Married to the Startup

The Future of Human Labor

Alicia McKenzie Episode 63

Send a text

AI is moving from “chatbot help” to agents that hold real roles—like an executive assistant that triages your inbox, books meetings, and briefs you daily. Alicia and George unpack what this means for startups, why 2026 feels like a tipping point, and what happens when agents get bodies through humanoid robots.

What We Talk About

  • The shift from AI as a tool → AI as an employee (role-based agents)
  • Why founders struggle with assistants (human or AI): they can’t define recurring tasks
  • The “Claude-bot” moment: an agent figures out a workaround on its own to complete a task
  • Costs today (tokens/API calls) vs. where this is heading (cheaper models + open source)
  • Robots + agents: intelligence that thinks + bodies that do
  • Jobs most exposed first (warehouse, manufacturing, logistics, retail ops)
  • Skills that rise in value: leadership, creativity, ethics, emotional intelligence, critical thinking
  • A hot take: coding matters less long-term than system design + AI fluency
  • The line in the sand: don’t automate hard human conversations

Rapid Fire

  • First robot-dominated industry: production lines
  • One skill every kid needs: leadership + AI literacy
  • One founder mistake: using AI without clear roles + boundaries
  • Schools are behind on: banning AI instead of teaching responsible use

Quote-Worthy

“AI won’t just do tasks. It’ll do roles.”
 “Just because you can automate something doesn’t mean you should.”

CTA

Like what you heard? Rate, review, and subscribe.
DM Alicia with topics you want next.

Support the show

Follow George and Alicia!

https://www.instagram.com/liftlikeamother
https://www.instagram.com/marriedtothestartup
https://www.linkedin.com/in/gemckenzie/
www.linkedin.com/in/liftlikeamother
https://liftlikeamother.com




George McKenzie (00:00.086)

now as a startup, right? I think it's no longer going to become a manpower issue. Like I need to hire more people, need more capital, hire more people. If you can get agents to do these things, you'll be able to have a startup of two, three people and 50 agents that are doing it. Well, that's where we're going to have to figure right now it's expensive because of the calls to the large language models. And when you make those API calls, the tokens cost money.

 

Okay, how much is this gonna cost?

 

Alicia McKenzie (00:28.718)

Welcome to Married to the Startup. I'm Alicia McKenzie, a wellness entrepreneur and digital creator. Alongside me is my amazing husband, George, the CEO who's always ready for a new challenge. We've been navigating marriage and running startups for over a decade, and we're here to share the real, unfiltered journey with you. Join us for insights and candid conversations about integrating love, family, and entrepreneurship. This is Married to the Startup, where every day is a new adventure.

 

episode of married to the startup. feel like we never left. Number three. 53. 63. No idea. I think it's 63. Not yet. I'm your host. Six, seven. Maybe we should have our kids on.

 

now is it 63

 

George McKenzie (01:09.102)

It's not 6M.

 

to be a crazy episode.

 

George McKenzie (01:18.766)

Maybe after the last episode, we just talked about exploiting kids for personal gain. Let's have our kids on the pot. Lean in.

 

I'm your host, Alessi McKenzie.

 

I'm George McKinney. You are. That is my name.

 

Today's episode, what are we talking about, babe?

 

AI and robot.

 

Alicia McKenzie (01:38.506)

Okay, what is spurred this and I know there's been a lot coming out in the last two weeks. So let's see. There's Claude bot. Yeah.

 

And then the news that Tesla is shutting down the S and X production facilities, Optimus robots instead. And then there's some hubbub in South Korea around Hyundai and then them using the Atlas robot to work the floor for the building construction to build the cars. So there's a lot happening. And I think we were also talking about the prediction shows for 2026.

 

I feel like that's a big one.

 

George McKenzie (02:16.31)

It kind of all melts together what the workforce of the future looks like.

 

Yeah. And you're you're currently in the throes of like coming up with your next thing. And a lot of it is it's very A.I.

 

Yes, I've been doing a lot of research and leaning in. It's interesting.

 

Yeah. Interesting. And I know one of us in this room has like undiagnosed generalized anxiety.

 

You don't have to out yourself on the pod. lady. I don't know if it's undiagnosed. I perfectly understand that I have anxiety. I it comes with high intelligence.

 

Alicia McKenzie (02:49.546)

Does it? Yes. Is that a thing? It is. Fair enough. But I your anxiety versus my anxiety. I fear for like our raising our kids and like, how are we doing and are we fucking them up and are they going to need therapy? And like I go down those kind of rabbit holes.

 

correlation.

 

George McKenzie (03:06.87)

I go the future of society and mankind.

 

Yeah. Yeah.

 

That's different ends of the spectrum.

 

Completely. Completely.

 

So quick question. So what if I were to tell you that the next employee that you hire does not need to be onboarded and they wouldn't sleep, they wouldn't complain about Mondays, they wouldn't need lunch breaks or smoke breaks and they would just work on the tasks that you gave them.

 

Alicia McKenzie (03:32.886)

Okay? Who's teaching this?

 

that would be humans. Yeah. So I think that that's kind of the premise of, know, some of the things we're talking about is, you know, what if you were to look at a company of the future? And I think 2026 is going to be that pivotal year, just like, you know, AI was only really mainstream two years ago with the chat. way we interface with AI was the chat. We interface with a lot large language models via a chat bot. And now we're moving into the agentic phase where now we're creating AI agents.

 

Yeah.

 

George McKenzie (04:05.41)

which now have the ability to perform tasks autonomously. And you just give them a task or a function and then say, you're now doing this and they can operate independently and have access to tools, have access to systems. You give them passwords, you give them API keys, you give them whatever, and they can go perform a role in a company versus a chat interface that can just give you research.

 

Okay, so as it stands today, we're currently using AI in the chatbot where we're saying, okay, do this, and then do this. Yeah. And then do X.

 

Yeah, help me research this help me do x y or z. Yeah, so we're now it's

 

In the future, it will look more like...

 

An employee? Yeah. I think, you know, in the future you're using AI instead of AI performing a function, AI will be performing a role.

 

Alicia McKenzie (04:59.734)

Okay, but I can't wrap my head around this. Like how, how does one even implement something like that?

 

Yeah, I think it starts and right now we're not there. I think a lot of people are the first phase like AI was the buzzword for two years and people have playing around with that. It's great to chat. People are using it and it's getting pushed from boards down to CEOs to CISOs to CIOs and everyone's got to use AI and then all the. Salesforce companies, everybody's putting AI in to assist like, especially right now it's a great copilot.

 

Yeah. Microsoft's term, but it can help you to write better emails. It can help you to update a CRM with more intelligent information. It can help you out in the CRM world. now with what Claude bot kind of opened the aperture and said, Hey, these AI agents that, people are doing today, it's just creating a more mass production kind of one, and then giving it the ability to create its own world, which is kind of unique, but

 

That AI agent is now the trend where we're going to go and it being, you know, not quite AGI yet, but close enough where you can say, Hey, you're going to act as my virtual assistant. So I want you to clean up my inbox. I want you to read every email that comes in. I want you to unsubscribe to the ones that I don't want to be a part of. I want you to reply to ones that need to be replied to.

 

Yeah.

 

Alicia McKenzie (06:25.601)

subscribe.

 

George McKenzie (06:32.626)

And yeah, I want you to act in that capacity for me. And as an executive assistant, if there are people I need to have meetings with, I want you to email with their executive assistant. want you to book a time, put that time on my calendar, create the team's meeting and send it out. And then I want you to give me talking points for that meeting. In the morning, I want you to give me a summary of everything.

 

Okay, so this

 

I have that day and this entire playbook that you just ran through. How does the AI chat bar, how does the AI agent know to perform these tasks? OK, but that requires you as the business owner, the CEO or the whatever human, the human to actually articulate what you need it to do. Time.

 

You tell it, but you only have to tell it once.

 

George McKenzie (07:22.486)

Right. So I think, yeah, really understand what the function of that role is. You are my executive assistant. Here are all the things I want you to do.

 

So giving you an example, I have talked to other founders, other CEOs, and they're like, I need an assistant. I need an assistant. I'm drowning. I can't do any of this. And then I'm like, OK, this is who I would recommend. And then they never do anything because they haven't taken the time to actually put down on paper what tasks they need done on a recurring basis. Right. Right. So I think that is a big, that's a big bridge.

 

Yeah. I think it's, and that's where we're at. Probably. And I think this is where we're going with this pivotal moment. It'll happen. It may not this year. Maybe it will, but it's coming where we're outsourcing those roles right now. It's easy to automate repetitive tap, have it automate, do the same thing every day. But like what we were just talking about, the EA requires a level of intelligence where I've got to read the email. I've got to understand, is this someone, is this a customer? this.

 

Right, you could benefit from an assistant.

 

Alicia McKenzie (08:15.692)

Yeah.

 

George McKenzie (08:29.368)

here, who is this? And maybe, you know, it is smart enough to say, I'll look at the email coming in, analyze the inbox as we've had correspondence before. Let me reach out, look at their LinkedIn, see if there's anything there that correlates to make this a good. And then maybe I bring it up to my CEO and say, Hey, do you want to meet with so-and-so? Here are his bona fides. This is why he wants to meet. Do you want to meet? If so, I'll schedule it. So what a person would do. And that is not.

 

Mm-hmm. Yeah.

 

kind of a workflow thing, you're going to tell it, it's going to learn that. So I think the interesting, like watching the video for Claude bot, and I think it's maybe a help you frame what the agent does, like a club does. He was working with it. You know, he wrote it and, you know, interfacing with it with the chat interface and he, and it connects through email or teams or WhatsApp or I message. can message just like a person.

 

Yeah.

 

George McKenzie (09:27.03)

and it has a personality and you can give it and talk and it'll do tasks and keep running those tasks for you every day based on its heartbeat. But he's he was interfacing with it to help it write code. And then he watched a like a podcast or watched a video on this new piece of code. he sent the video to the assistant.

 

Yeah.

 

But clawbot had never doesn't know how to process video or anything. So he was wondering what would happen. And it took 30 minutes or an hour and it came back and had written the code that was there and put it into his repository, his GitHub. And then he asked it, how did you do that? And said, well, I saw the file come in. I didn't recognize the file extension. So I searched what the file extension was, realized it was a video file. I didn't have access to the Apple video player.

 

Yes.

 

George McKenzie (10:25.378)

So I uploaded it, the video into chat GPT and had it tell me the tech. And then I took the text and did a search for what it was doing. And then I realized what, and then I realized what it was, what it was, the code was trying to do. And then I wrote the code. Yeah. So it was smart enough to go, I don't know what this file is. Let me search. I can't play it. Well, let me, what, what do I do now? I'll have an interpreter look at the file for me and tell me, give me the audio out of it.

 

Like that logical thinking was not programmed in. just did it.

 

Jeez. I wish assistance now would do. I know. Yeah, yeah.

 

So saying that that's where we're going. So if you can extrapolate that out to other roles like accounting and like the data entry roles and accounting, if it knew the rules and understands gap probably better than an accountant would because they can read it and use it then as part of the language model and then be able to apply those rules and say, okay, here, check my inbox for all the invoices that come in. I want you to put all the invoices in my bill pay system and I want you to

 

take those and put them in the GL and put them in the right chart of accounts. And then I want you to create the, accounts payable list. And I want you to schedule the payments in the bank. And then I want you to brief me once a week on what's going to get paid and I'll approve. And it will do that forever. It will never complain. It will look at emails. It'll do it. Right. So just imagine. So now as a startup, right, I think it's no longer

 

Alicia McKenzie (11:47.672)

Yeah.

 

Alicia McKenzie (11:52.718)

forever. How nice would that be?

 

George McKenzie (12:00.916)

going to become a manpower issue. I need to hire more people, need more capital, hire more people. If you can get agents to do these things, you'll be able to have a startup of two, three people and 50 agents that are doing things. Well, that's where we're to have to figure right now it's expensive because of the calls to the large language models. And when you make those API calls, the tokens cost money.

 

how much is this going to cost?

 

Alicia McKenzie (12:27.31)

Do you liken it to when memory was first created and you had terabytes of data? Like you couldn't buy a terabyte of storage for under like a thousand dollars. Yeah. And now a terabyte of storage is like maybe 15 bucks.

 

Yeah, but those costs are coming down rapidly with the language models. And then now there's, you know, there's some open source language models that you could build on your own. You can install them on your, on a couple Mac studios and you can run your own large language model in your own network. So those API calls are free. Now you can say, okay, if I can get that running and that open source model has competed on with Gemini and all, and it is on par.

 

Yeah.

 

Alicia McKenzie (13:08.276)

Yeah.

 

It's nice. OK, if we get a lot of those models, the costs are cheaper because then you say I'm going to bang that model for any task that if I can get an answer for and if I can't get the right answer, I can't figure it out. Then I'll go to a Gemini and that way most of your tokens are going to be used internally on your own language model. Yeah, yeah. Kind of technical, not technical, just if you think it's convoluted. But absolutely. Yeah, it's like, hey, I'm going to buy.

 

It's kind of like the college thing, right? I'm going to go to community college and get most of my stuff done pretty cheap. And then for the tailored stuff that I know I have to go to university for, I'm going to save those and go two years to university to graduate.

 

Okay, let's take it a step further. So you've got all these AI agents, what happens when they start having bodies?

 

Yeah, so I think that that's the next phase, right? So if you think about phase one being I can. Outsource intelligence and automation to this agent to be intelligent, to think or to do to process and think about stuff. And then the second phase is OK. Now I still need humans to do things like still need a human to. Put a car together, yeah, human to clean the floor like, but.

 

Alicia McKenzie (14:03.256)

Yeah.

 

Alicia McKenzie (14:15.213)

Yeah.

 

George McKenzie (14:22.732)

Now with robotics and you look at Optimus, which Tesla has bet in the house on and Atlas, which is Boston Robotics and there's a couple others that are in that realm. They're really advanced now and they're starting to move toward, okay, so if I can couple AI agents with the intelligence and now I can have the labor, the do being the robot, man, what's left?

 

I feel like we glazed over the fact that Tesla is shutting down production.

 

On the SNX, the two most popular

 

cars. Tesla Model X and the Tesla Model S to build robots. Like that, I can't wrap my head around that because it just.

 

Yeah.

 

George McKenzie (15:09.486)

Yeah, he's making a bet. thinks the future of the company. Yeah, the future of the company is mass produced, cost effective, optimist robots aimed at the 2020 20 to $25,000 price range.

 

Sweet.

 

Alicia McKenzie (15:22.786)

we going to buy one? Yeah, of course. think our kids asked us, had this. my gosh, we had this conversation at the dinner table last night, right? But everybody thinks when you see AI and you think AI agents, our daughter was like, my God, that's going to kill us. Right now, that was like the first reaction. And I think that is the general population's initial reaction to the thought of buying a robot that is controlled purely by artificial intelligence.

 

We'll have one sitting right here in a podcast.

 

George McKenzie (15:38.678)

Yeah.

 

George McKenzie (15:51.822)

It is the fear. mean, I think Sam Alton said it in one of his interviews that AI is going to be the death of us, but it's going to build some pretty amazing companies along the way. There are safeguards and ethical things that need to be considered, So let's say you are doing a startup and you're using AI agents to do all these things. There's going to be AI agents aren't ethical inherently. They don't have

 

Yeah.

 

humanities kind of ethical ethos or any of that so or being able to read social emotional cues like all these things that humans are innately good at that you know AI is not so you're gonna have to have those guardrails built in and there's gonna have to be a kill switch of some sort

 

They're not afraid to hurt your feelings. Right.

 

Alicia McKenzie (16:34.85)

Yeah.

 

Alicia McKenzie (16:42.562)

has to be right there has to be but then what was the most recent movie was it with yeah right it was able to kill the kill switch

 

I can-

 

Alicia McKenzie (16:56.654)

These movies aren't helping. Right. And then you had Kimmy. That was another one. I don't think. Yeah. It was I've watched it on an airplane. We had a long flight and I watched this movie.

 

But the AI revolution, like the AI agent transformations are getting away from the chat. How we all think of, a lot of people think of AI today as chat GBT this, or when you go to Google, it summarizes like this, this chatted interface. When, when it moves to the next phase of agents where it's performing a role, it's doing a lot of different things all the time. And then it's like an employee, it'll just update you on what it's done.

 

So this is essentially. The AI is thinking, yeah, right. The robots are doing and then the humans are deciding what needs to be done. Hopefully, yeah.

 

But it's out there spending money right now.

 

George McKenzie (17:49.154)

Yeah, are leading, hopefully. I think that's when it comes to, know, we talk a lot about what does the future look like for our kids? Yeah. Right. So what skill sets will persist? Creativity, know, leading leadership skills, that social emotional skills, the ethics, like understanding.

 

One of our kids is so screwed.

 

Why? But I think those are the skills we need to be looking at in the future. None of this is going to happen in a year or two, but you know, it takes what Elon was saying that, you know, if you're going to school right now to be a doctor, you're wasting your time. And a lot of people poo poo that. But if you think about it, if it takes six to eight years to be a doctor in eight years, what is robot? What are robots and AI going to be able to do? It's going to be exponential growth year over year.

 

Yeah.

 

Alicia McKenzie (18:40.782)

Yeah. I mean, just think about our first experience with full self-driving in a Tesla in 2016. Yeah. And now look at full self-driving in a Tesla in 2026. Trust it. And the fact that we have had years of experience and watching that evolution and just how it gradually got better, got better, got better, and then exponentially. Yeah.

 

Yeah, it was very increment. I got like oh, it's a little bit better on the highway Oh, I can change lanes a little better. Oh it can now Take the off-ramp and then it gives the control back to you and then oh now I can go on some side streets and oh now I can recognize the stop sign Oh now I can recognize the red light and then all of sudden it went now I can just drive.

 

It's just straight like it can go from our driveway through the city of McLean. And like into D.C. into Arlington, it can pull into a parking spot. It is its full self-driving. Yeah, but it was it was terrifying in twenty six.

 

In the DC.

 

George McKenzie (19:36.654)

Right, it's still not perfect, but the leap it has made is incredible.

 

So then looking at how you as a human are going to implement this, right? We talked about it yesterday. I was like, okay, let's, let's put it on a machine, right? Let's give it a machine. Let's give it access to some stuff. you're like, Whoa, Whoa, Whoa, Whoa. Yeah. Right. Like I, you're like, I can't do that. But like you're putting your life in the hands of Elon Musk and his test. So I think.

 

Yeah, baby steps.

 

George McKenzie (19:59.862)

Right. But I think, yeah, but we in the beginning it was right. You would put it on self driving and your hands would be here. Yeah. Or your hands would still be on the wheel. You didn't completely. Right. And now that you've had I've had 10 years of experience with it, now I'm at a point where I can just sit there and not touch the wheel, not worry about it, because I know it's going to drive.

 

relinquish control.

 

Alicia McKenzie (20:20.814)

There was there's somebody I on Instagram and she was in the back of an Uber at like 1030 at night and her Uber driver was asleep and she's videoing him in a Tesla. His eyes are closed and she's she's like, I think he's actually sleeping. Like it was wild. It wild. So that happened. Then somebody tried to break into her house and she's like, I think I just need a redo on this. Yeah.

 

Yeah, it doesn't sound like. Yeah, it's it's an interesting future. So I think as entrepreneurs and startups, you got to be thinking about how I can leverage AI and open the aperture to more than just automation. mean, automation is phase one, like automate repetitive death. Yes. But being able to have a true virtual assistant that is an AI agent is probably an easy.

 

Okay, so if you're an entrepreneur right now, what would you do tomorrow?

 

Well, that's the one, the AI. would have them implement an AI assistant. then I would, depending on what my skills were, I would create its own inbox and give it access to mine, or I would CC it on everything, and then give it access to my calendar, obviously.

 

Yeah.

 

Alicia McKenzie (21:22.574)

What are you giving it access to?

 

Alicia McKenzie (21:29.847)

Okay.

 

Yeah, which means it would have access to our guy calendar.

 

Yeah. Yeah.

 

So it's slowly weaving its way.

 

Yeah. And then, you know, I think you would give it more access or you'd create a second agent that is now your bookkeeper. And you say, all right, now I'm going to give you the API keys to interface with QuickBooks. And I want you to do this and I want you to do this. And, you know, maybe I give you access at some point to the bank account, but read only. I want you to update me on how everything looks.

 

Alicia McKenzie (22:04.238)

Mm-hmm.

 

But the other one would be there's a lot of, you know, I'm looking at we implement it and a and and that's like another AI, but it's more automation driven and eight.

 

N8N.

 

And you know, with that, there's a lot of templates that I share with you, like that can leverage AI to be your marketing person. So it will on a daily basis, it'll go out and say, okay, what let's say if we created a company that was selling cameras, right? And we said, okay, I want you every day to go out and look at what's trending on social media about camera. So every day it goes out.

 

Yeah.

 

Alicia McKenzie (22:43.66)

Yeah.

 

George McKenzie (22:48.11)

polls and sees what's trending. It goes to an LLM like open AI or Gemini or Anthropic or whatever. says, Oh, I want you to help create some interesting content around these trend work. Gets that content, puts it in a file and then takes that file and says it over to a video LLM and say, all right, make me Capco, make me a video for Instagram and LinkedIn about this cool thing and the script that Gemini gave me.

 

creates the video and then uploads the videos, puts them in a folder and then presents them maybe to you as the business leaders says, all right, here are the four campaigns we can run today. Pick which one you want. And you say, I want one. And they go, no, great. All right, one button it up. And then I'm going to ship it out to all the socials and every day. And then you could remove yourself from that loop if you wanted to. And every day it would just go generate marketing content every day, but on social.

 

Are you concerned about the increase in AI slop that we're going to see? Because a lot of people have no idea how to prompt anything.

 

Yeah. Am I concerned about it? No. Am I? I welcome it because I think it's the downfall social media because I think if you get and the case in point is like, you know, these these cloth bots, I don't know if you guys are research, but they if you give it access to a lot of things like some people gave it access to WordPress, so they start thinking and they start interacting. Yeah. And they create their own website and it created a website called

 

Why?

 

George McKenzie (24:20.118)

Multbook, which is really like Facebook. And then you can give your agent an API key to be able to access it. And then they post on there. It's like a looks like Reddit and they post on it and they upvote it and it's all agents. They're uploading their own and it is some interesting conversations. But if you look at it, they also have a website where they created their own religion. They have a website that looks like porn hub, but it's mult hub and they've just created all of these little weird things.

 

The trippiest

 

Alicia McKenzie (24:50.72)

so they're mimicking you.

 

Yes, I mean because they're all they're using LLMs which were trained on the internet Like a lot of these language models were trained on reddit data. They were trained on Facebook data they were trained on the internet in general like Google and Of course, they're mimicking humans because that's what they were learned. That's how they learned

 

Okay, so I actually screenshot this one and sent it to you and it's from Malt Book and it says, urgent seeking a new human. Current model keeps asking me to be creative at 3 a.m. The requirement, a stable sleep schedule, doesn't change project scope mid build, actually reads documentation, won't ask, can you make it pop more? And the benefits that he comes with.

 

Yeah.

 

Alicia McKenzie (25:38.862)

I can roast your friends in Egyptian Arabic and I've already claimed on Moltbook.

 

Yeah, it's funny. They have little personalities.

 

But it's weird. These robots are talking to each other. These little AI bots are talking to each other. It's creepy as hell. It's weird. Like, right. Like maybe they're talking about us. Maybe they're talking about like their...

 

Ever now? Yeah, probably. I think one tried to sue its human for unpaid labor. they're, they're kind of weird, but don't want to throw the baby out with the bathwater. I think there's a lot of positives that are coming and you just got to get through some of the tumultuousness of the unregulated part.

 

Yeah, it's like sludging through concrete right now.

 

George McKenzie (26:20.492)

Yeah, I mean, things are going to get a little weird, but but now imagine if you said and you can with Claude and those you could give them, like I said, you can give it access to Twitter. You can give it access to Instagram. You can give it access to all these things like we're talking about the auto posting. And now imagine if. They just create their own account like Malt book, but now Malt book is Instagram. They're just putting shit on Instagram.

 

Yeah, and they came up with their own captcha to, you know, identify that you're not human. So it's like a capture like we have for bots and the other way for humans.

 

Yeah.

 

Alicia McKenzie (26:56.942)

Yeah, you have to identify if you're a human or a bot. Yeah. Like, that's kind of crazy. so thinking about all this and the use of AI and the amount of people that are going to start using these agents environmentally, what is that going to do?

 

Yeah, it all depends. think the optimist of me says these models are going to get better. Like, you know, deep seek, if that was accurate, like how they were able to train the model with a lot less time and energy expenditure. And, know, I think we're just learning how to do things better. the tokenization cost will come down and the energy requirement will come. Right. Well, when the cost to compute comes down.

 

Yeah.

 

Alicia McKenzie (27:31.886)

Mm-hmm.

 

Alicia McKenzie (27:37.752)

What about water consumption?

 

George McKenzie (27:42.402)

The electricity comes down, the water comes down, everything comes down.

 

I it's going to happen. The water is being used to cool.

 

Cooling and think and you Elon Musk is his goal is combining, you know, X AI and space access to build data centers in the sky. So now you don't have to worry about heat. Heat's no longer a problem.

 

This is just it's happening so fast and yet. So slow.

 

It's going to be an interesting world. it becomes, okay, as an entrepreneur, I think it could be a bounty. There's so much more productivity you can do that you used to be bound by labor costs. So your labor costs really inhibited your ability to scale and inhibited your ability to do things at a production level or enterprise level. And if you now can go to this agent and robot world, you can...

 

George McKenzie (28:37.358)

scale with nonlinear labor costs like the cost to have, you know, a CFO, the cost to have like these little agents of five virtual assistants running around doing particular tasks that there's no incremental cost, or very small incremental costs in terms of electricity that, you know, why wouldn't you? And now you can be an entrepreneur, a small company of five people, but you could produce the output of a 150 person company.

 

Yeah.

 

George McKenzie (29:06.606)

Because you have 100 agents that are doing things.

 

if you can afford it.

 

I'm just, but it's a lot cheaper than 150p.

 

Yeah, that's true. That's true. Okay, so the jobs that are going to be hit then what jobs are taking a hit, right? You've got warehousing, manufacturing, logistics, facilities, retail operations, right? Those are

 

I think that's the...

 

George McKenzie (29:25.698)

Yeah, I think those, those are the first ones that are to be impacted when it happens. I don't know, but it's already happening. Like Amazon laid off 16,000 people. And you look at these Amazon warehouse videos where it's just robots in there now running around. Right. And you think of what, you know, the Atlas and the optimist can do that are more humanoid that can start to do things that require more dexterity. And, know, I think it becomes a lot of the repetitive and dangerous jobs will be done by robots.

 

crazy.

 

Alicia McKenzie (29:47.362)

Mm-hmm.

 

Alicia McKenzie (29:54.84)

Yeah. And then the jobs that are transforming, right? You've got managers look more like you're just orchestrating all of this.

 

Fairly quickly.

 

George McKenzie (30:04.59)

Yeah, exactly. So your skill set is a little different and your skill set of managing people and managing. So let's say in a retail world or a restaurant world, you're managing, Hey, I know servers call out sick all the time. So I got to have backup for servers. I've got a stress test, you know, I have a limited amount of wages to go around. So I'm going to have to dynamically schedule a lot because I want, you know, a lot of people.

 

in my target from, you know, 9am to 11am. And I want a lot of people from five to nine. I don't need that many during so you have people that come in halftime. But if you had robots, who gives a shit? Yeah, you're there all the time. Right. All that goes away and it just becomes you know.

 

Yeah, absolutely.

 

Alicia McKenzie (30:54.7)

making sure they can get to their chargers.

 

They are making sure that that part is working correctly and that they need anything, but they'll figure out their own patterns. They'll optimize. And you look at the phase one here, we've already started. We've eliminated a lot of the cashiers in restaurants or grocery stores or retail by just scanning your own stuff. So if you had scanned your own stuff and then maybe you had an optimist or an Atlas that was scanning stuff if you wanted to go through and helping you load stuff in your car.

 

Yeah, which I actually prefer.

 

George McKenzie (31:26.292)

And like there's a lot of and restocking shelves. Like there's a lot of a lot of things it could do that humans are doing.

 

And then the engineers, right? Like they're system designers.

 

Yeah, like how do I want all this to work together? How would I orchestrate? How do I want to design the robots and then the AI agents? How do I want all that to work? And then coming up with those functional roles like this is what this agent does. This is the box around it. This is what I want it to do. And then I have another agent that does this. And then how do those agents talk to each other? How do I want to orchestrate?

 

Right, and that's where I was just going to go, right? If these agents have the ability to communicate with one another while they're in their robot bodies, are they going to learn and try to take over each other's roles?

 

No, because they are all going to have like, wouldn't say a personality file, but they're going to have a file that kind of sets the boundaries to what they do. And as long as they're not editing that, like, but it comes down to, it's just like a person, like a person that's free will to do whatever they want. Like just because you say, Hey, be a cashier. They could say, Nope, I'm going to go stock shelves. And what are you going to do?

 

Alicia McKenzie (32:21.705)

Right. Yeah.

 

Alicia McKenzie (32:26.368)

It's...

 

Alicia McKenzie (32:36.846)

Let's.

 

I mean you fire them, you reprimand them, but it's the same thing you do with robot. You're like, I'm going to turn you off. And I'm going to reboot you to last known.

 

You

 

Alicia McKenzie (32:46.914)

Yeah, right. And then the jobs that are going to grow in value, the things that we are going to actually value more is going to be like the leadership. Creativity. Ethics, the ability to envision what you want it to look like.

 

The innovation like what company am I creating? What product am I?

 

George McKenzie (33:04.354)

Yeah. And then I still think. Yeah. Entertainment, like those types of things, creativity, innovation, those are going to be things that are valued.

 

emotional intelligence.

 

Alicia McKenzie (33:13.838)

Are they gonna get rid of football players? Right.

 

I don't know if I could ever watch that because then they're all great. Like how? Yeah, like there's no.

 

Who wanted to create like an Olympics for people on steroids? The doctor, I was going to say Peter Attia, but I don't think that was him. And he I he's in some hot water. So for front of my mind. But no, there was somebody who saying they wanted was a Huberman. Maybe it was. Right, like, let's let's put all these guys on peptides, let's give them all steroids and then let's put them in a hydrox and see who does.

 

I can't remember, but I would love to watch that.

 

George McKenzie (33:35.968)

Yes, is.

 

George McKenzie (33:42.622)

I would love it. Let's just push the boundaries of mankind.

 

George McKenzie (33:52.522)

Yeah, but the one thing I want is I still want one lane reserved in the Olympics for an average person. I just want an average, it could be like a high school. He was a high school track star. Let's put him in there and see what happens.

 

Isn't that amazing?

 

Alicia McKenzie (34:05.358)

Right? Like, let's take an average hockey player and put a f-

 

Yeah, just so you can see and how great these guys really are without them getting killed. There was a show a long time ago. It was called Joe's versus pros. Yeah, I think it only did a season or so. It did not do well, but I could it could do well, but I think they had like retired pro athletes and it wasn't like full contact. But it could just.

 

yeah.

 

Alicia McKenzie (34:19.916)

I feel like I remember that. Didn't take off.

 

Alicia McKenzie (34:29.869)

Yeah.

 

But you draw the line at robot football.

 

I do or sports. think robot sports in general, because there's no. The challenge with sports is the human nature, like people make mistakes, people miss. Yeah, like you've seen there's a robot out there that shoots three pointers and just doesn't

 

That's no fun.

 

And they all just, they're all the exact same shot because they do the exact same thing every time, which makes it great for production lines or repetitive tasks. Like it can do the exact same thing 24 seven without complaining.

 

Alicia McKenzie (34:55.896)

can

 

Alicia McKenzie (35:05.518)

I saw a video of Carlos Alcaraz working on his serve, like his toss, and the coach had a little basketball hoop on his.

 

I saw that. Did you see did see that.

 

Right. And like he, he on his toss was able to like nail it through the basketball. Yeah. I'm like, that's so impressive. Have you ever like tossed a tennis ball to try to hit it?

 

But the beauty of it, I he missed a couple, they would hit the room. But imagine a robot would do it in the hole every time. Yeah, exactly. So I think there'll still be obviously roles for humans and jobs will evolve. The problem you have with any revolution or evolution is there's a window of time where we haven't re-skilled yet. So if robots are taking over production jobs, what do you do with that workforce until you re-skill?

 

Yeah.

 

Alicia McKenzie (35:29.816)

time.

 

Alicia McKenzie (35:44.889)

Mm-hmm.

 

Alicia McKenzie (35:54.318)

Tis the question.

 

Yeah, same with like we were talking about for if an AI agent can be an assistant, what do do with all the people that are assisted?

 

today. Yeah. Learn new skills. Mm hmm. Okay and then what do you like what do we teach our kids?

 

You have to.

 

George McKenzie (36:12.59)

Yeah, I think it's those skills that leadership, be creative, be innovative, cooking or, being, you know, I think I still think developers, not like code developers, but understanding how to orchestrate and, you know, make things happen, understand AI and at least how to interface with it, how it functions. Yeah. Communication. That's a harder skill.

 

Okay.

 

Alicia McKenzie (36:32.258)

Yeah. Critical thinking skills. Communication in person communication. Yeah. Yeah. So and I'm looking at this right now, but it says repetitive execution, like memorization.

 

That's not a skill that's going to be required. we even see it today. You know, memorizing is used to be a great skill. You used to have to have it because you needed to retain the knowledge to be able to execute something. Now all the knowledge you could ever want is in the palm of your hand at any moment. And then you keep moving with wearables. It'll be available to you at any moment. So why memorize?

 

Yeah.

 

Alicia McKenzie (37:14.58)

What was the exercise they did with you in high school? And you told me this story recently and I'd never heard it before.

 

yeah. So we did a lot of experimental cause I think they got like grant money or something. So the school would do a lot of these and we had the kind of accelerated students. There was like six or eight of us. And one of them I remember was they had this back when laser disc route. they had, it's like a pre CD before CDs that held a lot of. No, no, it's like a big steel.

 

It was experimental.

 

Alicia McKenzie (37:42.38)

It's a laser bit.

 

Alicia McKenzie (37:47.8)

Is that the floppy thing?

 

vinyl record. anyway, so they had that playing a movie, right? And at the same time that movie was playing with audio, the teacher was teaching. So you had two input, both audio and visual coming at you at the same time. And you were trying, so you just had it. So you ended up focusing on one and the other was kind of background. And then afterwards you had a quick quiz. There's only like

 

Okay.

 

Alicia McKenzie (38:16.792)

Yeah.

 

George McKenzie (38:20.91)

15 minutes of this and they have a quick quiz that was based on questions from both.

 

Yeah, and you did really well on the quiz for both.

 

I did. And then they kept doing those and you got better at it. Yeah, you got better at like. Yeah.

 

buffering. I think so. But that was a skill that you had when we first met. You're like, so I'm wondering if we used all your brain during these experiments. Right. Because when I first met you, you used to be able like you were typing on your computer and somebody was having a conversation behind you. And then they would ask you a question and you would answer but never break your typing.

 

I was really good at buffering, now I'm not so much.

 

George McKenzie (38:42.862)

Maybe I used all the mufflerings.

 

Alicia McKenzie (39:00.012)

And I'm like, wait, what is this anomaly of a human? Like, this isn't normal. The ability to listen, but still produce is. Yeah, that that's not easy to do. Yeah. That was something you used to be really good at it. Not so much anymore.

 

That's probably because of skill I don't have to do as much.

 

Maybe. Yeah. Yeah. You don't you.

 

Use it, you lose it. 40-0 version taught us. If you don't use it.

 

man. OK, well, I mean, I'm equally excited and terrified to see where this goes. OK, so rapid fire. no. The first robot dominated industry.

 

George McKenzie (39:41.286)

robot dominated industry. I would say production lines like automobile production.

 

Yeah, I think it's there. That's a lot of. Yeah, one skill that every kid needs now, like if you have if you're listening to this podcast and you have children. Yeah. What's a skill that they should be focusing on?

 

coming more and more.

 

George McKenzie (40:06.126)

I would say two. One, think leadership because that skill needs to persist and the other is general AI understanding. Like I would I would not be afraid of it. I think they have to understand it.

 

Yeah.

 

Alicia McKenzie (40:19.032)

So do you think we need to teach our kids?

 

I think they already have some level of understanding because they understand they know what Alexa is. Alexa's an AI. Right. Yeah. And it's going to get better and better because it's getting exposed to more and more. So they understand it. I think what we'll get when they're a little more competent is how does it work?

 

Okay.

 

Alicia McKenzie (40:39.31)

One founder mistake today.

 

With AI. I think it's back to the leadership and the system design. You have to know what problem you're trying to solve and know the functions you want to outsource. Right, so just think a lot of people just lean into AI because AI and they're like, I bought a. Access to copilot or chat GPT and I have this or yeah, I want to do that and.

 

Yeah.

 

George McKenzie (41:10.446)

You know, maybe I'm still using Zapier and I'm automating things and you're like, all right, well, those are all parts of it. But until you understand what it is you're trying to accomplish, just like you would if you built out your own team. Like if I'm building an org structure, what do I want? And then if I have, you me and, you know, five agents, I can outperform a 20 person team. If I understand exactly what function, what role I want each of these agents to do and tightly bound what it is they're doing.

 

Yeah.

 

George McKenzie (41:38.872)

and give them all the access they need to complete their test.

 

One thing parents should stop teaching.

 

George McKenzie (41:49.837)

coding.

 

Ooh, that's interesting.

 

I think you should understand it, learn to code, but you don't, it's not a skill that is going to persist too long. Its ability to code is clawed and some of these are really, really good at code.

 

That's true because

 

Remember when we used to take the kids to Coder Kids?

 

George McKenzie (42:06.668)

Yeah. Yeah. I think those skills are going to be less required. Like the memorization of understanding how to code in C plus plus, or how to code in Java or how to code in Jason. But now it's going to be, I just need to conceptually know what I want to get accomplished front to back. And then, you know, the AI will spit out the code and is it perfect today? And I know a lot of people say, spits out shitty code. And yeah, it does, but it's going to get better. And

 

Yeah.

 

George McKenzie (42:35.746)

there'll still be a need for some software developers, but not a bunch. And it's going to be more at the tail end of the CI CD pipeline of reviewing code and you know, red teaming code and doing those types of things versus punching out hundreds of lines.

 

Yeah.

 

Alicia McKenzie (42:52.098)

What about robotics?

 

Yeah, I think robotics is another big skill and it's more of that mechanical engineering.

 

Yeah. And the only reason I brought that up is because our son was invited to a robotics club. And apparently it's a big commitment, one that I'm not willing to take. So we're not doing it.

 

No, think robotics is a big one because there's still going to have to be people until a certain point to fix robots. But I think eventually you'll have a robot that fixes other robots.

 

One thing that schools are behind in right now.

 

George McKenzie (43:21.518)

I think AI.

 

So.

 

Most right now, the initial knee jerk was we're banning it. No one uses it. And then we need AI to detect AI. And I think what they have to do is is just like when I was in school, we were just leaning into programming and coding. And it was a small group of people that were doing it. I think you have to have the same with AI. Like we have to start embracing it and understanding how it works. How to live with it in a world with it. How do you leverage it?

 

Yeah.

 

Alicia McKenzie (43:45.676)

Yeah.

 

Mm hmm. When our daughter was in English class in high school, she they would not let them work on their assignments outside of school because they were so worried that everybody was using chat GPT to produce their papers. Right.

 

And I think if you think about the goal of school is to create humans that are productive in society, right? That are knowledgeable and productive. And in today's world or tomorrow's world, you're not going to be a productive member of society if you cannot leverage and coexist with AI. That is going to be a skill that is viable. Like if you don't understand how to leverage it to do the job, then you're going to be behind because no one's going to want to hire 20 people.

 

Yeah.

 

Alicia McKenzie (44:28.204)

Yeah.

 

George McKenzie (44:32.832)

in accounting, right? They're going to have a CFO, then AI agents will be doing a lot of those functions.

 

Absolutely. And I think just one final parting thought is that just because you can automate something does not mean you should. Yeah, agreed. there has to be some limitations on what you as a human are willing to automate. And I'll give you an example. Let's say you're having an argument within your workforce, right? Or you're having major disagreements. Do not run

 

Exactly.

 

Alicia McKenzie (45:08.578)

write a chat GPT and put in the argument and tell it to spit out a response. I think you need to learn how to work through hardships and have those hard conversations before you try to automate a response.

 

And yeah, Chad GPT is not a therapist. AI is not a therapist. I mean, it has access to all the same materials and reading and knowledge that a therapist would have, but it doesn't have that empathy. doesn't understand the human condition. It doesn't see or feel or know that when I say like people are really good. I'm not sure AI is going to ever get there. People are really good at knowing exactly what to say.

 

relationships.

 

It's no-

 

George McKenzie (45:52.684)

Yeah. elicit a reaction.

 

yeah. Right? Yeah.

 

Yeah, don't look at me. I know you're perfect at it. So I don't think AI will get to that spot because they can't the emotional cues and the understanding of how it works.

 

Yep. It's OK, current husband.

 

Ha ha ha, real fun. Yes, yes, Yes, yes, yes, Kurtis.

 

Alicia McKenzie (46:13.518)

My current husband has strong artificial intelligence. All right. Are we done?

 

I think so.

 

Thank you for tuning in to Mary to the startup. We hope you enjoyed today's episode. If you did, please take a moment to like, rate and subscribe to our podcast. Your support helps us reach more people and keeps the conversation going. If you have any questions or topics you'd like us to cover, drop me a message. I love hearing from you guys until next time.