your career writing work?

Jordan Keller:
Yeah, it's okay.

Mm-hmm.

Yeah.

Yeah, the big the big the big thing that it has been helpful in is the research part which Again connecting this back to like Groupon that was that the balance was supposed to be I think it was a 60-20-20 60 % researching 20 % Writing and 20 % editing that was kind of the balance you wanted to take on an assignment and I still kind of hold that is like the

Matt Stauffer:
Yeah, you mentioned.

Okay.

Jordan Keller:
The bulk of what you're doing is learning and the bulk of what you're doing is trying to understand more than you need, more than you need, so much more than you need. that, I told this to people back at a certain point in my career, I get to become a temporary expert in one subject a day and then I throw it out. I have to write my one video a day so I learn everything I can and then, but, ⁓ and so AI has been really effective in,

Matt Stauffer:
Yeah, 100%. Yep. Yeah. You gotta be like a little mini expert so that you write it well,

Yep. Yep.

Yep.

Yeah.

Jordan Keller:
Helping that process Allowing me to but but this goes back to a thing we've talked about a lot that's because I know how I want to use it because because I Already knew how I want like I'm I worked really inefficiently and can I swear on this podcast? I worked really inefficiently and shittily, you know many times throughout my career and so to know I've got a reach for

Matt Stauffer:
100%. Yeah.

Yes, you can.

Yeah.

Jordan Keller:
this tool to do this assignment, that comes from experience. And this is a big thing we harped on in the blog post we worked on. Your expertise, there's knowledge and there's wisdom. And the wisdom is the thing that you get from applying your knowledge and having it not work and reacting to the problem. You're being like, oh crap, that didn't work. And you develop this robust set of experiences that allows you to react at a given moment. So yeah, the big thing is researching.

Matt Stauffer:
Yeah.

Yeah.

Jordan Keller:
⁓ Following ⁓ random ideas and quickly pulling out of the rabbit hole, that's another thing because AI is so sycophantic that it can be like, that's a great idea. And if you're not careful, you'll be like, I'm a genius, whereas I assume I'm a moron. like, know, explain this to me like I'm, often, my biggest thing is like, explain this to me like I'm an 18 year old, you know, who's like, whatever, I don't need to know.

Matt Stauffer:
Mm-hmm.

Let's go just, yep. ⁓ huh. ⁓

Jordan Keller:
Because that'll get you a concise version of the idea. It's like, nope, throw that away. Don't need it. What about this one? And then eventually when you find the thread that's worth keeping, that process might have taken four hours and would have required reading 10 articles. Now it's, I can quickly glance at 100 articles, chunk that down to 20.

Matt Stauffer:
Mm-hmm.

Jordan Keller:
you know, maybe summarize those 20 and then read 10 and that time is really helpful. It's helpful in the research phase for sure. The writing side of it, absolutely not. There've been like maybe a handful of times where I'm like, I just need to get this done and off my desk and this one sentence and I'm exhausted, just give me like one sentence here and I'll look at it and I'll usually hate it so much that I'll be like, no, it should be like this. So like out of spite, you know, I'll be like,

Matt Stauffer:
Yeah.

Something on the page? Yeah, uh-huh.

And you're inspired to write.

Jordan Keller:
I'm not gonna let the AI win this. ⁓ But yeah, no, writing, God, every time I've tried to even just attempt it, it's just mortifying. I've even loaded up some of my writing and then be like, write this in my voice, and I'm like, that's what I sound like? That's what you think I sound like? Yeah.

Matt Stauffer:
Yeah. Yeah.

Mm-hmm.

Please tell me it's not true. Got it. Okay. So,

so at least in your kind of world, there is an impact because this expectation from clients and from other folks that you're going to need to be, you know, faster or whatever. But let's, let's say you've got a job where the person has taste. ⁓ you know, this is in the programming world. I've been saying for a while, ⁓ you you're familiar with the word fungible, but for those who are not familiar, you've probably heard it through NFTs. Fungible means

something can be exchanged with something else that does the same value and you don't care which one you have. So usually for most folks, gas is fungible, hair gel is fungible, ⁓ milk brands are usually fungible, unless you're like, only buy this one milk and then it's non-fungible, right? And so we've said for a long time that as professional service firms, as experts, you're trying to be an expert writer, I'm trying to be an expert programmer, we don't wanna be fungible. We don't want people to say, I'll just...

Jordan Keller:
Yeah, right.

Matt Stauffer:
get whatever programmer is giving me, because fungibility means race to the bottom price, right? If they're all the same, you want the cheapest, right? Because why not? ⁓ So you're obviously working to not be a fungible writer here, because I believe that fungibility means replaceability with LLMs pretty easily, right? If you are writing the same code as everybody else,

Jordan Keller:
Yep.

Matt Stauffer:
That means that's a really predictable set of work, which means an LLM could probably write that code. Whereas if you're writing code with a unique set of perspectives and experiences, whatever that other people don't have, then it's very good chance that LLMs can't bring that as well. And so I imagine you're doing the same thing from a writing perspective. You you're trying to say, I am able through my decades of experience as a writer, as a producer, as a director, as an editor, to do things that, you know, mean that an LLM can't replace what I'm doing. Okay. But I still imagine that

for you like for me, there's an expectation that, okay, cool, you're an expert, but that doesn't mean that the advent of AI shouldn't still make you cheaper, faster, better, right? You mentioned using AI for research and not using it for writing, which is one of the reasons why you're working for Tighten because I could not, if you're like, I'm just gonna spit out a whole bunch of LLM generated crap for you, I'm gonna be like, cool, I can do that myself, right? And so I think that's where the thing is, like, there's just an angle of like, if I can use the,

Jordan Keller:
Mm-hmm.

Ahem.

Matt Stauffer:
the thing the same way that you're using the thing then then I don't need you to do it and if you have the ability to do something that I can't that comes from your background then cool you know are you finding it possible to talk to people who are expecting you're going to be three times faster whatever and for you to be able to establish yourself saying I'm not just the expert at writing because I'm the expert at writing I'm the expert at knowing what using an LLM in writing looks like and then are they listening if you're trying to establish that

Jordan Keller:
Right. Right. Yes.

I had actually the short answer is I've found that to be very hard and it has been kind of the defining problem of the last few years of my life is...

Matt Stauffer:
Yeah.

Jordan Keller:
articulating the value of that, what I believe strongly to be the value and also like, think depending on the mood, I know it's true because I get to work with good clients who ⁓ value that and see it and implicitly understand it, but I also feel that it's not true and it's just a thing I'm hoping is true because otherwise I'm screwed. ⁓

Matt Stauffer:
Yeah. Yes. We're all there. We're

all there.

Jordan Keller:
Right, right. That's

the thing, we're all in the same on fire ship. ⁓ But actually the conversation I had with your friend Carl, I an intro conversation with him a couple weeks ago, and it was the first time where I felt like all of the threads were coming together of the last few years of my experience where the problems he was trying to solve were about...

Matt Stauffer:
Mm-hmm. Yeah.

Jordan Keller:
It's not just producing content because again, I think like the output often is like fungible, whatever. You know, like I've written a lot of fungible copy in my days because I had to to survive and get through periods of life and periods of a company where I didn't need to be art or high value or whatever. ⁓ That's just the way it goes. ⁓

it was less about the output and more about the how did we get there and why are we doing it? meeting people who are, again, we're all on this on-fire ship trying to figure it out, meeting people who are trying to solve that problem and see that that actually is where the sauce is, how you think.

why you do things the way you do and having really rock solid convictions and hypotheses about why you're doing things the way it's not just because your taste. That's easy. You can be like, I'm a good writer. I like good music. watch good movies. Like everyone thinks that they have good taste, whatever. It's I have the lived earned work experience to say, this is one of the things I pitch all the time. one of my biggest values I can provide is here's what you shouldn't do.

Matt Stauffer:
Yeah. Yep.

Jordan Keller:
You should not start a TikTok and start, you know, this was years ago now, but like don't channel chase. Don't put all your eggs into one channel's algorithmic basket because someone can buy it and upend it. And then your whole communications pipeline is gone. Like you have to own your work. And because you own it, you should care about how you made it and what it's saying and what value it's providing and ⁓ making people...

Matt Stauffer:
Yes. Yeah.

Jordan Keller:
not making people, because that feels like it's a evangelical thing or something, like I'm preaching, but when I meet people who understand the value of that, I get excited about AI, because in the right hands and with smart people who are doing what they're doing for the right reasons and are genuinely solving problems, are genuinely invested in the work they're doing,

Matt Stauffer:
Mm-hmm.

Jordan Keller:
It is an exciting time to imagine.

Some of the, like, there were so many layers of bureaucracy at various points in my career where you'd have to, like, run things through the team of 40 or 80 people and ultimately it's about, like, a video script. Or it's about, like, you get it, if you have layers of middle management, you're gonna have someone who needs to put their thumbprint on something and say, well, otherwise my job doesn't count and why am I here? Like, it's a good question, like, why are you here? So, yeah, yeah, yeah.

Matt Stauffer:
Well,

I appreciate because one of my last questions as we're wrapping was going to be what's exciting for you? ⁓ And I've even though you just said this I am gonna kind of throw it at you one more time which is ⁓ As a programmer not every programmer many programmers. know I've been like this is an exciting This is just more crap I have to deal with but a lot of programmers I know I think the more entrepreneur ones who I've said this in previous episodes who I think are not Able because of our roles in the community to write code all day long

are very excited because this is allowing us to generate more code than we could because we're CEOs or founders, whatever. And it's a lot of us who are like programmers who don't get to program all day anymore are obsessed with LLMs because now all of a sudden more of my ideas have the ability to come to reality. That's the exciting thing I think for programmers. Where, what's the exciting, not the I have to, or the it's exciting for the employer because they can save money, but like as a writer, what's actually exciting?

about AI. Is there anything?

Jordan Keller:
Man, this is a... This is a... Clearly I'm ⁓ good at running off on long winding tangents here, so I'm trying to rein this one in, because there's a very long answer I could provide here, but...

Matt Stauffer:
Uh-huh.

Jordan Keller:
I want to connect this back to a thing that I said earlier, which was that this whole experience, this process, ⁓ the last few years of AI started with the death of Twitter for me. It was like, wait, this thing can go away. It got me thinking about...

Then the other thing is I released an album that I'd been working on for it was like my big it was the biggest thing I'd ever worked on the most amount of time I put all this time and money and energy and effort into getting people to record and engineer and had this big thing and it was a it was my baby and then I was met with the the reality of having to promote that and having to make content and like how do I Post on Instagram and looking at other artists and being like well I don't really want to do that and I don't really want to do that and actually don't want

do any of this. ⁓ And all of this has come around to like, what has me excited, it's a thing my friend Clayton said, he wrote a piece right around that time in 2023 about the importance of taste and curation in this next wave of the internet. I actually feel like there's been a big ethos

shift back into kind of like web 1.0 ethos. And the thing that has me excited is I'm not a programmer, but I can build websites now that allow me to put my work out there and talk about myself in a way that isn't running through someone else's algorithm. I'm not framing myself up for Instagram. I'm not framing myself up for TikTok. I had one briefly. I deleted it. It wasn't my vibe. ⁓ But there's a way that like you have to speak on those channels. ⁓

Matt Stauffer:
Mm-hmm.

Jordan Keller:
there's

a culture that forms around the way people communicate and it gets defined really quickly and then it's kind of locked in and there's phases and things change but what excites me about now is I get the sense that everyone is building something somewhere ⁓ whether it's a website or an app or whatever and I hope it encourages people to

Matt Stauffer:
Mm-hmm.

Jordan Keller:
be themselves a little bit more and to not try to keep up with the expectation. There's so many experts and content creators that are going to tell you what you have to do in order to break through. I've never been able to do it. just can't. I've tried at times in my life to do it. And what has me excited about as a creative person finally now on the cusp of 40, my gosh, is like I'm finally comfortable with just like putting my work out there and LLMs are allowing me to

Matt Stauffer:
Mmm.

Jordan Keller:
come up with an idea for how do I want to serve this up to people? How do I want to share this with people? And it doesn't have to go through some social platform or some that anyone else owns. I get to decide, for better or worse, again, this is how it looks. This is how it feels. This is how I talk. And now it's my voice too. I'm talking one to one to you instead of, well, anyone on Twitter could see this. Anyone on Substack could see this, whatever. I do think that that how, again, changes how

Matt Stauffer:
Mmm. Mm-hmm.

That's so interesting. Yeah.

with the experiences. Yeah.

Yeah.

Jordan Keller:
I approach sharing the work, which I think eventually the next few, you know, we've been in a very tumultuous moving and having a baby and everything. When I get back to doing a lot of creative work, I think that's going to trickle down to like.

Now I know I'm not posting this on Spotify. I'm posting this on my own website and then sending people to my website. Molly White, you're familiar with her, she was a big kind of, she writes a lot about cryptocurrency and she just has mollywhite.net and it's everything she does goes through there. And she's a big proponent of Posse, the publisher on site, Syndicate Elsewhere, that idea. I think that's bleeding back in. So yeah, what has me excited about LLMs is just.

Matt Stauffer:
Yeah.

Jordan Keller:
I get to decide how to use it. ⁓ It is ultimately another centralized thing that's being forced in everyone, but you do have some say in how you're gonna navigate it. So, yeah.

Matt Stauffer:
That's cool.

Yeah,

you're using that centralized system to be able to build systems for distribution so that you're not relying on other people's systems for distribution. I love that. I had not thought about that. We're at 50 minutes, so I can't talk about it more, but that's a fun thing. That's a fun thought experiment to kind of end it on is it's enabling, even though big AI is a big tech centralized thing right now at least.

Jordan Keller:
Yes.

Yeah.

Mm-hmm.

Matt Stauffer:
Is it potentially enabling decentralization in some other areas? Fascinating.

Jordan Keller:
Yeah. that's, mean,

this is again, I connect this to the blue sky and the open protocol idea of like, if we all have websites that can talk to each other, and we all agree on like the protocol, well then things get really interesting because then we just kind of have to agree on like, well, what are the, what's the grammar? What's the syntax? What's the lexicon? And then interoperability, I think is the big obsession I have over the next few years of like, how do all these things talk to each other?

Matt Stauffer:
Mm-hmm.

Mm-hmm.

Jordan Keller:
and help us connect more with the things we want and not be fed stuff because it's making someone money somewhere, you know, because it's good business for them to show. Yes, I know, it's coming back, baby. It was glorious. It was, I wish I was there, man. I was there for Web 1.0, man. Yeah, yeah.

Matt Stauffer:
Yeah, that is Web 1.0, rings, web rings and RSS, baby.

Yep. Wow. Okay, well this is great.

Several people have told me that I need to stop cutting these podcasts at 45 minutes. So to the listeners, I'm thinking about it because every time I'm I'm like, there's more, but I have not made the decision yet. that's another vote.

Jordan Keller:
I don't want you to cut them off. Every one you've had with ⁓ the

previous guests, like, damn it, I could have listened to another hour of that. Not today, no.

Matt Stauffer:
Okay noted I'll consider it not today I can't make the decision the spot but I'll consider it

well Jordan if people are fascinated by you they want to learn more about what you've done they want to find your album they want to learn more about your Concepts about open protocols and stuff like that. How do they follow your work?

Jordan Keller:
Well, I'm in the process of vibe coding my own site, which is jordan-keller.com. ⁓

All the programmers who listen are welcome to go there and look at my code and just ⁓ reflexively flinch at all the mistakes I'm making. But yeah, that's it. And I'll have links to all my, I'm on Blue Sky in the same place for the same reasons that we just talked about, but I'm trying to keep my ⁓ connection, my contact portal is my website now. So jordan-keller.com. There are other places there, but that's the main one. So yeah.

Matt Stauffer:
You

Okay.

Love it.

Thank you so much for hanging out today. I'm looking forward to talking more about decentralization and open protocols in our personal friendship. ⁓ Don't know if there's gonna, we were just gonna have to get back in the podcast later or what, but either way, I'm looking forward to it, so.

Jordan Keller:
Yeah,

get someone who can talk with authority on that because I'm the middle of the normies and the tech people. Two normie for the tech people and two tech for the normies. So get the heads over here.

Matt Stauffer:
You got it.

will do. Well, thank

you so much for hanging out today. really appreciate it. rest of you, we will see you all next time.

Jordan Keller:
Thanks, man, it was great.