Ep 73: AI as a Scaling Tool - What Leaders and New Grads Should Do Next
Watch the YouTube video version above or listen to the podcast below!
Episode Summary
Stop the hype framing: The team pushes back on “AI will end jobs” headlines, arguing that having powerful tools isn’t the same as achieving real end-to-end automation inside organizations.
Why adoption lags anyway: Using the “free electric car” analogy, they explain how legacy infrastructure, incentives, habits, and change resistance slow AI rollouts—even when the tech is available.
The messy reality of automation: Clean data, consistent processes, and resilient systems are prerequisites most companies don’t have; dashboards go unused, spreadsheet workflows persist, and real work stays complex because business is problem-solving.
Tasks vs. jobs is the key distinction: Automating individual tasks can be transformative without eliminating whole roles—this nuance changes how leaders plan and how individuals prepare.
What to do next (leaders + workers): Leaders should treat AI as both runway and competitive threat (especially from AI-native entrants), while individuals—especially new grads—should start experimenting now, build “AI sensibility,” and create reusable workflows as tools evolve (ChatGPT/Gemini/Claude).
Ep 73: AI as a Scaling Tool - What Leaders and New Grads Should Do Next Podcast and Video Transcript
Dave Dougherty: All right. Welcome to Enterprising Minds. The latest episode, first episode of 2026 with the three of us. So exciting times. We are seeing as though it's about halfway through the first semester or the spring semester of colleges, and I've been doing some. Conversations at and guest lecturing at some colleges.
We're going to talk about the job market what that looks like, and also how do you lay the groundwork given all the market environments in your current chapter. While you're looking forward to whatever your next chapter might be Alex per usual, I'll kick it off to you to queue up the stuff you put in chat in our.
Pre-recording meeting, so on you.
AI Job Loss Claims
Alex Pokorny: Yeah, it's been kicking around the line. There's been a number of kind of well-known individuals, so futurists, you name it all kind of have a similar phrase around AI saying oh, it's going to end a bunch of jobs.
Dave Dougherty: Right.
Alex Pokorny: And the definition of exactly what they're claiming is going to be ended. The timelines all vary. It's almost always a year, two years, sometime like that out. Microsoft, CEO at one point said that white collar jobs could be automated by AI in 18 months. He said that in February. I have a little pushback against that at the beginning, but I do want to pivot to something a little bit more useful instead of just ranting against statements like that. I guess I've always had that kind of concern in the past too, of make wild statements for publicity and stock reasons versus, we should really be listening to them because they know everything. There's a pretty big, separation between those things and the bias there.
Dave Dougherty: It's not like Microsoft is in a good position with its AI compares to the big three,
Alex Pokorny: If you want to talk about Azure and all the rest and hey, how should we, who's going to be on top and enterprise sales? Yeah, I can see a reasoning,
but my quick pushback on it is just simply there's a human aspect here that is completely missing from those statements. There's a technology aspect that is, and a human element that's missing.
Dave Dougherty: Yeah.
Why Adoption Is Slow
Alex Pokorny: The point is, let's imagine everyone in the entire world today. Got an electric car for free. Shows up, everyone got one. Wonderful. the cars that currently exist also going to be driven, and it's going to take 30 plus years for those to finally die. So, will we have 100% electric cars on the road in two days?
No, we won't because people are still going to be using the same cars that they already have. going to be a whole bunch of people that are highly resistant to the idea of an electric car. They will maintain their gas car for as long as it possibly can, we can all imagine the audiences and the individuals that might be around or in our own lives that are going to hit those different populations.
Some of them will be super excited and will love the thing. A bunch will complain about it and a bunch else will just refuse it outright.
Dave Dougherty: Right.
Alex Pokorny: That's the same thing with AI technology within any company incorporation. Just to say that you have technology does not mean that you have automation. That's two different things.
Dave Dougherty: Oh yeah.
Alex Pokorny: AI is, yeah, go ahead, Dorothy.
Automation Needs Data
Ruthi Corcoran: just because you have the automation doesn't mean you have the proper data and processes to put in place like.
Alex Pokorny: Also, doesn't mean that it isn't fragile as heck in the moment that any change happens. The automation breaks, like the idea that there's like step one through seven of every task is just simply not true because over time it shifts. Things change, needs develop, or needs end, like it's just not needed anymore. That's, think of all the past inventions that used to be part of offices. They're gone. Things have moved and moved significantly different. So, the idea of technology being released in 18 months from now that could do some part of a white-collar job. Yes. But. I still push back against that because you're dealing with humans again, whenever we think of a business, there are certain elements that are always the same, having to do paperwork, taxes, dealing with employment law, real estate, maybe. then there's the actual trade itself. Plumbing, pizza, making, corporate accounting, whatever it is, there's elements of that always can be automated or are sent to somebody else who is maybe more of an expert. And there's other elements that can't be. So, have a plumber doing plumbing and you're going to have the accountant doing accounting, right? So, to say that you can have a piece of technology that can do all the jobs in between, including one person, communicating with another person, talking about whatever needs to be discussed, strategy workday today, just. Elements of admin that's not getting automated. We're not having cyborgs walk around.
We're having a computer system do computer related work that's
Work Is Problem Solving
Dave Dougherty: there was a a guest on a financial podcast that I was listening to that I thought had a really good way of putting it where, business is about solving problems. So, you know the jobs you're trying to find solutions to different problems.
And essentially what you're saying, if AI is going to take away all of these jobs, is that AI is going to solve all of the problems of those jobs. And that's just not going to happen. Because to Ruthi's point earlier, you need to have perfect data, which, all right, I've been in this game, 15, 20 years, I've never seen.
Perfect data ever, anywhere in any form, right? To your point, the human level thing is like, all right, we have this dashboard. Great. Who uses it? No one. What do they use? These four separate spreadsheets that are redundant because they only have one unique column across all four of them.
But then the 17 other columns are the same data from the three other data sources they pull from. Because somebody has a preference for a green header instead of a purple one, and all the ridiculousness that ensues from that. So yeah, I don't, I'm not too worried about that. Is the technology capable of doing it in 18 months?
Maybe we're three years into ai and. Most people are still using it as like a Wikipedia search or an answer engine. They're not actually leveraging it to its full capabilities of, creating landing pages or connecting different data sources or strategic information like pulling in multiple data sources from public spots.
Like the capabilities most people are using are so surface level. So when I see things like that I'm thinking, all right it's about the pr it's not about, actual reality. Especially when you're a company that's targeting enterprise organizations that are going to be five, 10 years delayed anyway.
Ruthi Corcoran: All right guys. This is great. Love the. Question.
Tasks Versus Jobs
Ruthi Corcoran: And also I don't, I per usual take a very different read of the actual quote that Alex put in or the things that are being tossed around, which is like most, if not all white colored tasks could be automated by AI within 18 months, which is quite a different sentiment then. won't be any white-collar jobs in 18 months. It's, hey, this is the possibility space. Now, we've already talked about automation and processes, and so I, I won't cover that too much, and we can always dig more into it. That's more of this idea of what is stopping the diffusion of the technology, like what are the barriers for it to be spread? I want to take it in two slightly different directions.
Also, circling back to Dave, you mentioning like you've been talking to college students and what they've been thinking about. So, what, who cares about this statement? Like why would we care about this statement? the two audiences that come to mind are first like leaders, planning. What is possible? What are my competitors doing? Where are my risks? What are the opportunities? And that's so key. If you could foreseeably, automate a large number of activities within your company in the next 18 months or 18 months out and that gives you a runway to say, what do I need to have in place in order to do that? We're the biggest cost areas of the biggest. Problem spaces right now, and how could we use this technology to think differently about those? That's like a huge opportunity. But then also, are my competitors doing? And perhaps even more importantly, what are my potential competitors doing? Because what's always interesting about new technology isn't so much like there's the piece of, you can always update your existing processes and make them go faster or automated. guys have noted a little bit of why that's hard. You, maybe your data isn't structured in such a good way. Maybe you don't have internal processes that fit it, but now all of a sudden you can have new companies come up with new ways of working that can utilize. The advances in the technology that are starting from scratch with this innate and that allows them to come out of left field and really eat your lunch, which, as an end customer, that's awesome, right? Because we will, that's competition. That's the engine of innovation, which is also an engine for prosperity.
From a, we're consumers. That's a really awesome thing that's happening if you're an existing company who's trying to think about this. This is, was the challenge and opportunity. So, there's that lens which is, hey, what's roughly our runway for thinking about when some of the major disruption might be coming from potential CUS competitors?
Or one, we might be able to start really seeing advantage of the opportunities and, okay, that's an 18-month runway according to Microsoft. The other lens I was thinking about is like the lens of individual people, employees, like, how do I need to change my thinking, both as an internal employee but also as a prospective employee entering the job market?
Like is this new technology doing differently? How do I think about it? I read this really cool blog post that we can post in the notes. I can't remember the name of it just now. But talking through here's how to think about and how not to think about what you do. With ai, you could get really good at prompt engineering, but that's not going to be helpful in six months because the LMS will be good enough that you don't have to be, that you can be a little bit sloppier with prompts, right? Or you could be really good at figuring out how to automate particular parts of your job.
But again, tools just, it's like website building like it used to be. You needed to know how to code HTML. Now you just have a wizzy wig and you just do it. So, any of your advanced knowledge about. Excel spreadsheets, H, TM L, all these types of things. Those are slowly going to just be eaten away at. Which is both terrifying if you're like, I've spent the last 10 years really honing my skill in this area, but also really awesome if you're like, I've never had time to learn that. Now I can just do it for me.
AI As Scaling Tool
Ruthi Corcoran: So instead, what you have to be good at is finding opportunities for how you can use AI to scale.
And this is an idea that I've been playing around with a lot lately and I'm still trying to wrap my head around it, which is like this. AI sensibility. AI is like a scaling technology. Instead of doing one campaign, now you can do 40 and you can test the best. So how do you start thinking in terms of, I could do this one limited thing, now I can do it a hundred times.
How does that change the way I work and the potential output I could get? So that's really cool. And the last one I'll pass, pop in here this idea of companies not hiring or just replacing workforce. Like in some places that may be true, there might be like pockets of the organization where a company just goes, oh, we don't need a team of that.
We can just have one guy plus AI and it'll be great. But that is those are like productivity gains. And when you have productivity gains, this is Jensen Wong, now, CEO of Nvidia. Like you don't just stop growing. What you do is you get to utilize more of your ideas, and if you have more ideas, you could expand what's possible.
You're likely to hire more because it's ideas that allow you to grow. So it could be that in some areas you like, you might go, oh, AI is replacing my team. That's true. But then there's going to be other areas where all of a sudden you can do a ton more than you could before. Because of the productivity gains.
Dave Dougherty: Go ahead, Alex.
Alex Pokorny: I was going to take those, each one, Dave, did you want to jump in quick beforehand?
Dave Dougherty: Yeah, no, I was just going to, a couple of quick examples, Ruthi, with what you brought up. It made me think of it anthropic releasing the AI for Excel sheets and having it massively outperform, Microsoft. How embarrassing for Microsoft to not be able to do what its users have been begging them for years for, and here comes this AI startup that's just yeah, this sucks.
Here we did it. It's a quick plugin. That makes me hopeful for that innovation side for sure, because it's like, all right, cool. We're getting more useful things. We're doing it. And then, even with this podcast, we would not have. A week, a weekly newsletter. If it still took me two weeks per episode in post-production, that's just no way.
I don't have the time for that. The fact that we can, I, with AI and creating optimized processes and doing that, going from two weeks per to about four hours, now I have. More than a week and a half of time back. Okay, cool. Let's work. All right, let's get after it. Let's expand it. Let's do more things.
Ruthi Corcoran: Hashtag AI sensibility.
Alex Pokorny: I like it. Yeah. I really agree with and that.
Startups Disrupt Giants
Alex Pokorny: Points definitely echo in my mind too, because new technology, I think of them like Kodak is a famous example of it too. So, Kodak, of course, I mean it was a film company really primarily, and that was massively their manufacturing and business and everything. they also created the first digital camera and then they basically shelved the product for like over a decade while competitors started to come online and starting to absolutely replace the need for film disposable film. And I keep thinking of that same thing here of the massive amount of overhead and stuck processes and individuals who are still in roles.
The company's not going to move them out of, and they're stuck. So, you're going to have new companies, faster moving companies, you name it, starting to come up and starting to basically be able to go against these large players. Like it's almost. It used to be like, man, that there's a giant, corporate in this particular market, and you're thinking, man, it's never going to happen.
Like they own it. They're there. They got all the money in the world. We're never going to have a chance. But that actually is different now with, they're likely to be so slow moving that we have a chance to get a foothold in and find an angle and get our foothold in and start to climb against them.
And there's so many. We've talked to some of the, the great kind of business speakers in the past have the exact same example of just these small, companies that corporations like to push to the side and not think about because the dollar amount isn't very high rapidly eat them away because the new wave comes in.
I think of also Steve Jobs and any account Bill Gates, any ear, early kind of PC makers doing stuff out of their garage. was perfect versus the larger corporations at the time who were so stuck on their current model because it was selling and they had jobs and profits dependent upon it. But the other one, like philanthropic is not dependent upon sales of Excel. it's usage of their tool. Especially at
Dave Dougherty: It's innovator's dilemma yet again.
Advice For Graduates
Alex Pokorny: So, I think from a college student perspective too, if you're concerned about this, start to think also the companies that are going to move past. And if you want to basically stay with one of those, so SaaS companies like Salesforce, those who are and kind of software side of things, they're going to have to move and react quickly and they're going to have to change.
And those people will develop in new ways, way faster. The companies who haven't changed yet, like large B2B manufacturers are going to be probably really slow on the uptake. other things would be regulated industries, government, healthcare, you name it, who forced into a particular process or system and they can't get out of it. that's going to take a really long time before they get out of it. So, there is some job security in terms of, I would almost say like classic roles. within those industries, and then there's going to be industries that are going to be moving faster as well as the startups who are absolutely, they don't have the data problems, they don't have the politics, they don't have any of that because none of it exists yet. have a lot of green space to grow.
Ruthi Corcoran: And frankly, go have fun. Like to your point, Alex, about you get to choose fast moving companies, maybe you should put more weight on those than the sort of like old school ways of working. I mean it coming out to college right now seems. Awesome. It doesn't seem like the recession doom and gloom situation.
It seems really cool. Why? Because you have the opportunity to join at this sort of emerging wave of new ways of working that anybody who currently has a job is trying to figure out, like, how do I both do my job now and figure out this new thing? Whereas you get to step into the new world. Figure it out together.
PS you might be the most AI knowledgeable person on your staff. It's like when the three of us came out to college and they're like, oh, you're young. Social media. Here go. You're now in charge. Like you've got such a cool advantage from that piece. And you like, cool thing about the AI tools is you can just use them. You don't need an enterprise license in order to use it. Like you don't to give a specific example. If you wanted to use and get trained for Salesforce, it's super hard unless you're already in an organization to go learn how it's used, et cetera, et cetera. You just have access to the tools. You can go use them.
You can go learn them; you can go figure it out. This is such a cool time to be graduating so. To all those who are like nervous about the job market and only having half a job, I say go after it. Like you are the master of your destiny. You've got agency and this is a really cool time.
Dave Dougherty: That's the end of Ruthi's Ted Talk. Thank you for joining.
Counterpoint On Opportunity
Dave Dougherty: No, I think you're mostly right as with most of our friendship, right? I'll take the counterpoint to Ruthi's. I do think it's a cool time to grow, to be graduating, to your point that there's more opportunity, but I view it as more opportunity for the individual to figure out what it is they're good at and what it is they want to go after.
Because I think I, I look back at a lot of the students that I went to school with. Now, granted I was a liberal arts major. I wasn't a business or anything useful. traditionally viewed as useful, I’ll be happy to debate that any at any other time than this episode. But graduating into that, it was where, what can you do?
Liberal arts degrees teach you how to think, right? Not just do a bunch of tasks, but how to actually think and be creative and go through, and that was amazing, right? So then going through and then hitting my, master stuff that taught me how the world actually worked. Here's the financial underpinnings, here's how people are actually making decisions in organizations.
This is why something that doesn't really make sense does make sense if you know how the dark matter is, working through the system all of that, it doesn't really change.
Scaling Yourself With AI
Dave Dougherty: But to your point with, it's scaling yourself, scaling what you're good at, scaling what you want to be good at is now easier than ever.
Because you now have your own, personal assistant or assistance depending on how far you take like agent tech work or setting up different workflows and like how nerdy you want to get in, with that. And I do think the vast majority of people just getting knowledge that you can do more than just ask it a question like you would a librarian and actually start doing some strategically important work.
Once we get more people on board with that, then it'll be fun because there'll be more competition to go against. Yeah.
Alex Pokorny: I think we're about a year to two years out from that right
of two main factors. One is stuff that was released a year or two ago that people are now more comfortable with, that are more advanced stuff with it, right? So, it takes kind a little adoption period to, for be able to get their arms around it. Claude Code, Claude Cowork. OpenAI's orchestration, I think are kind of like the three main shifts that have happened. The most recent one being orchestration, which one is more of an enterprise level tool, but it's basically as you set up a number of agents and you're having them learn from a combined set. So that they're starting to build up actually a combined knowledge database and knowledge store so that each new agent isn't having to relearn the same mistakes as the other ones. So, they are more progressing, like a more complete team or individuals, or complete, set of individuals within a team because they're starting to actually understand more of what's going on.
You can also tell them more of full posts of this is success. This isn't. So, you're also giving them more critique and more guidance, so they're becoming more advanced. So, they're more helpful, if I could say more for a thousandths of time. The other one is Claude Code. And Claude Cowork code was first, Cowork was second, and thus we're out for now a little while. and those are amazing. Absolutely amazing. The one thing that I think was like a big shock to me was I always bring him up with Ethan Monik again. he was testing out Claude Code and he had to do a test and gave it this like paragraph long idea of his, basically building a small business or something like that.
And it's used my profile, online profile and find a business for me and build it. Basically, it was the prompt, and it took 10 or 12 hours. Came back with a fully working site that was trying to sell like a pack of prompts based upon him, because he is well known within AI space. And that was the thing that was mind blowing to me was like wait.
This is not just some Google search summary kind of tool where you're just doing this chat question or Google can't quite comprehend the question that I'm asking and I'm using CLA or ChatGPTinstead. Or I want something more in depth and research like, so I'm using Gemini instead.
I'm basically having to summarize Wikipedia documents at that point. But this thing, 10 hours and it's building something like that to me was like, that was a huge step up.
Dave Dougherty: Yeah.
Alex Pokorny: And there are, copilot, studio Foundry. There's a bunch of other things that are similar to that also come in our price space. So, I think once we're about a year out from now. Then we'll stop thinking about AI like we did about two to three years ago or a year ago, which was more of this Q&Achat bot kind of thing. And I'll start to see it as like this thing that's way more capable. I think also makes it tough corporations, new businesses and individuals in companies; you also don't really know what AI is going to be doing to put it.
So, speaking of Ruthi's like thought process of like, how do you think about it? That's tough. I don't know. I'd love to hear your guys' thoughts on that. Like how do you, considering it changes, do you keep up with your mental model of oh, this would be good use of AI, or this is possible.
Now That one I think is a struggle.
Rethinking Productivity Goals
Dave Dougherty: I think that's where, how do you define productivity and what are your goals? Everybody's going to have to sit down and do the work to figure that out, right? Because how many people do you know that just don't want to be their own boss or don't want to start a company? Or don't you know.
There, there's a fetishization of that at least in the US, right? Of be your own boss, do this, do that. And then appeals to certain people. I know a lot of people that just want a job. They want to punch the clock and that's good enough for them. Fantastic. Good for you. I don't know how that works because that's not my mental model.
That's not how I am. But more power to you. Once you know what you're going after then you can automate stuff. But you need that north star. And I would be curious to you guys, both being economy majors how do you guys, do you think we need to redefine what productivity is? Because if you can do eight hours of work and two hours, like cool, I'm going to go take the afternoon off.
Or are you going to be working extra projects now for your companies?
Ruthi Corcoran: I am going to give you just real world stuff I’m and leave this sort of academic question aside. If I can do more, I'm going to do more. because that's just what I do. Like I think. The productivity allows me to run with more ideas than I could before, so we just do them.
I am a little bit at odds with your, you need a North star, you need a direction. Because sometimes we don't have it. And with the
Dave Dougherty: Great.
Ruthi Corcoran: like we don't even know what's possible. Like you might get a glimpse of it here and there. PS there's a really sweet Nvidia podcast series. I highly recommend everybody listening to. That just gives you a sense of who's doing what, so you can broaden your horizons.
Just Start Using It
Ruthi Corcoran: But like, when I think about getting started and how do we even begin to think about this and start using it in different ways, it's like. Nike, just do it like the, with these sorts of things. You just need to start doing stuff. Whether it's a trivial thing, like I used Claude. No, I used Gemini, excuse me.
I used Gemini to help me plan my seed sowing schedule and to tell me which trays, which seeds should go in and at which point they should start in the spring, in the summer. And where in my dining room they should be positioned for the proper amount of light. And I just started with hey, I, here's the things I have seeds of.
Can you help me get started? I have six seed trays and from that basic question I got all the things that we listed, which is like, when should we start? Where should I put it? How should I think about this? What should I direct? So, what are like, you just get started and there's this really cool cycle that I'm noticing in projects at work or personal things where. Just do a thing and then you learn a few things, and then you do a thing and then you learn a few things, and you get that sort of feedback loop going and that. After I realized this, I was like, that's lean startup. So read the book, lean Startup. If you need a guidebook, it'll walk you through like academically how you think about this thing.
But really just get started with something and start playing with it because it's through the actions that you're going to learn most and you're going to uncover things you don't know. And I think what's really cool about LLMs and the place that we're at now, again, going back to my little hype pitch for college students, you had to go Dave to a master's program. In order to learn how to think those things
now you can just ask. You just be like, what do managers care about? Why do they think of this way? What are their incentives? How do I think about this? Like you can just now ask. And so it's like what you can do is limited to A, the amount of time you take for it.
But then B, like ability to just questions, use your imagination.
Dave Dougherty: Yes. And.
MBA Versus LLM Context
Dave Dougherty: So, I totally agree with you. If you are a naturally curious person you can now go to ChatGPT, and you don't necessarily need the MBA. However, for me, the MBA was an important persuasion tool for the audiences that I was targeting, right? Which in that point was hiring managers. because coming off of creative writing and music and being entrepreneurial and doing ghost writing and freelance work and all of the things that.
I had to do graduating into the great recession, to pay the bills. How do I level that up? People don't take the musician entrepreneurial thing seriously. So how do I demonstrate that? I know, the ground rules of business. I have an MBA, right? And it shouldn't matter, but because people are people it does, right?
Ruthi Corcoran: That's a really important clarifier.
Dave Dougherty: Yeah.
Ruthi Corcoran: still matters in many cases, but in terms of the knowledge gained, I think this is a really cool space to be in now, where
Dave Dougherty: Absolutely.
Ruthi Corcoran: context for free.
Dave Dougherty: Yeah. Alex, you've been awfully quiet.
Alex Pokorny: Yeah, I was going to throw my 2 cents on this. This is kind of an interesting point, so this might go in a slightly different direction, but Ruthi, your explanation of thinking through an activity using ai. I do wonder how much of that is simply an active sounding board that's available at any point. Speaking of activity and just do it like you have an idea, put, a little bit of thought toward it, but there's also a million other things going on in life that are going to be asking you to do the next thing. So, it's on your list, your to-do list, your mental checklist of like things that you want to accomplish at some point. But now can stop for the next five, 10 minutes and be tapping away on your phone and start having an active. Basically, sounding board thought conversation toward it, building out some of the details and teasing out some of the things that you need to think about and consider when going through this process.
And I some of these I was wonder of like how much of that was already going to happen if you had spent that time, just going through that motion, that activity, and out loud, literally would get you most of the way there. And then there's the little tidbits and tips that it's basically including along the ways which, some of which you might decline and deny.
I've done that with Claude a number of times where I had advice that I was like, I, I don't agree with that. That's not what direction I want to go.
So, it was much more of activity towards the accomplishment of the thing that I wanted to do versus this oracle of knowledge that I now have access to. It was much more of, I don't know, maybe that's just a personality trait of talking things out, which I like to do. Maybe that's part of it too.
Ruthi Corcoran: But I
Dave Dougherty: You had talked about it, saving your marriage with the design features.
Alex Pokorny: That's true. And endlessly annoying my wife with interior design questions versus and mostly tapping away on my phone. Far better to annoy the phone.
Ruthi Corcoran: But there's an important speed component and then there's this other because you can do this in snappy real time, right? Whereas like my little seed starting thing, like I would've to sit, I would've to look up individually how, where does each seed want to be and how does it go? And like I can do that and I have a pretty decent mental repository, but it takes longer. then also there's this other element that's new, which is like at the. Throughout the response, you're going to get little bits and pieces that spark new ideas. But then also those like three helpful responses at the end of everything is do you want me to do this? Do you want me to do this?
Do you want me to do this? Like it, it gets the forward momentum that you wouldn't necessarily have it. It's like next steps in a nice way that doesn't make you cringe,
Switching Models and Migrating
Dave Dougherty: The one thing that I've been wondering is with the changes in how the models are working found myself going from leveraging ChatGPT early 2025. To then migrating more towards Gemini because I prefer the responses. And Gemini 2.5 came out and it was just better than what GPT was doing now Claude's out and took the leaderboard.
And so, I've been finding myself migrating over to Claude because of the extra capabilities and the writing. Since I use it a lot for writing and thinking through. Creative writing stuff. I'm leveraging Claude more, and then I'm sitting there looking at all the stuff that I have. It's two and a half years of stuff on ChatGPT and I'm like, man, do I migrate any of this?
Do I save this? And just create a master prompt for like the main use cases and based on what I talked to you about for two and a half years, summarize it for a system prompt or something, or. And then delete everything or do you just delete all the context and then rebuild it every year so that this is my 2026, whatever.
Have you guys migrated the knowledge base or thought of that even yet? I,
Alex Pokorny: Yeah, I did that. OpenAI shifting from that common usage to Claude.
Dave Dougherty: yeah.
Alex Pokorny: I have edited the rules of responses for how GPT operates a number of times. My most recent set, it finally starts to respond in a way that I prefer very deterministic statements, it will state its level of confidence in parentheses around different statements, stuff like that,
Ruthi Corcoran: you have to share this, man. This sounds awesome.
Alex Pokorny: Yeah, I
Dave Dougherty: It is so Alex. Yeah.
Alex Pokorny: Oh, I got so frustrated. It was like, you need to do this right now. This is the only way to do things. I was like, what? freaking program tells you what to, so much of that rebelling that doesn't work. Plus, also it just wasn't, the prompt, speaking of like the prompts that come afterwards.
Afterwards, response of should I do this next? It was also extremely confident with basically running down its own line of thinking versus shifting away from it, which Claude
Dave Dougherty: Right.
Alex Pokorny: easier to a better conversation with and shift it to something that's the output that I'm actually looking for. So, the way that I did that was basically is I went through OpenAI's history, so you can look at basically the file that it has on you. And it's a pretty long summary file that it constantly uses. It refers to, and it adds to whenever you tell it something like, oh, I have a dog, or I have a cat, or, whatever else, or I live in this state country. It starts to add that to that file. And what I've done with chat GPT, but also Claude, is create separate projects. So, I have ones that are more work focused projects and then I have more personal life projects. So, in general, that's the two big But I actually have. Based upon things that I've done, like chat pulse, things which are automatically updated newsletters and stuff or
Dave Dougherty: Right.
Alex Pokorny: that it does on a recurring basis. You've got three or four of those that run now, and those are in separate projects as well, because stuff is already either around like job search stuff or keeping up with my particular set of industry and skillset kind of things so, it's not really work, it's not really, it's an in-between area, so I made different projects for that.
So yeah, it's totally possible to take that history. Import it over, throw it in, it still takes quite a while. You can also ask, any of those to request that it makes a master prompt for you and that alone, that phrase. Will kick off a gigantic interview process that I basically ask you who are you?
What do you do? What are these your sort of job activities? Okay, you said this, but do you mean this? And it'll start going through all these different questions until it basically compiles this master prompt that it will then you could use for any of the other tools as well. But also, it has now collected basically a really good idea of the thing that you want to talk about, the information you want to give.
Dave Dougherty: That's really interesting,
Ruthi Corcoran: That's cool.
Alex Pokorny: and the master prompt really helped because there was like, trying to, I don't want to ever give company information, but trying to like, line out where I am comfortable being able to like, line that out and give it a much more better context. It's responses. Were far more realistic with what it was suggesting at that
Dave Dougherty: right?
Ruthi Corcoran: Just off of that last little bit that Alex walked us through and then going back to our. who are looking to into the job market, and this could be applying for anybody, people to talk to people about what you're doing and what you're experimenting just like we are today.
Because all the things that Alex was just like, and then I found this and I tried this in no world would I have done those. Similarly, with what Dave's doing with how he's using the transcript technologies and the recording trans to then magically link things. I would, it would never occur to me to do those.
I would do other things that I can potentially bring to the table and the. Ability to talk through with other people and compare notes is such a cool learning and innovation engine that I think is super important. So, find some friends who are excited about nerding out and
Dave Dougherty: The only seed trays that I do is roasting. Yeah, no, I think that's the big takeaway for this episode, I think is, that it is a unique place to be for the new grads. Even with middle management, mid-career people like ourselves, right? Where it's all right, we have to do our old job, but we have to account for the new. And do I want to be part of the change or, what am I comfortable with? I think it's a really interesting time. And I'm, and as always, I'm interested to see where we end up. Obviously, we'll continue to figure this out. We're going to have a number of kind of special episodes coming based on the talks that I've been doing.
And I know Alex, you brought up wanting to do certain things, be on the lookout for those. And as always comment, subscribe, go check out Pathways and also first of its time we now have a website for the podcast. So go check that out. It is enterprising-minds.com. Again, it's enterprising minds.com where you'll have all of the episode transcripts, all of the Pathways newsletters.
All in one spot. Go check that out. Let us know what you think. I think there's a really cool video on the homepage if you don't mind me saying so. Yeah. Thanks for hanging. Have a great couple of weeks. We'll see you in the next episode of Enterprising Minds. Take care.
Ruthi Corcoran: Cheers.
