The Horizon, Bot vs. Agent, and The Cost of AI

Welcome to wake up with AI, the podcast where human powered meets AI assisted. Join your hosts, Chris Carillon, Niko Lofakas, and George b Thomas as we dive deep into the world of artificial intelligence. From the latest AI news to cutting edge tools and skill sets, we are here to help business owners, marketers, and everyday individuals unlock their full potential with the power of AI. Let's get started.

Good morning. October 30, 2024. It's time to wake up with AI here with George and Nico. How are you fellas doing today?

Doing good. Yeah. Doing good. Super excited. I always wake up now curious to what this 30 minutes is gonna bring and how it's gonna impact my life.

That's a good place to be.

Catching up to some news this morning from some of the big tech sector people and some really, some really interesting news. It's a lot of future now. We've kinda moved a little bit away from the present and have kinda embarked on the future. So, folks, it's getting to a point where if you're gonna start asking me about AI stuff, you're gonna start hearing answers that are 3 to 6 months from now, not stuff today or, like, I'm sure most people have noticed. And I I don't know whether or not I apologize.

I I think you guys would probably be a better judge of character for this than than anyone else, but I'm sorry not sorry that I'm not your AI tool guy. Right? Like, I'm a little too in the weeds trying to understand what's happening under the the hood of all of that than, like, the person that's and you guys know how it is. Like, I get too much leverage out of what's already there. Right?

There's out of 3 tools, I get all the leverage I need. So it's really throws me for a loop when somebody like, George has a different story. Like, he got a a sick AI app that he uses for presentations. He got sick AI apps that he uses for organizations. Chris has sick AI apps that he uses for, organizing stuff and for putting stuff content together.

Different story. I'm not a content creator, so, like, that's that's why you ask those guys.

And there's plenty of people that you can watch for that. What I like, Niko, is I do like the under the hood stuff. I do like the let me ask questions that most maybe aren't asking yet. And, like, I like the future crystal ball focus because I think while a lot of what we talk about here is, like, adoption and skills, there has to be a level of, but, like, what's coming? I mean, listen.

When we drive down the road, we don't look at our front bumper. We look at the horizon. Right? And and so when I think about this show and I think about you and the part that you play in the Slack channel that keeps coming and coming, it's like, oh, there's the horizon. Like, let's look where we're going.

And so I love that part.

Yeah. It's such a good analogy. I do what I can, brother. I do what I'm saying. We look down and check the speedometer.

Yep. Like, how fast are we going right now? But briefly Like, right now. Right?

But if you keep looking at speedometer, you're gonna wreck.

Right.

If you're worried about your speed versus your direction, yo, you jacked up. I'm just gonna say it.

Well, I think that's that's the point of and hopefully, that's why we can deliver some good value and content is, like, most of the time, you should be looking out front, but, occasionally, we're looking in the rear view. We're looking down at the speedometer. Whatever is telling us what's going on right now to confirm what's going on right now. And that's where I think, like, this balance needs to be played, especially after what I saw last night, Salesforce's thing. It's just for most of the people I interact with and this is where there's, like, a huge disconnect between the SaaS and tech and then, like, everybody else.

Right? And if I could just this example that they show last night where it's like, alright. Agents are here. Like, everybody, like, so easy to make them, and they show this website. Like, this is what you do.

It's not a chat. It's not a chatbot. It's an agent. Nobody's gonna know that. It looks exactly like a chatbot that we're used to.

And how many people have you talked to who were like, don't touch me with a chatbot. Like, I don't want anything to do with chat. I haven't had any good experiences. So it's like another situation where we're trying to deliver something awesome and new, and there's baggage on back because the thing that you want it to look exactly like, a lot of people don't like at all. Like, that container that it's in, which there really isn't any other way to do the use case because it was really cool.

It's like start talking to the agent, and by the end, you've got, like, applications for university filled out and a scheduled, you know, intro session. Like, all the things that you would have normally just gotten links to and had to go do yourself. Like, we're done. But I still gotta click that chat icon and, like, go down make

a request.

Right? And that's the part where, like, hey. Here's great tools. Right? So I I think that's where it's gonna be this balance between what could be and what is and, like, understanding that what is is a lot of freaking baggage about how this stuff is supposed to help us and and has really gotten in the way for for so many people.

Yeah. And if you guys aren't privy to what Chris is talking about, Mark Benioff, master of attempting to be disastrous to the to the industry, announced and did the, Salesforce Salesforce announcement yesterday for AgentPharce. And so Agentforce is, like, is the same thing as Breeze for HubSpot. It attempts to be the same thing that is. And it's interesting to me.

I I you know, hats off to Dharmesh. So anybody that's paying attention to the game, HubSpot beat Salesforce to market. First time in my career working with HubSpot. I don't know if that's overall. If it was overall, it absolutely should have been shouted from the rooftops, but, you know hopefully other marketing mags will do that justice.

To Chris's point, if you're following any of the mainstream people that are talking about news, talking about what's coming, talking about what's out, and what's changing, you will definitely know that the tide has changed, and the stories have gone from talking about what's possible and and I mean it's still about what's possible, but it's gone from talking about application and how we're going to embed this in our everyday lives to being more about okay how much disruption is this gonna cause? What is this going to do to the job market? What is this going to do to future way of life, future way of living? It's finally, you know, it's at the point I shouldn't even say finally. It should have been here, you know, a year ago.

We should have been having these discussions, but, you know, we never we never do. Right? We never really worry about the the earth getting destroyed until we can actually see the comet breaking through the atmosphere. Right? And then we start worrying about it when it's 70.

Movies go. Yeah. Yeah. We talked Claude about this kind of stuff. It tends to find that pattern real.

I think one of the most interesting things, if you if you talk to Claude and and have open conversations with it is you always have to keep in mind it knows all the things. It knows all the things that you're thinking of going to Google. So one of the really awesome things that I've been doing lately is telling it, hey, just look across human history and see if you can find a pattern of this, that, or the other thing. Because it can. It can do that in less than a minute.

Can do it in less than 30 seconds. And that's why I say it's definitely the smartest thing around because I don't know a historian who could just straight out the gate be 100% positive, 100% accurate. Yeah. Definitely. There's there's a pattern of that that happens here.

I'm pretty sure they'd have to crack a couple books at least.

See, but you have to come to that understanding, and you have to have that open mindset, though. Like and that's not the skill that pays the bills, but, like, an open mind and a level of curiosity and being able to even leverage the right questioning. One of my biggest worries is, like, people will have a hard time getting there. Right? Because you have to remember, we're coming off of the system that built manufacturing workers.

We're coming off of the, you know, retain enough information long enough to take a test and then forget it. These are not the skills that you're looking for. Right, in in the best Jedi arm wave that I can give you this morning. And so, like, you say it because it's native to you. You're doing it because it's who you are.

But, like, in the future, these are the things that people will need to make their selves into or not, but I don't like to think about the or not. I like to think about what's the best case scenario for the humans to be able to leverage the most to provide more.

I'm sure people are thinking, okay. Well, that's not news. It's not doesn't translate. I didn't learn anything new. What you are going to learn about and why all this is important and why I bring any of this up is to understand that, you know, we we've been talking about it.

It's not gonna stop. At what rate is it going to continue? There's been, obviously all the time, I think yesterday or day before, Elon Musk was doing a webcast talk, if you will, at a conference. I'm watching so much of this stuff. It's pretty difficult for me to remember all the names of so I will post the links in our repository and we'll try to get that up so you guys can take a look at it.

So asking Elon you know where where do you think the the progress of this stuff is going and Elon's response was, well essentially at this point and again think about the fact think about where we are with 01 preview compared to GPT-three. Okay. Elon's answer was, well we're basically doing 10x per year. So at 10x per year 4 years from now we are 10000x ahead of where we are now in terms of like intelligence level on the model right which is why I say like GPT7 is already probably solving things that we cannot figure out not like crazy high level stuff, but probably solving things like whatever it might be. Whatever the lowest level of unsolvable stuff is.

Right? So that translates to SoftBank that was at the international conference for AI and tech investment. Several months ago, Sam Altman said, and everybody laughed at the fact that it would take about $7,000,000,000,000 as a group investment in order to get AI to a point where it would be super intelligent. The I wanna say it was the CEO for SoftBank made the statement and said, given where we are and given the rate of progress, the soft cap for getting there would be nearly $9,000,000,000,000. And he said that that was being conservative.

That's just investment. That's not 9,000,000,000,000 in, like, anything other than building infrastructure and investing in. Semiconductors and and chips and graphics cards and compute and electricity. Because they've done the math and they're like look the the amount of power I'm gonna finish this and then I'll cap it off. The amount of power that we're gonna need for this stuff just for 1 or, like, for 1 digital super intelligent model, will take more power than the entire United States grid.

Currently, we're using enough electricity for 1 model. For the highest capable model, we're using enough electricity to power 1 I wanna say I'll go back and double check this number. I don't wanna be wrong. I wanna say it's something like close to 10,000 homes for the entire year is, like, 1 hour of processing power for a one preview.

See, and I don't think people think about that side of it.

One one little little sprinkles on the cake, we're not at quantum computing yet.

Yeah. Which is funny because this conversation I go back to one of the interviews you shared where is it Eric Schmidt? Yeah. Eric Schmidt, and he was talking about how he loves Australia because Australia would be like a dope place because of, like, energy and windmills and, like, just the way that it would be a great. And so to kinda tie things together of, like, the amount of power, we just don't think about that.

Right? At the point that we're at with most humans, we're like, oh, cool tool. Let me go ahead and do some chat gpt in today. Meanwhile, we're like, hey. Turn the light off.

Close the refrigerator. Burn electricity. But we're like, let's t p t it. And we're not thinking about that.

And what have I told you a couple of times? It's a bloody idea.

In trouble. Are you talking about when I got in trouble? Because I wanted to go GBT something, and you're like, just do it in Excel. Is that what you're talking about? Yeah.

Let the listeners know I got in trouble.

Well, look. It's it's illustrative. Right? So and and it's educational. I wish you guys could could be in on on the, the behind the scenes tutorial calls that we do and teaching each other stuff.

At one point, I was working with George, and we were doing something together. And he was just like, oh, yeah. I'll I'll just take this, and I'll give it to to GPT, and I'll have it do. And I was like, woah. Woah.

Woah. Woah. Woah. Woah. What you were trying to do is just put data in one column and then ask them something.

Like, no. No. That's more what it was like. Let's be honest.

So you were halfway through hitting the prompt, and I was like, oh my god. Stop him. Yeah. Come on, guys. Come on.

Just because it can doesn't mean you should. With great power comes great responsibility. So understanding that this thing is super powerful. Right? But if it's a trivial task that's gonna take you 30 seconds, don't save the 20.

The cost difference between the two is insane. It costs almost not like pennies on the dollar for you to use Excel or Sheets and do that formulaic thing. I'm talking to the listeners. Or it could cost, I don't know, an entire home worth of electricity for a year for you to ask GPT to do it. Right?

I'm not talking like people have to get away from token cost. Token cost is one thing, but it's trivial by comparison to the natural resource cost. And yes, one day, even already, Slade's already admitted it, you know, the models are are going to start getting and they already are. They're getting more efficient. We're getting more out of less.

Sadly, it's still gonna take about 4 to 5 years until we get some real, like, juice from the squeeze on efficiency.

So in, like, 4 to 5 years, I know I don't have to worry about Nico yelling at me?

Yeah. Yeah.

Alright. Yeah. I'll I'll let

it go. I'll let it go. Brother. I don't know that that message is ever landing the way that you want it. Because I heard the same stuff about Google search, like, at at these data center conferences.

Like, one Google search equals, like, one light bulb. Right?

See, but, like, okay.

That's that's

the thing is, like, that one, that's not so bad. Right? But

Well, the difference is I don't know that it's only gonna take me 30 seconds in Excel, or I don't know how to do it in that 30 seconds. That's where the balance between, at this point, we just start trusting that this is the place I go to do things.

And, actually, when I think about this, your guys' thoughts, is it irresponsible of Google? Every time you do a search now, it gives you the generative response, which means that it's using Gemini and doing a light bulb. A generative query. Right?

More more than a light bulb, but that one, they all said, it became more expensive. That's interesting.

So now, like look. The industrial revolution is what put us in the place we are currently with climate. Are we once again about to trivialize natural resource in the name of doing things faster?

Yeah. That's another show probably, but, yeah, I'd say, like, there's an element

of the news is bigger than you think.

That's a deeper culture. That that's a problem. Right? And we should probably solve that problem at some point. But go ahead, Chris.

Well, I think that's where it's I'm sure some minds out there, like, especially the ones that are pushing all this stuff forward, like, as quickly as possible. And if you're coming at this from a place of abundance, like, the faster we go, will we be faster to innovate our way out of this problem? Especially if if AI is 10 x, it's 10,000 x better. Like, has it solved climate change problem for us at that point?

Here here's an interesting bit to that, only because I keep hearing the same thing being put forward. Just a reminder that Claude also kinda gave you while we were talking about it. We have to change the climate, not Claude, not gbt. It will give us the answer just like any guru can give you the answer on how to fix your life, on how to be better, on how to do what you need to do. But you need to do what you need to do.

So while I look forward to the those problems getting solved, I think we are absolutely gonna solve medical issues. We're absolutely gonna solve, like, things that are tangible that we actually that a small group of humans can affect and grow into something greater. Will we solve climate change? I don't know.

I love that you're talking. I love that you're talking about solving, though. I really do.

Yeah.

The fact that you're talking about solving and we're talking about AI and we're talking about problems because hey, folks. Today's AI skill that pays the bill is AI facilitated problem solving so that you can attack challenges with speed and precision. Listen. In business, challenges pop up every day from project delays to customer complaints, and the key to tackling them effectively is having the right insights at the right time. By the way, we're talking about business in this segment, but as you can tell, it is much larger, climate control and all those things.

But this is where AI shines, turning complex data into clear, actionable solutions so that you can problem solve faster and with more precision. Imagine you're faced with a sudden dip in customer satisfaction. Instead of combing through endless reports to find the cause, AI steps in to analyze customer feedback, pinpoint patterns, and flag the specific pain points that you're feeling. Within minutes, you've got the insights you need to address the root issue, not just the symptoms. And here's the power of AI facilitated problem solving.

It doesn't just give you data, it connects the dots using predictive insights and real time analysis to show you the best path forward. But like Niko just said, us humans are gonna have to take action. Maybe you're trying to cut costs in your supply chain but aren't sure where to start. AI can break down the processes, highlighting inefficiencies, and even suggesting areas where automation could have resources. In a world where every second counts, AI lets you see the big picture and the finer details, helping you find solutions that are practical and effective.

The best part, AI adapts. Let's hit the highlights. Data driven analysis. AI cuts through complex data identifying patterns and root causes so you can tackle the real issues faster. Predictive insights for proactive solutions.

AI forecast potential problems before they escalate, helping you stay ahead of challenges rather than reacting to them. And clear actual information, AI presents information in a way that's easy to understand, making your decision quicker and more confident. That's the AI skill that pays the bills. I need to go GPT how to actually be able to talk today. But other than that, we're good to go.

It's all good, man. It's all good.

I'm like going I'm like, mouth, just work. Can you just work today, mouth? Like, what

is going on? Like, I could see it. And I do the same thing. And after, like, the second or third word, it's you're sitting there, like, I swear I can talk today. I swear I know how to do this.

My usual go to is, is it Friday yet?

Yeah. Yeah.

Speaking of things it can't help us with yet. No. I think this is such a good topic and why, like, critical thinking and, like, the way that you learn and grow right now is gonna be the same way that it's gonna greatly impact how AI, like, supports you or empowers you to to grow and and learn moving forward. In an example, I was showing, like so I've got a couple of Spanish speakers coming into the house next week, and they're gonna be around or this week, and they're gonna be around for a while. And, you know, previously, we're both in the same situation.

We can understand the other language, but it's really hard to converse, like, with each other in that language. So, you know, you have WhatsApp and you have Google Translate, and you you talk into it. It translates, and then you show it to each other and you read it. While I was testing out the the advanced voice mode, I still need to work with you, Nico, on the exact, like, prompt I need to give it so it doesn't get confused on who's talking and when it should translate and things like that. But I was putting it out on the table at lunch, and the one of the immediate responses is like, oh, now you don't have to learn Spanish.

And I was like, well, kinda. But for me, it's likely going to help me learn Spanish faster.

There you go. Well

yeah. Because you're hearing it and hearing

it. Yeah.

It's like

And I'm use yeah. And that's the how many things are like that where there's definitely gonna be people. It's like, okay. AI's got that. I don't need to engage my mind right now.

For better or worse, I can never not engage my mind. So it's like whatever is happening in front of me, I'm trying to take something away from it and to build on it, which also means I'm constantly focused in the present moment. So this like, doing that, being able to do that while also having everything documented while it happens is just, like, amazingly powerful right now. Because documentation and review and reflection, like, that's one of my biggest areas of weakness. And since that's happening in the moment, it's just like, man.

I love where you're going with this, Chris, because the fact that it's being documented, but I would even say it's documented in a way that you can then iterate on it. So, like, one of the things that I'm loving right now, again, bringing the superhuman framework to the world and what that looks like from, like, content creation and assets, and being able to create that is then documented to that then I can listen or read and then iterate and recreate, and the recreate is documented. And so, all of a sudden, you start to get these layers on top of the thing. But what's crazy is while you're bringing something to the world because you're trying to teach what's inside of you, you're also learning from what's coming out of the creation process. And so, not only are you layering and iterating on the thing you're creating, you're iterating and layering on you, the human, the new information that you're providing and bringing, it's you have to have a learner's mindset.

You have to be curious. You have to be willing to test everything. You have to be willing to iterate. You can't fall in love with your baby. You have to be able to pivot.

Like, ah, like, this these are the types of humans that we have to be creating ourselves into to move forward into the future where we're going.

Yeah. And, I mean, I I I love that you're trying to push things in a in a direction in terms of like like you mentioned with prompting and like and getting better at it. I was thinking about the fact that like, yeah, we haven't really thrown out any prompting tips at this point, but

Oh, we should do that.

You know, I I think about it and it's definitely I I think we definitely should. It's that the type of tips that I'm going to give out are not gonna be what you would expect. So in other words, like, I'm not really gonna give you a prompt so much as I will I will give you, like, a way of working it. Right? So for, you know, Chris's purpose in particular just kinda reminds me that when you're giving directions, there's 2 ways of of doing it.

You can, you know, give direction. Right? And you can tell the person what they're supposed to do or you can give direction and tell the person what they're supposed to do and what they should avoid. If you've ever if you've been working with these models and you do that and you tell them it's a complex, you know, the more complex the task is, the the more the supplies. But if it's a complicated task where there's multiple steps involved, absolutely give it a shot where you're telling it, hey, I need you to do this and I need you to get this done and do this thing and make sure that this is like this.

But I want you to make sure you're not doing this, I want you to avoid doing this, and I want you to keep keep away or, you know, make sure that you're not doing this.

One of my favorite things to do is please stay away from bullet points unless they are extremely needed to add educational value. Stay in paragraph style, in conversational tone, educational, and have a little fun along the way. Man, the stuff that comes out anyway, like, dude, I would love to have, like, prompting, pieces in here because and not like, hey. Here's a fish, but, like, hey. Here's a rod, a reel, and a tackle box for so you can go catch some dope fish with your AI tool over the next day, week, month, whatever.

Like, that I think that would be so valuable to the people listening. And if you agree, by the way, listeners, hit us up. Let us know. Like, we're here to to build a show that you feel comfortable with, valued, all of that good stuff.

Don't forget that the easiest way is what? Wake up with AI.

Wake up with AI.

That's a wrap for this episode of wake up with AI. We hope that you feel a little more inspired, a little more informed, and a whole lot more excited about how AI can augment your life and business. Always remember that this journey is just the beginning, and that we are right here with you every step of the way. If you love today's episode, don't forget to subscribe, share, and and leave a review. You can also connect with us on social media to stay updated with all things AI.

Until next time. Stay curious, stay empowered, and wake up with AI.