Soft Skills Engineering - Episode 501: Vibecoding CEO and doing to teaching

Episode Date: February 23, 2026

In this episode, Dave and Jamison answer these questions: A listener named Derek asks, I am the CTO and cofounder of a startup. Now that vibecoding is a thing, our CEO has kind of gone r...ogue, and and he’s vibecoding a bunch of random stuff, one of which he bought a domain for and has pushed a potential customer to pay for, without talking to our team. I feel like this is fragmenting our focus, but I don’t want to ban our CEO from vibecoding and being creative. how should I handle this without damaging relationships? AdmiralFox asks, Hi Dave and Jamison! Longtime listener, first-time question asker here. After 14 years at a consultancy firm, I’m moving to a major retailer to become their Java Learning and Community Lead. Instead of shipping code, my new role will be shipping knowledge. I will be managing learning paths, organizing internal knowledge sharing events, and help managers screen candidates. Basically, I’m moving from a ‘Maker’ role to a ‘Multiplier’ role. I have 13 weeks of notice period (Standard European “I’m not leaving yet after 14 years” protocol) and I want to use my free evenings to prepare. My questions for you: How do I transition from “the guy with the technical answers” to “the guy who helps everyone else find the answers”? How can I use the remaining time of my notice period to prepare for the people side of this role? Love the show! Keep up the ‘quit-your-job’ advice coming (although I’ve already taken it!)

Transcript
Discussion (0)
Starting point is 00:00:00 it takes more than putting a token in the slop machine to be a great engineer this is soft skills engineering episode 501 i'm your host dave smith i'm your host jameson dance soft skills engineering is a weekly advice podcast for developers engineers who just want to pump those tokens into the slop machine isn't that its job it's supposed to do the token stuff not me you got to give it tokens to get tokens yeah no i produce i produce value i operate above this solid plane of tokens that's right you're definitely not just a cog in this machine now i'm definitely not just taking every problem in my life and typing it in to an llm and then doing what it says yeah ignoring what it says yeah man what a weird world i don't get it i still reject
Starting point is 00:00:58 i like having a hard time cognitively accepting that computers can do things that i never thought they'd be able to do i have never been more excited to be a software developer or software adjacent person it's so fun the problems it brings up are fun it's it's so cool yeah it is wild things that are possible finally i've uh okay i promise i won't monologue on this long but i i've spent the last whatever 20 years watching people use computers and just struggle to do basic things like oh i've got this file it's an excel spreadsheet but i need to put the table from excel inside of this like slideshow thing and make a pdf and then it's like can't do it you know and i'm like ai is making all that crap easy all these dumb things you know just so many
Starting point is 00:01:51 things i think i think we're like on the cusp of having computers actually be useful for everyone and you don't have to be quote computer literate or like a computer whiz you know to do basic things like oh yeah i can make a pdf for you or yeah i can answer that question or i can i can make a picture of someone riding a bike for you pictures of pelicans riding bikes market is cratered exactly oh i can make a trillion dollar sass company for you you're absolutely right that is a good idea and i will go build it yeah um dave should i thank our patrons please do thank you to these folks who contribute at such a reasonable yet admirable level that we shout them out every single week thank you to brian wolford is in germany old man yells at claude open parentheses
Starting point is 00:02:38 followed by an actual open parenthesis character joel jamie sundance which by alter ego i believe I'm going to say that's the bad one because I'm the good one wow Jamie Sundance that's great yeah go ahead and run your own mail server what are you chicken bok bok my actual name on LinkedIn is Yami debugging the dark candy the mobile development ordinary first of his name
Starting point is 00:03:00 quote or one equals one drop table meow mix Jacob Shandling Larry Ellison coffee is for closures the missing semicolon Christy the world's okayest programmer I apologize for the technical issues in last week's announcement the internet in the soft skillet nation is not the greatest
Starting point is 00:03:15 yeah you said words and now i hate my job nick molyneux embedded engineers treat assembly the same way typescript engineers treat javascript javier gonzalez chewy ted timbril i like chicken i like liver meow mix meow mix followed by a single closing and then an actual closing paren and then will angel and then another single closing paren good luck parsing that a separate a separate sponsor who is a single closing parenthesis character this is getting so good dan from drone to play never is not just a crater on mars flamingo emoji i like chicken i like liver meowix meowix please deliver swiss python okay i will measure your text bandwidth beautiful is better than ugly implicit explicit
Starting point is 00:03:54 is better than implicit is better i don't know why i feel like i have to read that one really fast yeah yeah i gotta get the bandwidth numbers up uh kyle boss can see dodds this podcast was recorded at 2026-02-10, 5.53 and seven seconds or later. That is correct. Jenny Kim, the Stochastic Parrot, Ira Chan,
Starting point is 00:04:15 Jonathan Kings, and I, beautiful functional user documentation. I don't understand the patrons in this list who don't change their name regularly. What are you even doing? Should Will Angel start a blog?
Starting point is 00:04:24 Vote now on your phones. Does that mean call him? How do you? Just call him and say yes. Just open up Gemini and tell it you vote for it. Ragnar, Brayden Cain. I vote yes. Ragnar, Brayden Cain, John Grant, Brittany Ellick,
Starting point is 00:04:41 and close parenthesis and then a single closing parenthesis. Thank you. Thank you. You are building the world's ugliest lisp, and I am excited to be a part of it. Thank you. Thank you for your contributions. You help keep the show going, help keep us happy and sane.
Starting point is 00:04:59 And if you want to join them, you can go to softskills.audio, click support us on Patreon. Any amount gets you an invite to our Slack team. Enough of an amount gets you a weekly or one-time shout-out. I never thought that the hardest part about hosting a podcast would be learning how to pronounce parentheses. Yeah, I feel like the word is going to lose all meaning. It's like if you just say Apple over and over and over again.
Starting point is 00:05:20 Yes. Linguistic saturation or whatever that's called. Oh, is that what it's called? That's cool. It has a name. Well, if you trust Ted Lasso. Sounds like a trustworthy name. Semantic satiation.
Starting point is 00:05:33 That's it. Does that also sound like a trustworthy name? Because that's the right name. And with that smooth transition, Dave, will you read our first question? Yes, I will. A listener named Derek asks, I am the CTO and co-founder of a startup. Now that vibe coding is a thing, our CEO has gone kind of rogue. And he's vibe coding a bunch of random stuff, one of which he bought a domain for and has pushed a potential customer to pay for without talking to our team.
Starting point is 00:06:00 i feel like this is fragmenting our focus but i don't want to ban our cv our ceo from vibe coding and being creative how should i handle this without damaging relationships oh what a cool world i don't know i'm sure some people look at this and say this is horrifying but i think it's so cool i assume the cto is not technical uh the ceo probably not or sorry ceo yeah hopefully the cto is technical so the cto is the one asking the question and probably reports to the ceo Yeah. Which is why I think the CTO feels a little bit slighted. It's kind of like, wait a minute, you went and built a product and sold it?
Starting point is 00:06:35 I thought building the product was my job. This is such a cool problem to have because part of the job of a startup is to learn, and you're experimenting, you're trying stuff. The feedback loop can be quite long, and it can be compressed enormously by lots of funky startup tricks, like you just put up a landing page and tell someone to pay for it and then go build it. But also, a new arsenal or a new weapon in your funky trick arsenal is just chucking LLM at it. Turns out you don't need to be an engineer to do that to some degree. It's cool.
Starting point is 00:07:08 I just think you like it because it's taking GPU time away from crypto scammers. That's true. But it also makes GPUs more expensive for video games. Oh, you're so torn. Yeah. What do I like more, playing video games or punishing crypto people? Yeah. yeah i think the 50 90s or what like 3k on ebay their retail price is like 2k no so we're it's
Starting point is 00:07:38 it's just as bad as deep in the crypto days anyways you have more ways to experiment now than you did before like that is a capacity you can purchase now yeah with tokens yeah now now you have the next bottleneck which is you need like judgment to learn from those experiments and it's fragmenting your focus absolutely it absolutely is fragmenting your focus and maybe that's right like maybe the new product idea is awesome and it's cool that you can experiment with that but i do think you need to have a conversation with your ceo about how to make this real and an analogy i like to use is free as in puppies like it's it's free to build this product you get a free puppy yeah exactly congratulations now puppies are great everybody loves puppies
Starting point is 00:08:28 and it it brings along a lot of work if you want to do that and the human like your cognitive capacity is not any larger so at some point you need to decide do we want to keep investing brain power in this or not and if we do that as a trade-off which is our other product will go slower so wait hold on the the humans are the problem in this scenario right if we just had puppies and machines taking care of the puppies everything would be great yeah i think so people talk about paperclip maximizers i want puppy maximizers yeah if you're going to destroy the universe to maximize something why not have it be puppies why not puppies yeah and they never grow up they're just stay puppies
Starting point is 00:09:12 AI is superhuman it figured out some benevolent way to keep them puppies and they're immortal and they don't poop no pooping no shedding and you don't have to bathe them otherwise it's a actually a poop maximizer because over the course of one puppy lifetime
Starting point is 00:09:30 they probably produce a lot more weight and poop than puppy yeah I think so that sounds right to me it's like the monkey's claw just ruined your metaphor yeah no i feel like i've experienced this a little bit with we've felt some increased productivity from llms we've used some of that to invest in new product directions and it turns out it doesn't make maintaining uh we're spread more thin now we've got extra extra capacity to do stuff but the decision making capacity and the judgment and
Starting point is 00:10:03 like cognitive capacity to deeply understand stuff is the same so it is a trade-off i think depends a lot on the stage of your company and it's okay to make this trade-off as long as you are explicit about it and it's bad to get okay worst is if your ceo says no it's fine like i'll just vibe code it and this is just a net positive don't worry about it it's not going to fragment our focus because i'll just work on it it's fine don't worry about it hopefully your ceo is not saying that but it's possible that the ai hype has has made him downplay the costs i guess yeah and i'll tell you like as a cto i think um one of the understated job responsibilities of a cto is managing a relationship with a ceo and most people don't become ceos because of personality
Starting point is 00:10:53 traits such as compliance easy to work with yeah it does whatever you say yeah it doesn't rock the Like these are not CEO attributes generally. And so you've got to know what you're getting into as a CTO. And part of it is working with someone who fundamentally might see the world a little differently than you do. Especially if you came up kind of through the traditional engineering ranks where you are very analytical. You're a little bit more risk averse than the typical, maybe a lot more risk averse than the typical CEO. Maybe you're even more introverted. And so here you have a CEO who's like thinking big.
Starting point is 00:11:29 And that's exactly what you want your CEO to do. identifying new opportunities i don't know vibe coding a bunch of random stuff like i don't know why this is such a bad idea if you get a potential customer who like gives you strong market signal that they're going to pay good money for something maybe your team should pivot to work on that product anyway so i don't know i'm kind of with you jameson like let this ceo run wild and maybe they'll tap into a gold mine that is going to make you very happy yeah i feel like this is fragmenting our focus i don't want to ban our ceo i think it is fair to call that out and to say this is the trade-off we're making i think it's it could be a good trade-off but i i want to
Starting point is 00:12:08 recognize it is a trade-off are you okay with the costs right yeah basically yeah yeah that's i think that's what i was getting at before if your ceo thinks there's zero cost to this then they're wrong and you could be led to believe that from uh some some loud voices but i don't know it's their company, right? Yeah, exactly. Exactly. Also, I think there is a tendency among engineers and even CTOs to think of their CEO as someone who is not thoughtful, just a bull in a china shop, who just runs roughshod over everything and is always like super distracted. And while there is an element of truth to that in many cases for CEOs, I think that a lot of times they are more thoughtful and more careful than we give them credit for. And don't underestimate the amount
Starting point is 00:12:55 of burden that your CEO is carrying. This is a startup, so your CEO probably took a whole bunch of venture money, probably gets a ton of pressure from the board of directors regularly and from their investors who may be one in the same, and feels an intense obligation to deliver value and get the investors paid, make sure that all the team members get their paychecks and they clear the bank every week or every whatever. And don't underestimate that. Try to empathize a little bit with the CEO because they could be under an immense amount of pressure. So to see them reaching for new ideas, this could be really great. And I think when you kind of humanize the CEO, so it's not like they're just this erratic machine that just randomly lashes out
Starting point is 00:13:39 at you and makes your job harder every once in a while for no reason, then you can humanize them as a person who has these concerns like I was talking about. It might help you understand them better. There's another problem to be concerned about, which is it can be really addictive to build stuff. And depending on the stage of your company, it's very possible your CEO is avoiding critical stuff that nobody else could do around operating and running a business because they're cranking out things in a way that maybe they've always been jealous of engineers' ability to build stuff and now they can finally do it right so i make sure that the lights are still being kept on but i think if if if the ceo is bringing some excitement and if you can be excited with
Starting point is 00:14:27 them and say here's the trade-offs we have in front of us and here's how i think we can make this more mature and operational instead of stop doing that because it's scary yeah i think you can have a a good partnership yeah and then maybe you need to say something your ceo like hey i'm gonna want to operationalize this like or sorry my your job is to come up with crazy new ideas my job is to help operationalize those ideas so go do all the vibe coding you want and do all the make all the promises you want but the moment we need to operationalize something let's talk you know and then it's kind of like it's not going to be clouding your delivery objectives constantly because your ceo might be like look i don't care if you ever turn this into a workable product i
Starting point is 00:15:07 just want to test the market and that's my job as a ceo just keep my finger on the pulse of the market. So don't tell me I can't do that, you know, but then you need like a clear line in the sand. And I've talked to my own CEO about this several times where I'm like, look, my CEO is actually an extremely gifted idea generator and loves to talk about ideas and can compute hundreds of ideas in like a few moments, it seems. And I'm like, I suddenly feel all this pressure to operationalize all those ideas into working product. And I just, I can't, like, I don't have the resources to do that. And then I self-impose this pressure, like I'm going to disappoint my CEO and he's going to fire me because I'm not doing all his ideas. But in the reality, he never
Starting point is 00:15:49 expected me to do all those ideas. He's thinking on like a 10-year horizon. And I'm thinking like, what am I going to do next sprint? And that mismatch creates tension. And you can resolve that tension by having rules of engagement around like, hey, make sure we talk and have like a clear vocabulary for when this is a must-do-now situation. Yeah. Have you answered the question? I think so. Good luck. Natural tension. It is normal. And I think with some good ground rules, you can have a great working relationship with your CEO. Isn't it great to have somebody who's excited and cares? I mean, they care. That's cool. Nice to work with people who care. All right, Dave, should I read our next question? Let's do it. This is from a listener
Starting point is 00:16:31 named Admiral Fox, who asks, Hi, Dave and Jameson, long-time listener, first-time question asker here. After 14 years at a consultancy firm, Dave, I read the number 14 years, and I thought, that sounds like a really long time. And then I realized, we started this podcast like 10 years ago.
Starting point is 00:16:46 Yeah, we did. Which is also a really long time. Apparently not that long, because I thought it was like yesterday. Yeah, that's how it feels anyway. Yeah, time flies. Anyways, okay. After 14 years at a consultancy firm,
Starting point is 00:16:57 I'm moving to a major retailer to become their java learning and community lead instead of shipping code my new role will be shipping knowledge i will be managing learning paths organizing internal knowledge sharing events and helping managers screen candidates basically i'm moving from a maker role to a multiplier role i have 13 weeks of notice period standard european i'm not leaving yet after 14 years protocol and i want to use my free evenings to prepare my questions for you how do i transition from the guy with the technical answers to the guy who helps
Starting point is 00:17:30 everyone else find the answers. How can I use the remaining time of my notice period to prepare for the people side of this role? Love the show. Keep up the quit your job advice, although I've already taken it.
Starting point is 00:17:42 Oh, nice. Nice. We often have people write in and say, I want to ask you a question, but you can't say quit your job. And usually we're like, well, quit your job anyway.
Starting point is 00:17:51 But this one is like, I already did. Yeah. You found the way to how to stifle us. unless quit this next job have you ever been in like a training role as your main responsibility uh no never as my main responsibility it's always been like a secondary thing for a while i was a freelancer consultant contractor whatever word sounded fanciest and one of the things i did was
Starting point is 00:18:20 some training and i only did it once and i did not like it boy did i not like it what did you not like about it what did i not like about it i didn't feel like i was good enough to be worth the money i i felt like i was excited about the technology and it had some cool attributes and aspects but it was a pretty niche technology it wasn't like a major mainstream thing okay and i think relative to them i knew a lot more but i wasn't like the world's expert on it huh so you had imposter syndrome i think i had a little i don't know or i was an imposter and teaching is just hard for a long period of time you don't get great reactions for eight hours a day for seven days in a week and yeah yeah it can just be it just felt like a slog yeah for sure i don't
Starting point is 00:19:07 know so i i admire people who can do this well it's uh it's a skill i guess that's my first point is uh recognize that it is a skill and and you're already in a good spot because you know a ton about it and presumably you've taught other people things about java in the past as just a smart java person but you're getting a new job you might be kind of bad at this job even though you're good at some parts of it yeah and that's actually to me one of the interesting parts about this is that you can this is like a complete reset and you're going to have a lot of opportunities to learn new things yourself even though your job is to teach apparently but learning how to teach is going to be pretty interesting. And I think probably one of the keys to success is setting
Starting point is 00:19:48 up feedback loops so that you can quickly learn when things work or don't work. And this could look like surveys. This could look like follow-up tests. This could look like metrics in adoption of technologies that you educate the company about. And I just got to assume this is a gigantic company. So all these things should be pretty natural to do. I was looking up how to pronounce the word pedagogy, and apparently I've been saying it wrong my whole life. What is the right way? did you just say it right pedagogy i don't know pedagogy pedagogy yeah that's how i say it the first the google pronunciation thing says pedagogy pedagogy well i guess you're not the world's expert on how to pronounce that i'm not but now you are you do learn an enormous amount when you
Starting point is 00:20:34 teach if you think you are a java expert now you'll you'll be even more one after this but then you do get to learn a whole stack of skills around pedagogy oh yeah which i think is exciting yeah i guess i'm asking questions that you can't answer but i i would try to find conferences or meetups or i would try to find groups of engineers to speak to as much as you can and maybe you're doing this already maybe not but i would practice teaching people stuff for now put together some some Java presentations? Is there a local Java users group? Are there conferences that you can go to or local or more regional ones? And that's sort of curriculum prep and public speaking a little bit and just sort of polishing your skills. I have one more suggestion for someone whose job
Starting point is 00:21:24 is to do technology education, especially with the target audience of software engineers. I think that I am not alone in sharing a deep skepticism of people who have turned their careers into teaching but have not shipped any product in a long time. Oh, absolutely. Yeah. And I think I was trying to think, like, how could you mitigate that negative perception? and I think the answer is you probably need to find a way to incorporate shipping software into your at least not maybe not daily work but at least into your regular work and maybe that's something where you go hey listen team I'm going to embed with an engineering team a couple weeks
Starting point is 00:22:09 out of every quarter I'm going to I'm going to ship working software I'm going to be in the trenches and then I'm going to surface for the other 12 weeks of the quarter or 10 weeks of the quarter, whatever. And I'm going to share direct learnings, not just informed by my amazing time in the ivory tower, but also informed by my time with you. And, you know, that could take a lot of forms. Like you could ship directly by making contributions yourself. You could also spend a lot of time interviewing teams and talking to them and taking them out to lunch and learning what are their struggles. Because one big risk on doing a teaching gig like this is that you're going to share a bunch of information
Starting point is 00:22:49 that either A, everyone already knows and then you just look like out of touch or B, that's not useful or not applicable to their particular roles because you don't know how they work. It's like, oh, I'm going to teach you all about annotations in Java and how you can make really great annotations.
Starting point is 00:23:07 Come to find out that like the head architects of this company banned the use of annotations a couple of years ago and you're like, ah, crap, you know? Yeah, I love that idea. I also think it'll help to have more grounded examples from code at your company. I feel like there's always some toy project or kind of pet code sample you work through. And it's always small because that's one of the constraints you have to have when you're teaching.
Starting point is 00:23:34 You have to be able to comprehend the whole chunk of code at once. That's not most of the code we work with. So you have a great real-world laboratory to pull gnarly bits of code from and show, here's how i would do this differently or here's how i don't know so i so i would solve this problem that we actually have and that your team knows about that kind of thing yeah totally and i'm gonna just assume here that given the sizes it's a major retailer and the fact that they have enough people to warrant hiring a java learning and community lead tells me that they probably also have a whole bunch of weird custom stuff and i worked at a major retailer some might say
Starting point is 00:24:15 the majorist well actually you know what we both worked at really major retailers didn't we the major retailers yeah the two major retailers and i gotta tell you there was a lot of weird custom stuff and it was like if you've never worked with this weird custom stuff and then you go and try to give some like vanilla java thing like java education program to the team there it would have been an absolute dumpster fire for all the reasons i said earlier like just not applicable or banned technology or whatever um so i do think you've got to come up to speed on whatever internal stuff they've got going you better become a master of it because guarantee they've got some of it i don't have a solution for this but i do have a problem which is it's easier to point to a software
Starting point is 00:24:57 engineering team and say look they build this thing what do they do for us we pay the money they make this thing you are abstracted a level away from that and in a different shape than someone like a manager or project manager or something where like i i support them in building the thing and i believe that there's there's efficiencies and improvements to be had from engineers getting better at this programming language and tech stack but i think you have to be able to demonstrate value in some way i don't know what that way is you could measure satisfaction with your trainings but that's approximate that's not actual value the dream is you can show because i taught this team they did this thing better and had this outcome because
Starting point is 00:25:43 if i look at the business at like the spreadsheet level as expenses and outcomes and stuff i feel like a uh community lead is a fuzzy role that i uh in the right financial circumstances i might squint and say what does that do for us do we really need that so um i'm assuming because they're hiring you and because they're a major retailer they're not like about to can this position but it's still just uh when times get tough these positions go i have seen that for sure that is that is true unless yeah that's probably true and you better be able to demonstrate that when times gets get tough you better keep this position to make sure you get like maximum value the guy who helps everyone else find the answers i wonder how much of this is going to be
Starting point is 00:26:33 being an expert in the code base of the company as in as opposed i mean i know you talked about they have to ship stuff to be grounded in reality but also like how much of their time is is just code-based level questions is that even a thing that would be useful i don't know probably not if the company's big enough then you probably can't learn enough of the code base to be useful at answering questions like that yeah people i mean you might be able to become an expert on some of the internal weird stuff though that's shared across lots of teams and you're sort of like a you're like a messenger from a faraway place like i come bringing news from the land where we use Lambdas all the time or whatever.
Starting point is 00:27:16 Yeah, that's a good point because you've had a bunch of years as a consultant, which means you've probably got a broad cross-section of exposure to lots and lots of different companies. And so you'll be able to say things that the people who have been at this retailer forever simply don't know and have no way of knowing.
Starting point is 00:27:31 And so that could help on making things, making important decisions about what technology to use, what patterns to apply, and how the company should operate going forward. And that could be a really valuable way to contribute and not be perceived as just, oh, the training guy. Yeah. This is an interesting role. I have never worked somewhere.
Starting point is 00:27:52 Maybe I have, and I just didn't know. I've never worked with an in-house, like, technical trainer like this. Yeah, me neither. It would be very interesting that you could do some cool stuff. Have we answered the question? I think so. Best of luck to you. Sounds like it'll be fun.
Starting point is 00:28:10 I think so. What can people do if they want their own questions answered? Go over to softskills.audio and click the Ask a Question button where you can fill out our form and ask us whatever the heck you want, anything. Even if you try to trick us into not telling you to quit your job, we will accept it. There's no validator on the form that stops that from coming in, and so you can submit it.
Starting point is 00:28:28 And we want to thank the many, many people who every week submit the coolest questions and the rest of you. Just kidding. Everyone who submits a question. everyone who submits a question we thank you so much you keep the show going without your questions we would just be staring at each other we would just answer we we would just we just start answering we wouldn't have a question we'd be lost yeah yeah just like tell an llm please answer any question okay what happens if you just give it white space as input over and over and
Starting point is 00:29:02 over again yeah probably more context spirals somehow oh that's a great idea all right thank you for listening. We will catch you next week.

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.