Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

Code, Cloud & AI

Code, Cloud & AI artwork

Human in the Loop: Tim Warner on AI, Coding, and the Future of Tech Work

Jul 28, 202636 minAudio

In this episode, I sat down with Tim Warner to talk about how generative AI is reshaping the day-to-day reality of technical work. We compared this moment to earlier platform shifts like the web, mobile, cloud, and virtualization, and dug into what feels different this time: the speed, the accessibility, and the way AI has moved from being a novelty to becoming part of the operating system for how many of us think, learn, and build. Tim shared how he uses tools like Claude Code, ChatGPT, and GitHub Copilot in real development workflows, why he still insists on keeping a human in the loop, and how prompting, context, sub-agents, and memory all change the way he approaches building software and teaching others to do the same. We also explored how AI is affecting infrastructure work, instructional design, and technical education, including how Tim uses it to improve Pluralsight course demos and certification content for Kubernetes. Toward the end, we zoomed out and talked about careers, kids, and the future of software development: if code generation becomes increasingly automated, what should people actually study, and what work will still matter? Tim’s answer was refreshingly human: focus on the intersection of passion and aptitude, and use AI as a force multiplier rather than a replacement for judgment, curiosity, and craft. Tim Warner: https://www.pluralsight.com/authors/tim-warner Barry Luijbregts: https://www.pluralsight.com/authors/barry-luijbregts Pluralsight: https://www.pluralsight.com/ Cornell University: https://www.cornell.edu/ Microsoft: https://www.microsoft.com/ Claude Code: https://www.anthropic.com/product/claude-code Claude Code for VS Code: https://code.claude.com/docs/en/vs-code GitHub Copilot: https://github.com/features/copilot Visual Studio Code': https://code.visualstudio.com/ Windows Terminal: https://learn.microsoft.com/en-us/windows/terminal/ ChatGPT: https://chatgpt.com/ OpenAI: https://openai.com/ OpenAI Codex CLI' https://github.com/openai/codex Gemini CLI: https://github.com/google-gemini/gemini-cli Everything Claude Code: https://github.com/worldflowai/everything-claude-code Getting Things Done (GTD): https://gettingthingsdone.com/ David Allen / GTD About: https://gettingthingsdone.com/about/ Microsoft Bicep: https://learn.microsoft.com/en-us/azure/azure-resource-manager/bicep/ Kubernetes: https://kubernetes.io/ Certified Kubernetes Administrator (CKA): https://www.cncf.io/training/certification/cka/ Camtasia: https://www.techsmith.com/camtasia/ TechSmith: https://www.techsmith.com/ ScreenFlow: https://www.telestream.net/screenflow/overview.htm Adobe Premiere Pro: https://www.adobe.com/products/premiere.html

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

Barry LuijbregtsHere we go. Yeah. So

Tim Warnerhey, Tim. How are you doing? Hey, Barry. I'm doing great. How are you?

Barry LuijbregtsYeah. I'm doing I'm doing very well. I've spent my day today. It's fifteen thirty, like we just said before we started the recording. My day is kinda almost done. I spend all day basically talking to my computer. Like to an AI, to various AIs. It is very strange how my days have changed from, like, browsing and looking stuff up to just talking to my computer, and it just talks back, basically. It's very strange.

Tim WarnerBut, yeah. How are you? It's a different world, isn't it? I'm doing great. Thanks. I really am. And I agree a 100%. First of all, I wanna say it's awesome that you're using voice chat with the AI in your computer. I tell that to my learners all the time. Some people are better with their prompting using their voice as opposed to writing. Hit that microphone button and and get her done. I use voice chat when I'm on the go a lot. But to your point, Barry, yeah, it's it's rapidly

it's a AI has has rapidly increased my knowledge, my effectiveness personally and professionally. It's it's an amazing time to to be alive.

Barry LuijbregtsIt it is. Yeah. And it's an amazing time to just be in tech as well. Yeah. Because, you know, this is the I don't know how many of of these big changes. Like, had mobile. We had the web. We had cloud. I've I've caught a couple of those now. The web I'm I'm a bit too young for. That was already there, kinda ish. So when I started using computers, we had the web. I feel like this thing is way bigger than that. I don't know. What do you think?

Tim WarnerWell, it's funny you mentioned that because that's something, again, whether I'm in a consulting or teaching context, I always like to bring up is when you think about the big disruptions in tech, and I am a little bit older. Actually, I'm wearing a a shirt of my alma mater. I was at Cornell in the mid nineties at the dawn of the World Wide Web, so I was right there in the center of it. I think that Gen AI has been following a very similar trajectory to all those other disruptions. Yes. There's a change in in economies

and job roles and and all of that, though there's definitely truth to that. It's not one of those things that's rapidly gonna fizzle out. It's it's like the mobile device revolution. It's like the World Wide Web. It's gonna be with us. And I I find that I see a lot of the same reactions to it that I've seen in the past, like when the cloud came in. I I still have a lot of friends and professional colleagues who are really skeptical and angry about the cloud because they prefer to have a screwdriver in their hand in their local data center, and that's cool.

But, fortunately, I've always been one who's willing to change and evolve, so it's been more enjoyable than painful.

Barry LuijbregtsYeah. Yeah. I find same same thing. And indeed, also from a Microsoft perspective where I work now, I do still see customers that are, you know, half in the cloud. Some are still like, hey, the cloud. Is that a new thing? I don't know. Right. But every everybody does it at their own pace. Right? So and and that's fine. That's totally fine. Have you been developing software yourself still?

Tim WarnerYeah. Absolutely. Yeah. More than ever before. Again, that's that's another level up with Gen AI. It's really twofold. One, being able to take an idea and push it down the field faster than than ever was conceivable prior to generative AI. And two, I feel my effectiveness as a developer has is far farther along faster than ever before because I make sure to have the AI teach as it goes.

I'm not one of those stereotypical vibe coders who just wants the AI to run off and and do all of its magic while I get my espresso. I'm gonna get my espresso. Don't worry about it. But I actually wanna know what's happening behind the scenes. And, again, I think that's a critical through line to this discussion. Human in the loop always.

Barry LuijbregtsYeah. And so you develop a production software as well for clients?

Tim WarnerNo. I would call myself mainly a hobbyist. I don't I've in my career of almost thirty years, I've worked with and on software engineering teams, but I I have yet to have developer as a formal full time job role. So I don't feel I would be doing clients and customers a service doing that. More more architecture, and and the stuff that I develop and ship is mainly just for my learners and my tech community. I do submit work to my full time employer, Pluralsight, again, and then automation

more than anything else, helpers, that kind of stuff. And in my, instructional design, of course, I need demos. I need sample code. And and I guess last of all, I have a 16 year old daughter, and we've always got a project going on. So there's something to do development wise in some stripe or another.

Barry LuijbregtsOh, that's fun. So she also uses it?

Tim WarnerActually, not. It's an interesting thing. Here here's my daughter. She's got a dad who's all into AI. She's very opposed to AI. She's an artist, and she's concerned about intellectual property rights as she should be. Yeah. And she's an animator, and so she resents AI. And, actually, intentionally, when we're doing like, we're working on a it's a pretty low tech, interactive web application. And, she quizzes me to make sure I'm not using AI. I mean, in full disclosure, I, of course, I'm gonna use it a little bit. For her I I had tell her that. But for her, we're kicking it old school. Fortunately, it's just light JavaScript,

mostly just old school HTML and CSS. So

Barry LuijbregtsOkay. Well, keeping it simple. That's how I like to do it as well. Yeah. Okay. So when you do use AI for, creating software, what do you use? What type of Copilot or thing do you use? Mhmm.

Tim WarnerAt this point, I'd say my 80% scenario is Claude code and the Claude Versus code extension. I I found of the LLMs, the Claude models not only are the most human like, but they they think like a developer more than any of the others. Now, it's ebbed and flowed with, my use of GitHub Copilot. Now that it supports multimodal, the barriers that might have been there in the past are no longer there. But for me, it comes down to effectiveness

first, but customer delight right behind it. And my complaint with GitHub Copilot and Versus Code, Versus Code is my code editor, is that those engineers keep making changes, particularly UI and and feature changes and renames way too often, and and it's anti delightful. So that's why I found Cloud Code to be the, the best fit for me currently, not only because of its effectiveness, its ability to think like a programmer, but just from the pure bucket of customer delight using the software as a delightful experience for me.

Barry LuijbregtsYeah. I I totally get that. Like, they move your cheese a lot. There you go. Yeah. Because things are so you know, they're developing very actively on it and new features drop every time and then, you know, maybe something changes. I I totally get that. So so you only use it from Versus Code, or do you also use the CLI?

Tim WarnerOh, right. Yeah. I guess I should have been more specific. Mostly from the CLI. Yeah. Yeah. Either in Windows terminal or the integrated terminal in Versus Code. Yeah. And that's something else. You can tell when when a vendor has a good idea when all the competition starts adopting it. I mean, all of the paradigms we have in Claude code from the slash commands to context visualization to sub agents, now we see going from the Google Gemini CLI to the GitHub Copilot CLI to Codec CLI,

they all behave largely the same, and that's something I like to stress in the classroom, that don't worry about having to reinvent the wheel, you know, the dry principle in development. These vendors are all cribbing on each other, so learn one really well, and you're 80% there for mastering the competition if and when you wanna go there.

Barry LuijbregtsYeah. Because they all basically work the same. Right? Mhmm. And so so how what I used to do is I used to try slash commands and ads and all sorts of things to try to make it work better simply because, I don't know, six months ago, they weren't that great. Mhmm. Now, when I just type a sloppy prompt, like, hey, make this for me, it just usually basically does it without much hand holding. Mhmm. And how do you use those things? Do you use, like, skills and and Oh, yeah. And stuff like that? Yeah. Absolutely. In fact, not to be skewing this, you know, anthropic wise

Tim Warnerspecifically, but I wanna give a shout out to a GitHub project called Everything Claude Code. I don't remember the dev's name, but this person really has gone to town creating a plug in that includes very useful agents and hooks, all the different parts and pieces that some of which that you were mentioning, Barry. And it might be a good spot for someone to look at and experiment to see and test how these various doodads and thingamabobs actually work. To answer your question,

it's probably my bias because I've been a professional writer for so long, but I I really, try to be as explicit as possible in my prompts. I do use the different triggers, like you said, the at in Claude code to tag a file. I I I don't generally trust the AI's ability to do inference. I've been with it since the beginning, and although it's gotten so much better to your point, I still take the Python approach of explicit is generally better than implicit

or belt and suspenders, insert cliche here. But, yeah, I I I will, I'll be pretty darn imperative in my prompts strictly to, try to reduce and hopefully eliminate the inference space of the AI because I don't want it making anything up. None of us do.

Barry LuijbregtsYeah. Yeah. Exactly. And I think as these models, become better, the need for that becomes maybe less.

Tim WarnerProbably. I I admit. I'm gonna have to evolve my prompting approach. Right now, I'm being in a curmudgeon, I guess. I'm gonna have to evolve that to survive, and that's fine. You're probably right. Who knows? In the future, maybe the AI will be able to work in conjunction with a neural implant, and we can it'll just follow our thoughts.

Barry LuijbregtsOh, I don't know do not want that.

Tim WarnerOr do I. But you know what? All the Black Mirror stuff, it's gradually turning into reality.

Barry LuijbregtsIt is. Yeah. It is. Truly is. And so, you know, you could do so much stuff now with it. Like, you can hand off a session to your phone, for instance. Do you do that and then, like, go on a holiday and then you're still working?

Tim WarnerYeah. A lot, actually. Not yeah. Actually, even for development projects. Just last night, in fact, I was driving. I had to run a couple errands, and I, I had a conversation in ChatGPT on my desktop specking out a a software project for for a course I'm teaching, and I picked it up on the road using ChatGPT voice mode, and it was super useful because we were able to explore stuff that I wouldn't have thought about. You know, basically,

require and that front end kind of stuff, front end proactive optimizations and stuff. And I was like, let's let's brainstorm and get this tight. And then when I got home, I got onto my computer and I said, whip up an excellent PRD document that I can hand to my coding assistant. And that's exactly what I did. So, yeah, that handoff is super, super helpful for sure.

Barry LuijbregtsYeah. It is. But for for me, it it also means that now I'm just doing stuff all the time. And I'm like I have, like, I don't know, 10 projects in my head. And I'm constantly like, oh, do this. Do that. Oh, maybe we can do this. Maybe we can do that. And so it's a lot of context that I hold myself, although the work gets done by all these robots, basically. But I don't know. I get a bit overwhelmed by all this stuff because now I'm I'm taking on too much because I can do everything. It's weird.

Tim WarnerNo. No. I I understand, and I I relate to that. It's very real. For me, I I think I would be lost if I didn't have a productivity methodology. I practice GTD, getting things done. And the whole principle there is to get stuff out of my brain and get it into a trusted capture system. And you know what? To your point, you could have all of those project stubs in your capture system and feel very overwhelmed indeed. And I've been there, done that, and bought the T shirt.

I can think of an example where Gen AI has helped me with that. I've I've pointed ChatGPT. Look at my projects. Look at my past conversation history. Find these unopened loops. Let's take a look at them. Give me your best guidance on which have some teeth to them, which should I maybe deprioritize or let go out of my brain. So using AI in that meta sense has really given me some some peace.

Barry LuijbregtsOh, that's a good one. I I'll try that as well. You know, to get some focus. Yep. That's, that's important. Mhmm. So so you can use it for software development. I know I also talked to platform teams and infrastructure people. They're like, yeah, we have GitHub Copilot for instance, but you know we just ride our bicep all by ourselves. Do you sometimes, have conversations with platform people and and info folks? And and do you find that they use, AI?

Tim WarnerYes. I have those conversations all the time, I'm fortunate to to say. And what I've seen as a trend is quite a few engineers, either by directive, you have to use it, or out of their curiosity, tried it out maybe in the earlier days when it was not very effective in many cases, and have a net negative and now a resistant attitude towards it. So it's an it involves the conversation of giving it a try now that we've seen this delta

from how it used to be even six months a year ago, and then b, having them feel enough return on investment to go a bit further and to learn the slash commands, the invocation commands, the shortcuts. And if and when they cross that bridge, yeah, you've probably got a a convinced developer at that point, and that's a good thing. But I I have seen a lot of v one after effects of devs who felt burned, cheated, afraid

of Gen AI initially, and now are super skeptical of it, and they haven't yet seen the improvements of which there are many and continue to be.

Barry LuijbregtsYeah. As in, I tried it with, GPT 3.5, and it sucked. So I'm not trying that again. Yeah.

Tim WarnerExactly. And you read the scary tech news of businesses having to uphold hallucinated read refunds, you know, warranty stuff because their AI chatbots hallucinated, or companies that went way too far and downsized their human engineering teams, replaced them with agents and got burned that way and then frantically tried to rehire, that those experiences don't help either. But I think, again, in in how we started the conversation,

this stuff happened and happens with cloud, mobile, virtualization, containerization, World Wide Web. I mean, I can go back to the seventies in elementary school when my math teacher gave let us use calculators. There was a big furor and uproar. The kids aren't gonna know how to do math because they're relying on these computers. It's all of a thing. It's amazing.

Barry LuijbregtsYeah. And now you mentioned it. I kinda forgot about that. During the cloud era, people were like, alright. All of the or we're gonna make lots of cuts to the the infrastructure people because, you know, we don't need them anymore. We just run stuff in the cloud. But, that turned out not to be true.

Tim WarnerYep. Exactly. Yeah. Yeah. I think AI is a is a wonderful, force multiplier, but I know it sounds like I'm standing on a soapbox or something saying human in the loop, but I I really believe that. When people in my life who aren't in tech ask my opinion of AI, First, they wonder about the sentience sentience or consciousness. I say, look. It's software running on a computer. Unplug the computer and the AI goes bye bye.

And number two, it's not the AI to be afraid of. It's the people who are governing that AI that we wanna be concerned about. You know? It's the operators. It's the humans.

Barry LuijbregtsYeah. And that's kinda always, the crux, isn't it? Mhmm. It's always the humans. Yeah. So keep the humans in the loop. So where do you think this is going with, you say keep the humans in the loop? Will companies still have humans, or will it just all be agents agents sending emails to other agents, calling agents?

Tim WarnerWell, already with software development, the question of writing code from scratch is almost a moot point because these agents, these GenAI products can write code so much faster, in many cases, securely, more efficiency more efficiently. But in my opinion so in other words, the the role of the software developer becomes more of an architect or a steward, a a guider, an operator,

for lack of a better term. So that that's one direction that AI seems to clearly be going. Because to your point, with some sub agents, you can, you know, fan out and delegate this fleet of virtual assistants. I want you to proactively look for bugs here. I want you to proactively look for vulnerabilities, security issues. I want you to follow behind and test. That's another important point, making sure that whatever you're having your agents do, you've got another round of agents behind them validating.

And then you've got an agent. I want you to keep the docs updated. Yay, docs. And update your your own memory file. That that's an issue that some people forget about. Make sure that you have the agent update its memory so it has a current snapshot of your code base. And the next time you start a chat, the AI is current and it's got that loaded and and ready to rock.

Barry LuijbregtsYeah. So so with all these agents, you think we're gonna have, like, a a lot less, for instance, software developers or other people?

Tim WarnerIt seems likely, unfortunately. And and, again, I think back in my life, I remember in the late eighties, I grew up in in Upstate New York with a lot of manufacturing, particularly auto manufacturing. And I remember how disruptive that was because once assembly line robots turned up, there was literally less need to have humans on the assembly line because the robots never sleep, and they do a deterministic operation twenty four seven, three sixty five.

So, I mean, there's a kernel of truth in the fear that I experience with developers in my life about is this Gen AI stuff gonna put me out of business? And I I would encourage back the guidance. It's about getting on top of the AI and learning how to use it to make yourself more effective and make, yourself a force multiplier. It's using AI. Because like I said, there's gonna be a need for those people to operate these agents going forward.

Will there be, as many people necessarily? Who knows? I mean, maybe if the cost of hardware goes down and we start getting closer to edge and on premises LLMs Yep. Then you're gonna need operators in those contexts as well. So, yeah, I I think it's just gonna be a case of increasing specialization.

Barry LuijbregtsYeah. It's it's nobody knows. Yeah. But Yeah. I heard somebody say on a podcast for developers, for instance, if you're a developer, your task is not to develop code. Those are tasks that you do. Your task is to create software or to create something that creates business value for your customer. Automate a business process or something like that. And then part of that is tasks that you do, is creating software. And now you can give those tasks to an agent or to Claude or whatever. And it can do that, but you still have all these other things to

form that into something that automates the business process, which actually provides a value for now, at least. Mhmm. So, you know, that that's a good way to think about it, I think, what this guy said on a podcast. But let's see if that holds true in a year. Yeah.

Tim WarnerYeah. For sure. And going further, the idea of an LLM running in, say, a robot. I don't have any experience with these humanoid robots that are under development now, but I have little doubt that the day of the domestic robot is in our future. And I kind of think that I I I'm I'm a fan of that myself. I think it's pretty cool. Again, you could take these things in any direction, and it would be the humans behind them that are doing the operating that are thinking of nefarious things. But I'm thinking of stuff like folding my clothes for me. You

know? Again, delegating tasks and freeing me up to use the full capacity of my brain to ideate and to add value and to help and to be kind and do all of the human stuff and delegate scut work and tasks to AI agents.

Barry LuijbregtsYeah. Yeah. Definitely. Embodied physical AI that's absolutely coming. Mhmm. Mhmm. Yeah. And then let's see how that goes. Like, would you trust that with your daughter? Like a a humanoid AI?

Tim WarnerYeah. Exactly.

Barry LuijbregtsExactly. I don't know. Maybe maybe after a few years. I don't know. Mhmm.

Tim WarnerYeah. Exactly. I mean, the v one issue is pretty universal, isn't it? Yeah. You you you need to iron out the warts and and get relatively stable. Okay.

Barry LuijbregtsYeah. So the main, thing that you produce now for Pluralsight is courses. Right? So educational courses about all sorts of things. Like I'm looking at your Pluralsight author page here and it says currently 179 courses and I'm sure there's a bunch of them that have been retired as well. Right. That's a lot a lot of material that you have been that you have created and you still continue to create, which is incredible because that means people can learn and then they can scale

what what they do. I think that's great. So how does using AI change how you teach? Mhmm.

Tim WarnerWell, the rapid development has been a watershed. I mean, for example, the courses that I'm doing now are on Kubernetes, particularly the certified Kubernetes administrator certification. And that's a lab exam, so it's entirely practical. And creating demos is something I love doing as an instructional designer. It's it's actually more delightful now because I'm able to include the AI and the planning of these demos, and it consistently gives me, I think, banger ideas that I might have taken a lot longer to think about on my own. Or maybe in some cases, wouldn't have thought about them

across the board. So the quality of my sample code, the quality of my demos has definitely increased, and it's not a question of complexity for the sake of complexity. Why? Because I put that down in my guardrails, human in the loop. You know? But the quality is is, again, it's it's really nice. I I tend to work quickly anyway. I'm I'm fortunate in in that I can learn pretty quickly. And and generally speaking, knock on wood, I'm able to avoid the analysis paralysis

that hits some fellow authors that are in my life where you get hung up on perfection. I'm so grateful that that I've been able to sidestep that so far. So, yeah, it just potentially makes me even faster, and and to do these things in parallel with AI is really helpful.

Barry LuijbregtsYeah. And do you also use it in course production? So, I don't know, voice cloning, video editing, stuff like that. Mhmm.

Tim WarnerI've experimented with that stuff. Just the other day, I got turned on to a GitHub project that offers an MCP interface into Camtasia. Camtasia's been my video editor of choice for couple dozen years now. And I thought, wow. That's cool. I could turn the agent towards my raw video and and see what it can do. So that's on my bench to take a look at, and I'm happy to share that project with you after we're done recording, Barry, if you're interested. Out of curiosity,

are you still on Camtasia, or do you use ScreenFlow or Premiere? What do you use for No. Still Camtasia. Nice. Still Camtasia. Okay. Yeah. And and it it has improved a lot over the last years. Oh, yeah. Exactly. There was a period of time where it felt like the camp the TechSmith engineering teams were not only ignoring feedback, but actively bugging their software. I know that sounds critical. But they've got something good, a lot better going on now in terms of transparency and effectiveness.

But to get back to your original question, no. I mean, I I I experimented with the voice cloning a little bit, but I'm I'm really not comfortable with that because I think some of my best value as an instructor is, I I inject my passion and personality into my training. And until recently, I never used transcripts. Now I I am using transcripts a lot more, not during demos, but during the PowerPoint stuff. I hope that I worry a little bit, frankly, that I'm coming across more robotic. I always

prided myself thinking, well, if I'm not using a transcript, I'm it's like a live classroom, so there's a built in spontaneity there. And that's true. And a built in energy, and you run the risk of missing stuff. Now I I use the transcripts because it's easier. I'm not missing stuff, but I worry about the human factor. And that's also why I'm not a fan of voice cloning or video cloning in my instructional stuff at this point.

Barry LuijbregtsYeah. Agree. I I've always used transcriptions. So what that means for the listener, when we create a Pluralsight course, what I always do is when I, for instance, present a couple of slides, I write down everything that I'm gonna say upfront. And then I'm gonna I re read that, I record it, and then I stitch it together with the video. Mhmm. Mhmm. Some people are very good at not having to do that. They can just do it off the cuff. Like Tim, for instance, he can do that, which is amazing. And that's a talent to do that.

But, yeah, I I use that myself as well. And I think when you do use voice cloning or something, people sniff it out. People feel it. And then they're like, this is AI. No. Thanks. I I don't like it. It just feels wrong or something. I don't know. Yeah. There really is.

Tim WarnerThere's a sterility. There's off would be about the most technical term. Yeah.

Barry LuijbregtsYeah. The like, the uncanny valley. That's kinda what it is. Like, it's not. Yep. Do you as you see users consuming your courses and you can see the analytics and stuff as well, do you still see that people have the attention span and the willingness to actually sit down and go for a course of, let's say, two hours to learn something?

Tim WarnerLet me see. I need to ponder that for a second. Yes. But I think more than anything else, it's the subject matter. Pluralsight, I think it's pretty reflective across the world with technical training. The notion is of, generally speaking, shorter form content, atomizing content so learners can dip in and dip out however they want to. You know, that's all well and good. It can be taken to too much of an extreme, but, again, I think content is king more than anything else.

For instance, I've I've had a lot of viewership on leading edge, bleeding edge stuff, like if I'm teaching Claude Code, for example. Yep. That people seem to be, for some reason, willing to sit down with a lot longer than maybe me teaching SQL Server. So, I mean, the instructor factors in there, but I think more than anything else, it's the relevancy and currency of the content and how that relates to the learner. They may be under guidance

under strict guidance. You get smart on this Claude code business, or you're out of here. Or they might think themselves. If I don't get on top of this stuff, I'm limiting my career advancement. So

Barry Luijbregtsthat'll Okay. And so, just as a as a sidestep here, I've avoided teaching things about Claude code, for instance, because it changes so extremely rapidly. How do you deal with that? Like, with the UI changes, for instance? Right.

Tim WarnerThat's a good question. One thing I always make sure to call it out, and it can't be called out too many times. I'll say it right in the recorded training for Pete's sake. I'll say, now please understand, and I'll repeat this. What you're seeing on your screen right now when you compare it to mine may be very different. We're talking about rapid development, rapid iteration. We're talking about companies competing and trying to get ahead. So focus on the underlying 80%.

Why are we doing these things? What goals are we looking to achieve? Don't get hung up on the specific invocation operator or the UI element because they're those are gonna change. Yeah. Yeah. Exactly. That's also what I usually try to do. Yeah. And learners thank me for that as well. I've had quite a few learners thank me for that heads up because, you know, it's a way to avoid stall out and somebody rage quitting your course. And and that goes to what I said a few minutes ago about somebody having a negative experience with AI and just saying, ah,

the heck with the whole thing. I don't need that AI stuff.

Barry LuijbregtsYeah. Exactly. That works the same with courses as well. Hey. This button isn't even there. What is what is he talking about? I'll I'll quit this course. Yeah. Same thing.

Tim WarnerMhmm. Okay.

Barry LuijbregtsSo maybe as a final question. Looking I'm for an answer and I don't have it myself. Like I have two kids, they're six and 10. I forget almost. A year ago, I would say, You should learn software development or something in IT. You'll be fine. Maybe there's a lot less work or something, you can be an architect or whatever. Now I'm not so sure anymore.

What would you recommend that people go and study?

Tim WarnerOh, gosh. At the sound at the risk of sounding like I'm spouting a cliche, what I the guidance I offer to those who ask and the guidance that I give, I have one 16 year old daughter, is what my grandfather told me when I was a young boy. And he said, Timmy, if you find a job that delights you he didn't use the word delight, but makes you happy, you're good at it, and it's something that you would do even if you weren't getting paid for it, then you found your professional calling. And that always struck with me, and I'm grateful that I found that in the merging of technology and adult education.

So what I advise is people to really self inventory and think about what you're most passionate about professionally or personally and see how that lines up with your aptitude. Of course, you'll wanna have aptitude as well for it. But Yeah. Passion and aptitude, find where that nexus is, and evaluate that as a potential career path. And whether that's in the humanities or the sciences or the arts,

whatever, but for me, and I'm just speaking for myself, man, I don't wanna be in a position where I'm watching the clock because when you think of how much your life you spend on the in the job, I need to do something that's personally fulfilling. Not everybody's privileged enough to have that. I respect, and I wanna put that out there as well. But if you are privileged enough to have that choice, consider it. Passion and aptitude.

Barry LuijbregtsThat's beautiful. Yeah. I like that a lot. Alright. Thank you very much. This was delightful.

Tim WarnerYou're welcome, Barry. Always great to chat with you. I absolutely love your instructional content. Always have.

Barry LuijbregtsOh, thank you. Yeah. Same here. Big fan. Thanks. Alright.

Open the transcript file

Hosts and Guests

Hosts

  • Barry Luijbregts Host Feed-stated

Guests

  • Tim Warner Guest Feed-stated

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. Why Senior Engineers Matter More in the Age of AI Agents with Albert Tanure

    Oct 25, 202638 minAudio

    In this episode, I sat down with Albert Tanure, a Senior Cloud Solution Architect at Microsoft, to talk about what AI is really changing in day-to-day engineering work - and what still matters just as much as ever. We...

    Credit the people on this episode

  2. From Big Data to Agentic AI with Microsoft Foundry with Heena Refai

    Sep 15, 202638 minAudio

    In this episode, I sat down with Heena Refai, a Cloud Solution Architect at Microsoft, to talk about how enterprise AI has evolved from the big-data and traditional machine learning era into today’s generative AI and...

    Credit the people on this episode

  3. Rob Bos on GitHub Copilot, AI Engineering Fluency, and Better Software Delivery

    Sep 8, 202641 minAudio

    In this episode, I sat down with Rob Bos to talk about what’s happening across GitHub, Azure DevOps, GitHub Advanced Security, and the fast-moving world of GitHub Copilot. Rob shared how his work evolved from DevOps...

    Credit the people on this episode

  4. From Autocomplete to Agents: How Anton Sizikov Uses GitHub Copilot Today

    Sep 1, 202649 minAudio

    In this episode, I sat down with Anton Sizikov from GitHub to talk about how AI coding tools have evolved from “smart autocomplete” into full-blown agentic workflows. We unpacked the early days of GitHub Copilot, when...

    Credit the people on this episode

  5. Agentic Development Is Forcing Us to Become Better Engineers with Jon Galloway

    Aug 25, 202637 minAudio

    In this episode, I sat down with Jon Galloway to talk about what agentic development actually looks like in practice inside Microsoft and out in the real world. We explored how AI-assisted software development is...

    Credit the people on this episode