Episode 8 AI and the future of work

Four pressure points define the landscape: how much energy AI consumes, the cost of hallucinated outputs, the fragility of agents in real-world tasks, and the gap between broad deployments that underdeliver financially and narrow, industry-specific uses that do work.

From drive-thru ordering failures to an AI-run kiosk experiment that spirals into giveaways and overspending, the limits of automation show up in expensive ways—especially when accuracy matters and humans have to re-check the work. The conversation also connects AI’s energy appetite to climate risk, property insurance, and the possibility of a WWII-style “all hands on deck” response that could reshape employment for decades. Along the way: capitalism, business cycles, “bullshit jobs,” basic income, and why nuclear—reframed as “elemental energy”—may return as a practical answer to rising demand.

Episode Show Notes

• Predictions about AI are uncertain, but its impact on work is expected to be significant.
• Four key attributes discussed: energy consumption, hallucinations, weak financial returns from broad deployments, and stronger ROI from narrow applications.
• AI as a research assistant and the importance of verifying claims with sources.
• “Agents” in practice: failure cases and the risk of letting systems act autonomously.
• Layoffs framed as partly economic and investment-driven, not purely AI-driven.
• “Bullshit jobs” and why employment can persist even when roles feel unnecessary.
• Climate change as an existential threat and how large-scale mobilization changes labor markets.
• Nuclear power and small modular reactors as a potential response to growing energy needs.
• Language, translation, and the idea that AI systems shaped by English may not be the only path.
• Operational risk: AI summaries with meaningful error rates and the cost of validation.

Episode Timestamps

00:00:00 Intro and topic framing: AI and the future of work
00:00:58 “Predictions are hard” and why impact is certain but outcomes vary
00:02:16 Four attributes: energy use, hallucinations, ROI challenges, narrow wins
00:03:46 Faraday/electricity analogy and delayed, world-changing effects
00:06:36 Extremes of the future: full employment vs reduced need for work
00:07:56 Big government, business cycles, and Hyman Minsky’s view
00:10:54 Will AI take jobs? Using AI as a research assistant with citations
00:12:20 Agents and reliability concerns
00:13:35 “Taco Bell” drive-thru ordering failures and automation mistakes
00:16:09 AI kiosk experiment with Anthropic/WSJ and unintended outcomes
00:19:50 Shifting work onto customers (self-checkout and customer friction)
00:20:49 AI energy consumption and abandoned net-zero targets
00:21:43 Climate risk, insurance markets, and cascading economic effects
00:24:20 Food systems, famine risk, and compounded disruptions
00:27:29 WWII-style mobilization as a template for existential response
00:31:18 Workflows that now depend on AI and what changes next
00:32:36 Layoffs, investment tradeoffs, and “bullshit jobs”
00:41:22 Why this shift may be different—and uncertainty about adaptation
00:45:33 Nuclear power, Project Pele, and small modular reactors
00:49:07 Quantum computing mention and future energy requirements
00:52:18 Geopolitics, AI “poles,” and language-first model assumptions
00:54:44 Translation, Esperanto, and meaning loss vs standardization
00:56:44 Costly AI errors: summaries with unknown 10% inaccuracies
00:59:06 Closing

About the Podcast

Hosted by Kevin Carney and Emanuel Petrescu, two curious minds exploring ideas, culture, and everything in between. Curious Pundits is a conversational podcast where each episode starts with a topic that caught our attention and unfolds into thoughtful, unscripted discussion. We follow curiosity wherever it leads, across disciplines, opinions, and perspectives, without pretending to have all the answers.

Listen to the Curious Pundits Podcast on Your Favorite App

Website: https://curiouspundits.com/
YouTube: https://www.youtube.com/@curiouspundits
Podbean: https://curiouspundits.podbean.com
Apple: https://podcasts.apple.com/us/podcast/curious-pundits/id1874614249
Spotify: https://open.spotify.com/show/4nvSoNRrgPPBkdZiLGanYh
Amazon: https://music.amazon.com/podcasts/26ab96d9-ab37-4369-bccc-fe2cf937f950
iHeartRadio: https://iheart.com/podcast/321051634
Soundcloud: https://soundcloud.com/curiouspundits

Follow the Curious Pundits on Social Media

Facebook: https://www.facebook.com/curiouspundits/
Instagram: https://www.instagram.com/curiouspundits/
Threads: https://www.threads.com/@curiouspundits
YouTube: https://www.youtube.com/@curiouspundits
LinkedIn: https://linkedin.com/company/curiouspundits
TikTok: https://www.tiktok.com/@curiouspundits
X: https://x.com/CuriousPundits
Pinterest: https://www.pinterest.com/curiouspundits/
Reddit: https://www.reddit.com/user/curiouspundits/
Medium: https://medium.com/@curiouspundits
Tumblr: https://www.tumblr.com/blog/curiouspundits

About the Podcast

Hosted by Kevin Carney and Emanuel Petrescu, two curious minds exploring ideas, culture, and everything in between. Curious Pundits is a conversational podcast where each episode starts with a topic that caught our attention and unfolds into thoughtful, unscripted discussion. We follow curiosity wherever it leads, across disciplines, opinions, and perspectives, without pretending to have all the answers.

Entities

People and works mentioned

• Niels Bohr
• Yogi Berra
• Michael Faraday
• Karl Marx
• Hyman Minsky (Financial Instability Hypothesis)
• David Graeber (Bullshit Jobs)

Companies, tools, and terms mentioned

• OpenAI ChatGPT (custom GPTs)
• Anthropic Claude
• Wall Street Journal
• Taco Bell
• UPS
• Amazon
• Google Translate
• Mistral
• Esperanto

Other references mentioned

• Demolition Man (film)
• Project Pele (small modular reactor program)
• Ecclesiastes (biblical reference)

Transcript

[00:00:00] Emanuel: Hi there and welcome to another episode of the Curious Pundits Podcast.

My name is Emmanuel.

[00:00:06] Kevin: My name is Kevin.

[00:00:07] Emanuel: And if you are not following us already, go to curiouspundits.com, and follow, share, and leave a note about what you think about our podcast.

Today’s topic is something that Kevin proposed, and I completely agree because somehow it feels very timely, one might say.

AI and the future of work.

What did you have in mind Kevin, when you come up with this topic?

[00:00:37] Kevin: A free rambling conversation that probably goes on forever and ever. In fact, the challenge of this topic is to be concise on a topic on which there are… there’s just a lot to say.

[00:00:56] Emanuel: We have all the answers though.

[00:00:58] Kevin: Oh, no, not at all.

In fact, that’s like the opening position is… I would never presume not just to have all the answers, but to necessarily have correct answers.

At this point, I’m going to channel, or probably more accurately quote, two people who at first glance seemingly have nothing in common.

Physicist Neils Bohr and baseball player Yogi Berra are both quoted as having said, making predictions is hard, especially about the future.

[00:01:36] Emanuel: Fair enough.

[00:01:36] Kevin: And I think the only definitive statement that I’m willing to make about AI and the future of work is that the impact will be significant.

[00:01:49] Emanuel: It already is.

[00:01:49] Kevin: But exactly what that impact is going to be… It’s still unfolding and it’s too early to tell.

[00:01:57] Emanuel: Yeah, I think it already is here, the impact of AI and the future of work… for some more, for other less.

[00:02:08] Kevin: I agree with that as a general statement, but I also believe that we’ve only just started.

And…

[00:02:16] Emanuel: Yeah.

[00:02:16] Kevin: I’d like to say there’s… In my mind, and again, I’m more of a student of this than a teacher, in my mind there seemed to be four significant attributes, or whatever the right word is, of AI that we need to consider.

One is how much energy it uses.

[00:02:41] Emanuel: A lot.

[00:02:42] Kevin: One is the fact that it makes stuff up and there are certain fields in which that’s just not acceptable.

[00:02:50] Emanuel: Many.

[00:02:51] Kevin: The other is that in the current deployments of AI, generalized deployments of AI by companies are not providing the expected financial return that these companies are looking for.

[00:03:08] Emanuel: It’s important to mention that the date we’re recording, this is December, 2025.

[00:03:13] Kevin: Yes. But the last piece is that very narrow applications of AI, very narrow industry specific applications, are working out financially.

So in terms of all of that, what does it mean? And, each of those four areas could potentially be an entire episode.

And we’ve got to try and squeeze this into 30, 40 minutes. So where do you want to start?

[00:03:46] Emanuel: I will start with reminding you of an anecdote you shared on a different show when you are a guest… about Faraday and electricity and the comparison.

[00:03:57] Kevin: Yeah. Yeah.

[00:03:58] Emanuel: And if you want to tell our audience what that anecdote was, it’s really short and that’s a starting point as well for today’s topic.

It’s still a valid, although one year and six months have passed since you made that statement, since you shared that anecdote.

[00:04:15] Kevin: So the anecdote in question has to do with the harnessing of electricity and the way it basically changed everything. I’d have to look up exactly when this happened, but the guy who initially figured it out was a guy named Michael Faraday, and he lived during Victorian England. Sometime between the… he lived longer than this, but he did his work in the 1830 to 1850 timeframe, I think.

And by modern standards, the generators that he invented and the motors that he invented… we would look at them as toys today. But they were basically a proof of concept.

Now in his day because there was no electricity, there were no movies, there was no television, there was no radio.

So one of the things…

[00:05:10] Emanuel: No podcasts, no internet, no nothing.

[00:05:11] Kevin: No podcast, no internet.

So one of the things that people would do for entertainment is they would pay money to go to a public lecture.

And he, amongst other people, would give these public lectures and he would give demonstrations.

And there’s a famous incident when he gave his lecture, he did a demonstration, opened it up for questions and answers.

And the famous incident is a lady in the audience said… that’s all very interesting, but does it have any practical use? And very honestly, he said… no, because at the time that question was asked, it didn’t.

Electricity or more accurately, the harnessing of electricity is something that we figured out in the 19th century, but it didn’t change the world until the 20th century.

And I have an inkling that AI is on a similar trajectory, that we’re going to find all kinds of things to use it for in the short term, but the lasting permanent changes that are going to come, we don’t know what they are.

They’re going to happen in a while. It may not take 80 or 90 years like electricity did. But it’s just, it’s too early to say.

We’ll see.

[00:06:31] Emanuel: And I agree on this one. You and I agree.

[00:06:33] Kevin: Yeah.

[00:06:33] Emanuel: On this one.

[00:06:35] Kevin: But the…

[00:06:36] Emanuel: So where are we now?

[00:06:38] Kevin: I was going to say, but the potential outcomes on AI in the future of work are literally across the spectrum.

There’s one possible outcome where it’s all hands on deck, full employment, nobody gets a day off. And there’s a another possible outcome where we don’t need most people to work, and AI and automation ultimately validates Karl Marx’s primary thesis, which is that capitalism will ultimately undo itself.

Like these are the extremes that I see.

[00:07:17] Emanuel: I was laughing because I was hoping that if that scenario will unfold, maybe they’ll start with the government employees or the government agencies.

[00:07:30] Kevin: You mean laying people off?

[00:07:31] Emanuel: About not needing people to work in the government or in some agencies?

[00:07:39] Kevin: As I read a lot of the history of economics and all that kind of stuff, and that’s actually not as desirable a future as people think.

[00:07:50] Emanuel: I just don’t, would like not to interact that much with them.

[00:07:56] Kevin: Yeah, this is probably a topic for another episode, but I’ll try and do this as concisely as I can. But the era of big government started during and after the Great Depression and prior to the era of big government when money was almost exclusively created by commercial banks, the monetary expansions and contractions were incredibly frequent and incredibly painful.

So the economist who I think has this dialed in the tightest, is a guy named Hyman Minsky. And he has this idea called the Financial Instability Hypothesis.

And basically he was able to, as far as I’m concerned, demonstrate that the business cycle… booms and busts… is an inherent part of capitalism. It has nothing to do with external shocks. A production based economy simply goes through periods of expansion and contraction.

And prior to the era of big government, we used to have pretty significant economic contractions every four to eight years.

For whatever reason, we’ve forgotten about that.

So what the banking sector does is procyclical. It makes booms bigger and it makes busts bigger.

So big government is countercyclical. So if we shrink government moving into an era of much less employment, it’s probably not going to be good for a variety of reasons.

[00:09:43] Emanuel: There’s a season in everything as the ecclesiastes teaches us, and I will call them seasons, but not cycles and compare them to nature.

It’s not necessarily before the first or second world war or before, whatever you might call big government. I don’t think that the Roman Empire or Alexander the Great Empire or any of the Persian empires would not be big governments as well. It is just that the technology obviously didn’t allow for such control and communication.

In such a timely matter to be able to move things forward faster and at a bigger scale. But I believe that it’s little too close to none in any ancient empire. So big empire or whatever you, we might call today, big government, but we do have this tendency to deviate. The topic of today is AI and the future of work.

So let me ask you this, are you scared that AI will take your job, Kevin?

[00:10:54] Kevin: It depends in what direction it takes us. And, again it’s too early to tell.

Now I’m a pretty big user of AI. I use it in a very specific way. I use it as a research assistant.

So I also have the experience of AI… I’m pretty comfortable is not telling me things that aren’t true because my use of AI is every time you make an assertion, provide a citation that I can follow to verify that assertion.

And, I literally… I’m a recent subscriber to Open AI’s ChatGPT, because I like the custom GPT capability.

[00:11:44] Emanuel: Welcome, welcome to the club.

[00:11:46] Kevin: Yeah, because I didn’t realize the power of a custom GPT. Which is basically… for people who don’t know, you can define a custom GPT, give it a very detailed set of instructions, and then every time you use it, you don’t have to repeat those instructions.

It’s embedded in the custom GPT.

[00:12:07] Emanuel: The closest thing to the agents that everybody’s talking about and we dream of helping us. But they were promised two years ago and yet we still do lot of manual…

[00:12:20] Kevin: Agents is one of the problems with AI currently, and I would like to talk about that.

But in terms of a custom GPT, in my personal… I call it like a financial historian or a financial historical researcher. Embedded in the instructions is every time you make an assertion, provide me for a link where I can verify it and it now just does that. And I click to the links. Like I don’t trust AI to tell me things that are true.

So I want to verify where the information came from and then I can personally assess the legitimacy of that source. So I’m pretty comfortable by virtue of me instructing it to do that, that it’s not lying to me.

[00:13:02] Emanuel: Only if you go deeper as I do East European… you may think that AI could potentially create a website just to cite a source for your query.

[00:13:15] Kevin: But can I talk a little bit about agents now?

Because there’s been a couple of like glaring incidences of… “What just happened?”.

[00:13:25] Emanuel: Yes. By the way, I call my agent Smith, all my agents, I call them Smith.

[00:13:32] Kevin: I take it, that’s a matrix reference.

[00:13:34] Emanuel: Yes, sir.

[00:13:35] Kevin: So are you aware of the… I’m going to call it the Taco Bell fiasco.

[00:13:41] Emanuel: That’s the fast food chain restaurant, right?

[00:13:46] Kevin: Yeah. So it’s…

[00:13:47] Emanuel: No.

[00:13:48] Kevin: I’m sure they have Taco Bell in Toronto, don’t they?

[00:13:50] Emanuel: Yeah. So funny enough, I never ate at them. I know about Taco Bell since the Demolition Man. If you know the movie with Sylverster Stallone.

[00:13:58] Kevin: I actually have not seen that one.

[00:14:00] Emanuel: Yeah, it’s futuristic movie from 92 with Westley Snipes, Sylvester Stallone, and a few… Sandra Bullock, and a few others.

Anyway it said that there was… it’s happening in the future, I believe around our time and all the restaurants were named Taco Bell because eventually the big California earthquake happened. It destroyed everything. Almost everything. All only Taco Bell restaurants were left untouched or kinda like surviving.

Then they named all the restaurants Taco Bell.

But coming back to your…

[00:14:35] Kevin: Last man standing.

[00:14:36] Emanuel: Yeah.

[00:14:37] Kevin: So Taco Bell corporate decided that they wanted to reduce staff by having an AI agent take orders at the drive-in and it just didn’t go well.

And there’s… you can find all kinds of video shorts online of people trying to order stuff and the AI agent just not getting it right.

And there’s one where the guy says… and I would like a large Mountain Dew. And the AI says, what would you like to drink? A large Mountain Dew. What would you like to drink? A large Mountain Dew? What would you like to drink? 18,000 glasses of water. No problem. 18,000 glasses of water coming up… right.

[00:15:21] Emanuel: Yeah. I guess some…

[00:15:23] Kevin: Somebody… and somebody inside Taco Bell… no, this couldn’t have been Taco Bell because they don’t have ice cream, but apparently… I don’t know… maybe they do someplace. I don’t know. But this must have been another company.

But somebody ordered an ice cream and the instructions were like to put a piece of bacon on the ice cream.

So the guy inside did it and handed it out to the guy, and the guy in the car was like… what is this?

And then my other favorite story was an experiment, and I think it was Anthropic. Anthropic wanted to run an experiment on having an AI run a small business. So they partnered…

[00:16:09] Emanuel: Claude… Anthropic is… are the… is the company that created Claude.

The…

[00:16:13] Kevin: Yeah.

[00:16:13] Emanuel: …LLM Claude.

[00:16:14] Kevin: And I think it was Anthropic. I could be wrong on which company it is, but I’m pretty sure it was.

So they partnered with… a team of journalists at the Wall Street Journal. Presumably it was the journalists who write about technology and AI, and they basically installed a kiosk in their office.

Now, part of the experiment was that the AI doesn’t actually dispense products. So there was a certain amount of honor system, but don’t take anything that the AI didn’t tell you that you couldn’t take. So basically it’s a refrigerated cabinet with a computer on the side, and the AI runs on the computer.

They gave it a thousand dollars and the AI had the ability to actually go out on the Internet and order stuff. And people are basically interacting with the AI and telling it what they want. So it’s ordering like Coca-Cola and Doritos.

In a matter of days, they like turned it completely communist or it’s ordering stuff, giving them away for free and referring to people in the office as comrade.

Now, again, this was an experiment and these people are intentionally trying to break it, but it didn’t take very long.

Oh, and then they got it ordering like wine for religious functions. And it even ordered a live fish for some reason because somebody felt that they needed a live fish.

[00:17:43] Emanuel: Oh, all this under a thousand dollars?

[00:17:47] Kevin: Oh no, it actually spent, $2,000.

[00:17:49] Emanuel: Oh, okay.

[00:17:50] Kevin: So it started with an initial budget of a thousand dollars. Initially it was charging a dollar for a bag of chips or whatever, and pretty soon it’s like free snacks for everybody. And it continued ordering stuff and spending money and when it was a thousand dollars in the hole, they basically shut it down.

The experiment had run its course. The people at Anthropic are probably back to the drawing board and they’re going to try it again later.

But agents are not necessarily reliable just to let them loose and do things on your behalf. You don’t know what kind of stuff they’re going to do.

[00:18:26] Emanuel: Again, at the moment, because…

[00:18:28] Kevin: At the moment.

[00:18:28] Emanuel: Basically we know we’re recording this in December, 2025.

And it’s all funny and we laugh and it about this, but I couldn’t picture six years ago, for example, actually wanting to go to the kiosk and order or pay for stuff myself. The system, the payment system and the self checkout system didn’t work at a shop where I do… the next to me. So we had to pay in cash.

First of all, most people didn’t have any, but it was such a huge line and it was taking forever that we actually forgot that was the norm just before the pandemic, which is like a couple of years before that, we actually forgot about that. And I’m still, I actually put on my resume to at, job descriptions, cashier and I list all the major businesses that allow self checkout essentially because for the past couple of years, I developed some skills that I didn’t knew I had in me, right? How to fast check out how to put the right things in the right, place, how to move fast, all those things. So that’s a skill in itself, right?

[00:19:50] Kevin: Well this is potentially a topic for another episode as well, but corporations at scale are pushing responsibilities onto customers that customers never used to do. That’s just simply one example and it’s causing frustrations with customers, but what are you going to do? Not buy your groceries.

So we’re, accepting of these things even though they irritate the hell out of us.

[00:20:19] Emanuel: Yeah. And also in a sense, as I realized, it definitely has makes small improvements in our lives, but as you said, that’s a topic for another conversation. And there’s so many things that I want to discuss also about today’s topic, but I do want to give the opportunity to you to finish, to share all the ideas.

[00:20:49] Kevin: And I’m never going to be finished with this topic, not for God knows how long. But I think something that is not being discussed enough is how much energy AI is consuming. I read just this morning that it currently consumes 4% of all the energy generated in the United States.

And that’s pretty significant for… considering what we’re doing with it. And when you also consider that, like five years ago, all these tech companies had these net zero carbon targets. And then AI happened and that’s just gone.

[00:21:33] Emanuel: Forget about it. AI and the new administration that also had different views that helped.

[00:21:43] Kevin: But, and here’s the big but… this is a… I was going to say dystopian, and that’s not a wrong word, but it’s not necessarily as terrible an outcome as it sounds.

AI consumes a lot of energy. We’re just committed to producing more energy. We’re doing it in such a way that it releases more carbon into the atmosphere. If the climate scientists are right, and I suspect that they are, we are heading towards… the word I’m going to use is a reckoning.

So as the atmosphere warms extreme weather events are more intense and more often. And they are starting to destroy things.

So the canary in the coal mine is property insurance. And apparently Florida is already, because it’s hurricane country.

[00:22:43] Emanuel: Yeah.

[00:22:43] Kevin: Florida is already suffering from… insurers saying we’re not going to reinsure that house for what, I consider to be like sound business reasons.

So if the private insurance companies back away, if you don’t have property insurance, you don’t have mortgages. And if you don’t have mortgages, you don’t have a mortgage market. So we can’t just tell everyone in Florida, sorry, your home is now worthless, you’ve got to move. So who’s going to step in and provide that insurance?

The state in some capacity. That’s who always bails out private industry when there’s a large failure.

[00:23:24] Emanuel: But if you, but you need to change code, you need to change a lot of things. And for those people who are listening outside of North America… houses in North America are not built like in the rest of the world.

[00:23:36] Kevin: Oh, they’re very flimsy by comparison. Yeah.

But if you take this to a… extrapolate this into the future, at some point, the buildings that are being destroyed by extreme weather events are factories where we build stuff, right? Are buildings where we process food, our fields, these aren’t buildings, but fields where we grow food.

So if you extrapolate this out far enough into the future, we’re going to at some point experience a significant enough loss in production that we’re going to collectively say, wow…

[00:24:19] Emanuel: Enough.

[00:24:20] Kevin: We need to do something. And at that point, we’re going to see it as an existential threat and… the… probably the most recent existential threat that we’ve responded to is World War 2.

So that’s probably the playbook that we would use.

[00:24:43] Emanuel: I would argue that’s actually the biggest threat I was reading… and some people confirm… that used to travel in Asia a lot for the past, I know, 20 30 years, that for the past five to six years they’ve been experiencing more and more changes in the climate turbulences.

[00:25:02] Kevin: Yeah.

[00:25:02] Emanuel: With the transportation, all those things in areas that I know from Thailand to Vietnam or some, I don’t remember exactly which routes, but they were like, like the safest, the smoothest ever. And now they’re experiencing all sorts of things that weren’t there. And that’s a direct effect of the change, the climate change that is happening right now.

And to the other point that you just mentioned, probably one of the biggest threats. For us humans right now is a famine.

It always has and it always… at least for the near future will be, unless we become the food again, matrix reference… the energy.

But things always happen together in order to have a cataclysmic event man made, with nature’s influence, there’s all, there’s never just one thing. There’s a bunch of things, right? Because there’s war here and there. There’s a economic downturn here and there. There’s a new technology disruptor here and there. Yeah. All those things. There’s a pandemic here and there. Could you imagine what would have happened if we had ChadGPT in during the pandemic?

But that’s a topic for another conversation in itself. Maybe we can hallucinate a little bit and see things that could have happened. But most importantly, when all these things are kind of like aligned, then one of the factors, either the war or either nature creates something that push people to starve… food.

I live in Canada. Saskatchewan is a province. It’s a state here. That’s the granary of many places in North America, not just Canada, but in the US as well. So if just for a couple of months something gets delayed. I’m not saying that you don’t have a crop or anything like that, just delayed or I don’t know, anything can happen, right?

Weather war trades, economic malfunction of a machinery or something, even stupid. Then we’re, in big trouble, and not just, oh my God, I can’t send an email or stuff like that. No… existential threat. I’m not I think they call it preppers, dooms day preppers. That’s the term they use.

Not necessarily.

[00:27:29] Kevin: Yes, but that’s not where I was going. Like World War 2, which I think would probably be the template for any kind of a response to an existential threat response. It was catastrophic, but it didn’t last forever and it didn’t diminish us and, bizarrely it had a very positive economic impact.

Which I know sounds weird, but consider for a moment that during the Great Depression, everybody was broke. Massive unemployment, massive homelessness.

And then this was from an American perspective, Canada joined the war earlier, but the Japanese bomb Pearl Harbor, and all of a sudden there’s no shortage of money.

Now, where did that money come from? Did all these broke people pull out other mattresses and give it to the government and say, we are now patriotic Americans, let’s go to war.

[00:28:27] Emanuel: The government took some.

[00:28:31] Kevin: No, the government makes money. The commercial banking system makes money. Money, and I don’t mean profitable… produces money.

That’s where it comes from. It comes primarily from the government, the central bank, and the commercial banking system.

[00:28:45] Emanuel: I was…

[00:28:45] Kevin: Again, that’s for another episode.

[00:28:47] Emanuel: Taking people’s gold essentially, right?

[00:28:50] Kevin: No. This is probably something else for a another episode, but…

[00:28:59] Emanuel: Like hydra, you cut ahead and seven other heads come…

[00:29:04] Kevin: Right.

People don’t have a good understanding of like, where is money produced? And the answer is in banks. So central banks produce money. Commercial banks produce money. So going into the mobilization for the war, banks started producing money. Money is an economic lubricant. We work for money and we tooled up to fight the war.

And the fact that we were broke like two years ago was irrelevant. We had an existential threat to deal with and it was all hands on deck and let’s deal with it.

So if we get there with climate change, and I’m pretty comfortable that the odds are pretty high, that we will at some point suffer a loss that we can’t deal with and we will do the five alarm fire, all hands on deck mobilization.

There’s going to be so much to do that there’ll be no unemployment for the next 30, 40, 50 years, because we just need to like deal with this.

[00:30:09] Emanuel: You think this will happen in our lifetime? You and I.

[00:30:13] Kevin: Hard to say. The short answer is maybe. Now you’re a little younger than I, so there’s more of a chance that it’ll happen in your lifetime than my lifetime.

It’s possible.

[00:30:28] Emanuel: Everything is.

[00:30:29] Kevin: It’ll definitely like I’m old enough to have grandchildren. I think it’ll definitely happen in their lifetime.

[00:30:35] Emanuel: Unless we somehow, and there’s efforts put towards this too.

[00:30:42] Kevin: Yeah, but we’re not…

[00:30:43] Emanuel: Colonizing other places.

[00:30:45] Kevin: Yeah, but there’s issues with that as well.

That’s another episode…

[00:30:48] Emanuel: So far, right? Yeah.

[00:30:50] Kevin: So that’s like the most dystopian possibility in terms of AI in the future of work, but it’s not… it would be a extremely painful transition. All of these major transitions are, but it’s not necessarily, quote unquote, the end of the world. And it might actually usher in a new way of doing things that’s better.

[00:31:18] Emanuel: Could you imagine doing anything without AI today? You, yourself, I’m asking Kevin, in your workflow. For example, I’m an a digital marketer, so I do marketing. I help people with their websites, with their content strategy, paid advertising producing videos social media, all this stuff. I couldn’t think of going back off to not using AI in my workflow.

[00:31:53] Kevin: I’ve got two different answers to that. And that’s, if for whatever reason the tools just got shut down tomorrow for whatever reason and nobody had access to them, I would adapt just like everybody else, right? But as long as the tools are there, I’m going to use them because… they don’t really save me time in what I’m doing.

They do, because I do like really detailed research on stuff that would take me longer to do without the tools, but rather than the advantage being like a time saving, it allows me to do broader and deeper research on topics.

[00:32:34] Emanuel: For what you do. But…

[00:32:36] Kevin: Yeah.

[00:32:36] Emanuel: I promise you that. In five, not even five years, three years, the conversation will be totally, different.

And I’m going to say something controversial right now. This is my rant my part of the podcast where I rant about, we’ve been hearing news for the past couple of weeks.

Again, this is December 2025. Couple of weeks, couple of months about companies, big companies laying off people, many headlines mentioning due to AI, which is not completely true at the moment.

Yes, I believe UPS and some delivery companies have a system that’s been… that replaces human with AI or robotics, Amazon as well. But most, of the layoffs are strictly economical, right? Moving money, because they need to be competitive. They need the chips, so they need to invest heavily.

And how, where do you get cash? That kind of cash? You look at your own company and say, okay, we’re going to let the people go. We’re going to use the money that we pay them to invest. So far it is just for the investment at this stage. And now the controversial part is… this… and I don’t want to trash anyone, but the reality is this, all the big companies that I know, if they laid off somebody due to the position that was no longer needed, eventually rehired them for a different position or anything like that.

Realistically. And then I’m going make a parallel with the government. I dunno if there’s a word in English. I can probably Google it or ask but we use “slujbe de căcat”, I think “slujbe de căcat”, it’s also adapted in English, right? It means that a job, your job is just sitting on your own as essentially, which is a parallel to government employees or many in some of the big companies as well.

Job that does nothing. And the reality, is that many of the people, unfortunately are like that. I’m talking about, think about the last customer service interaction you had with a human. Was that a pleasant experience? Was it a good experience or was it painful… where you just keep asking the same questions.

Keep saying that, oh, it’s our policy, it’s under the the rules or anything like that, which is totally fine, but where I’m going with this is… that’s the first thing that I would replace someone with. That’s the first position that I would replace someone with something powered by AI. At least AI can take the compassion.

I could think of so many examples that I’ve seen in IRL in real life. When, whenever I went through the cell phone to renew it and there was this young lady that clearly didn’t have patient patience for a senior couple, and clearly she was not supposed to be in customer service. You either have it or you don’t.

Even when she was smiling, it was a painful process to smile, right? All those interactions I would gladly have with somebody else because the human part of thing doesn’t do anything. Human doesn’t help me as a user, as a customer with anything at the moment. So AI has been more helpful and I’m still pulling here with lots of things, especially when you try to do some high level stuff.

But the reality is because the topic is AI and the future of work is that AI is coming for your job, which is something good, right? And realistically it all comes down to a personal level. What do you want to do with your life? Where do you want to go? Where do you want to be? It’s not 20 years ago where you could have a career when you can have a simply a job and a career and retire from that one. And just keep doing your stuff, and I believe in the long run that’s, a good thing. Obviously it’s unfortunate to have people lose their jobs, right? Obviously people are at different level of their careers, have different personal things going on. That’s why also the big tech also propose the UBS the, basic income. They’re going to give everyone $2,000, which is nothing in the grand scheme of things.

It’s like peanuts, cheap money.

And you can be… live happily ever after with that, which is not bad of course, but definitely I won’t be on that that side. I don’t want to be on that side… of things when this happens and who knows what will happen with me as well. Maybe I’m in a good position right now.

Thank God I’m very grateful for this. I work hard, I’m learning new stuff, I put the time in. So I like to think that those are some of the skills that will help me survive. Whatever is coming.

[00:38:32] Kevin: It’s entirely possible… and, I’m going to segue into a direct response to that in a minute, but let me briefly just make one last comment on the concept of how economies respond to existential threats.

I won’t go into too much detail, but responses to existential threats, which in the past has typically meant periods of warfare have been pretty extreme and have been an all hands on deck kind of situation.

Now, on the other extreme, what you spoke about… that idea for which there’s a word in Romanian, there’s a… an American anthropologist named David Grabber who actually did some study on that topic and the phrase he came up with is “bullshit jobs”. And bullshit jobs are not reserved for people who work for the government. Like lots of for-profit corporations have people working for them who feel that the job that they do is meaningless and unnecessary, and if they stop doing it, no one would notice.

[00:39:43] Emanuel: Yes.

[00:39:43] Kevin: He actually wrote a book on this topic.

The title of the book is Bullshit Jobs. I haven’t read it yet, but I’ve read another book he wrote that I thought was like, really great. So it’s like on my to-do list is to read this book.

Now, if you look at the history of the way employment has changed due to technological innovations and AI is a pretty significant technological innovation, it’s incredibly disruptive.

Lots of people get put out of work. Then lots of jobs that we could previously never have imagined come into being. If you’d have told someone 40 years ago that their children would make a living as a social media influencer, they would’ve said something like, what is that?

And the same thing is likely to happen with AI.

We have this notion that jobs exist because they’re needed when in reality jobs exist because there’s money to pay for them. They may or may not be needed. Ideally you want as many jobs as are needed, but if you work for a company who is not bringing in enough money, no matter how badly you need someone to do that job, there’s no money to pay them with and that guy’s not going to get hired.

If you work for a company like Google or Amazon that has money coming in hand over fist…

[00:41:12] Emanuel: Yeah.

[00:41:13] Kevin: They can hire people they don’t need, to their heart’s contempt. And that’s just the way it works. For better or worse.

[00:41:19] Emanuel: Yeah.

[00:41:20] Kevin: So things get shuffled…

[00:41:22] Emanuel: Because you don’t want, as a society, you don’t really want a lot of people unemployed and you definitely don’t want them being laid off.

For sure.

[00:41:32] Kevin: And this one might actually be different. So, far every technological… significant technological change, people have been absorbed into the workforce doing completely different things. This one may in fact be different, and if it is, we don’t necessarily know how we’re going to deal with it.

On the other hand, it might just be another majorly disruptive technological paradigm shift, and when it’s all over, everything will settle into something resembling what we call full employment… 4% unemployment or 5% unemployment… and we’ll say, oh yeah, this works.

It’s too early to tell.

[00:42:15] Emanuel: I have another theory because we read these news, people are getting laid off, which is obviously real, but in 2025, people being laid off, that doesn’t mean one… that they don’t have a job… sounds weird, but bear with me for a second. And it doesn’t mean that, let’s say a hundred people were laid off from a big company.

That doesn’t mean that all a hundred people don’t have a job anymore, and definitely doesn’t mean that those a hundred thousand people don’t have an income or don’t have any means of surviving anymore. It’s not the same thing as it was, let’s say 20, 30, 40, 60, 70 years ago.

We know for sure that all the people that I know with very good jobs, with very good jobs and earn a lot of money also have another thing going on.

They either invest in the market they like to do day trading. They like to volunteer and do something else that may or may not get paid. I know some people that get paid and they choose to donate the money back to the organization and all those things. So in 2025, I think we should look at things differently and see exactly what the economic impact these actually have.

And the president from my neighbors down south in this case Kevin’s president make it made it made a very, I would say, not bad decision, but employment, they don’t show… the employment is doing good because you don’t show statistics anymore. You had a gap. Now they show some statistics, I believe, in the late November, and I think recently what I read, they decided not to release… not to do those anymore.

Is that true? I might be.

[00:44:17] Kevin: Oh, they’ll get back to it. Yeah. So during the government shutdown, the people who compiled the statistics were not working. So they just made up numbers and… but…

[00:44:26] Emanuel: Talking about halucinations.

[00:44:27] Kevin: The people who compile statistics are back to work… and this administration is not going to be in power forever as much as they seem to like to think they will be, it’s highly unlikely.

But that’s, another topic, right?

[00:44:43] Emanuel: We’re close to an hour also, so we should probably wrap up.

We discussed energy, and it’s consumption. I didn’t say here, but I truly believe in nuclear. It’s clean, it’s cheap, it’s safe. Yes, very controversial thing to say, but it’s safe.

It’s just… don’t mess it up. Know your thing. Don’t let ChatGPT run… your facility. But maybe a rebranding, call it something different. Elemental Energy, as I’m a big proposal of, I believe in that. We discuss hallucination. And to your point right now, you don’t need AI to hallucinate.

Many people do that. What else is something that is important and we didn’t cover?

[00:45:33] Kevin: I’d like to actually talk about two things, but they’re on disparate aspects of this.

One of them is like the dystopianness, the existential threat. Like I firmly believe that if and when we get there, a part of the response build out is going to be a rapid implementation of nuclear energy.

And I mentioned that the tail end of the last episode about Project Pele.

[00:46:01] Emanuel: Yes.

[00:46:02] Kevin: Project Pale is a modular nuclear reactor that the US Army is funding. It’s an R&D stage right now… because they anticipate needing more energy on the battlefield. And it’s a 20 foot shipping container, or it fits in a 20 foot shipping container, and it generates somewhere between one and five megawatts steady state.

Now, roughly a megawatt is enough to power a thousand houses, so however many houses there are in the US and I looked this up. If they make five megawatt units, 30,000 of them will power all of the homes in the United States. And the other advantage of using modular, small modular reactors is the other expense involved in electricity is not just producing it but transmitting it over distance…

[00:46:59] Emanuel: Distribution.

[00:46:59] Kevin: What we call the grid, right?

[00:47:01] Emanuel: Yeah.

[00:47:02] Kevin: If you have these like nuclear batteries that you can drop wherever you want, we can spend a lot less building out the grid because we’ll be building municipal grids rather than an national grid. And we wouldn’t do this because it necessarily makes technological sense. We would do this because it’s cheaper.

[00:47:23] Emanuel: I bet China is already at least on par.

[00:47:26] Kevin: Oh, I bet they are. Yeah, I bet they are.

[00:47:29] Emanuel: With the states, if not a little bit further.

[00:47:33] Kevin: So.

I firmly believe that if and when we hit that wall, we’re just going to go there and we’re not going to think twice about it. Like that debate will end because the other option is turn out the lights and we’re like, no, we’re not turning out the lights, we’re going nuclear.

Yeah. And I don’t mean literally turn off every light in America, but Americans are not going to take kindly to four hours of no electricity every day. They’re like, what was, what are you talking about? This is America. We don’t do that. Right?

[00:48:03] Emanuel: That kills no. Even Romania, they couldn’t stand.

That’s what drove people to turn to take Ceaușescu down, essentially, right? We had two, three hours of electricity, 30 minutes of TV, which was him telling us that there’ll be no electricity and a couple hours of hot water a week or something like that. So if Romanians couldn’t take it only for 10 years, imagine American… any other place in the world.

Everything is a season. Everything has a cycle. I think it’s important to understand this and call it, let’s call it Elemental, not nuclear. Because nuclear may have some negative connotation, but again, reminder for people, we won’t go into details because I, first of all, I don’t know them, but remember this, it’s cheap, it’s clean, it’s safe.

Okay? So it’s not polluting. It’s not polluting like any other source of energy. And you said you wanted to share two things. I do want…

[00:49:07] Kevin: Yeah, I forget what the other thing was, man, it just flew out of my mind while I was talking about nuclear.

[00:49:13] Emanuel: Maybe I’ll say something and you’ll remember in the meantime, when we talk about AI, what we’re referring to are actually the LLMs, the large language models that we’re interacting with every day, which is essentially a… it’s not… nothing intelligent in it.

There’s just some machine learning, some highly sophisticated algorithm that can predict better and faster what the next word would be when the AI that will completely change the game. What happened is probably the moment one of these somebody will figure out the quantum computing stuff.

So we know that Google has made advancements, I believe Microsoft, IBM, all these big companies. China doesn’t talk about it, but I’m pretty sure they have a couple of them going on as well. I’m assuming other places on earth are trying to crack the code essentially because they’re not economically viable.

You can’t do much with them at the moment, but they, will happen at one point and that’s when things will probably change for sure. Now, can you imagine what kind of energy you’ll need to operate those things?

That’s important. What I wanted to say is how I envision the future will be looking will look like probably there’ll be a couple of influential poles around the globe like it is right now.

You had the US for the past 50, 60, 70 years the US has the major in influence, the major power dominating. You had a beef with USSR, the Soviets at one point, and then US was the dominant force allowing China to raise its power right now. It became on par with many things. It’s not there yet, realistically it has the potential.

Many leaders in any many fields do think that China will become the dominant force.

Obviously for the sake of the planet it’s good to have not just one, but more influence, more powers, more poles of influence, I believe.

I like to think that not all of them will be that evil and do have some good things in mind, right?

Health, prosperity, growth and all those things. But the future will probably look like a couple of… groups of influences around the globe, couple of superpowers that will include also an AI component, significant AI component.

[00:52:18] Kevin: Oh, absolutely.

[00:52:19] Emanuel: Symbios.. So it’ll be the North American AI with its own government and the Chinese Asian AI with its own government.

And I’m assuming India will have a world… would have something to say as well by their own and not just with the rest of the Asian countries. And you’ll have this, symbiosis between the humans and the AIs.

And I believe it’s really important, and I also came to the conclusion that it should, it’s important to create an at least work on an LLM that’s not built with English as a first language.

The language we all know that the languages influence the way we think, the way we walk, the way we do stuff, the way we exist. It influences at more languages at the bigger level. English has allowed us to systematize and create… have a common language that we can all interact with everyone around the world.

But for example, in my own language, which is a romance language based on Latin, it’s the closest actually to Latin because French, for example, has been influenced more by the Germans or the English than Romania was. I can say that I’m in the past, in the present and in the future in the same sentence with no problem, and that sparks different parts of the brain to construct your reality.

If you can translate that into technology, into AI, I believe that’s important. I’m not saying to build just that, but it needs to be an alternative. And it needs to push… Mistral… some of the French guys are doing a pretty good job. I was surprised to have some positive interactions with the… their LLM and I was getting some answers faster than with OpenAI, for example, which is constructed based on a… on the English language at its base and even the coding and even the programming, right? When you program the them, you use English. So that’s what I mean. I hope that made sense.

[00:54:44] Kevin: It does. Now, I’ve never looked into this particular detail, but I always thought it was plausible and credible. But apparently Google Translate, where it translates various languages into each other, it doesn’t translate directly from English to Mandarin.

Apparently they use a… this is going to be a ridiculous statement I’m going to make but I don’t know how to make it not ridiculous. They use a even more standardized version of Esperanto as the intermediate language. So everything gets translated into and out of a Esperanto variant, which is… been stripped down even more for purposes of computational speed, and I remember thinking, yeah, that makes sense.

[00:55:35] Emanuel: But for those who don’t know, Esperanto is a made up language that nobody uses apparently. No human uses, but apparently Google does, which is scary. It doesn’t have any practical use and God knows what the impact this might have.

[00:55:52] Kevin: I would argue that if this story I just told you is true, that’s a practical use.

[00:56:00] Emanuel: Yes, correct… correct, but up until recently.

[00:56:04] Kevin: Yeah yeah.

[00:56:05] Emanuel: I haven’t heard about any practical use in itself, and I think it’s stripping down a lot of the meaning of words, but there’s… some people don’t know a certain feeling. They don’t have a word for it either. Once they learn the word, they experimented the feeling as well.

It is a topic for a different conversation in itself that I propose will be recording soon, but last couple of words and let’s do a closure about AI and the…

[00:56:44] Kevin: I do remember the other topic I wanted to talk about…

[00:56:48] Emanuel: Oh.

[00:56:48] Kevin: And that’s the mistakes that AI makes.

So companies are reducing headcount and replacing their function with AI tools. So the existing people use the tools in theory to get more work done, but it’s turning out that these mistakes are very expensive.

And I saw a video just within the last few days, and I don’t recall what company it was, but a company is using AI to create summaries of communication. So summaries of meetings, summaries of documents, summaries of emails.

[00:57:27] Emanuel: I do that.

[00:57:29] Kevin: And then at some point they realized that approximately 10% of the information in the summaries are wrong.

Yeah, but they don’t know which 10%. So now they have people who are double checking the accuracy of the AI because accuracy matters to them. And now they’re wondering if laying off all those people in the first place, what’s the smart thing to do.

So this is a different manifestation of the AI agent not knowing that the guy wanted Mountain Dew, but thinking 18,000 glasses of water was a reasonable request.

[00:58:07] Emanuel: That’s not the AI’s fault of course, but this is strictly a decision made by somebody who did not have all the information necessary to make that decision.

[00:58:19] Kevin: Right.

[00:58:19] Emanuel: Or shouldn’t be in that position in the first place. And hopefully he’ll get replaced sooner rather than later by a symbiotic entity between a human with AI enhancements probably.

[00:58:35] Kevin: But it seems like what’s happening is senior executives are buying into the hype of AI and they’re looking for opportunities to reduce headcount without fully understanding…

[00:58:45] Emanuel: Yeah.

[00:58:46] Kevin: The ramifications… and we’re experiencing these unintended consequences.

[00:58:51] Emanuel: One Marxist could argue that’s capitalism, but…

[00:58:56] Kevin: Actually a lot of capitalists would argue that’s capitalism.

[00:59:00] Emanuel: Yeah. I think that’s topic for another conversation.

[00:59:04] Kevin: Yeah.

[00:59:06] Emanuel: AI in the future of work. AI is here, is part of it. Probably will have a bigger impact in our lives in the years to come. I’m looking forward to it. I can’t imagine editing this podcast and putting it out there without the help of AI.

That being said, I’m Emmanuel.

[00:59:27] Kevin: I’m Kevin.

[00:59:28] Emanuel: We’re the Curious Pundits.

We… this is our podcast. Go to curiouspundits.com and follow us, like us, and drop us, drop a comment, tell us what you like, what you didn’t like. If we said something that you feel it’s not accurate, or we made a mistake, feel free to reach out and until the next episode.

Bye.

[00:59:50] Kevin: Take care.