Calculus of IT - Episode 35 - "The Value of the Virtual CIO"

The AI assistant is not a new concept.  It’s been made “new” again through the GAI zeitgeist effect, but we have used AI assistants for quite a while now.  In Episode 31, Mike and I discussed using GAI for IT strategic decision-making. Tonight, we took it a step further in connecting the dots with episodes 33 and 34, which focused on redefining the IT Paradigm and the future of the CIO.  The net result is that in Episode 35, we at last examine the worth of the virtual CIO. This assistant can help any moderately seasoned IT professional create an entire CIO program (in about 45 minutes) and, by this, have a forever assistant to help construct the IT paradigm.  We have some additional questions on the substack post that are worth discussing further, and maybe we will tackle these in a future episode.  I have also released the entire vCIO Bot script code package (which can be re-used in Make, Zapier, Botpress, or your favorite bot-making automation machine) and posted it on Substack. For now, enjoy Episode 35 - The Value of the Virtual CIO.

Transcript

Nate McBride 00:00 
What are you talking about? It's fancy. Anyway, we do have a lot to cover tonight. I think that we're hitting this one moment in time where we're going to have another divergence of the theme of IT leadership, but I want to try and bring it back to exactly that point, which is we need to talk about AI again, but in the context of how it contributes to leadership development and strategic development, 

Nate McBride 00:31 
basically, because there are a few things that have happened recently. Before you even get into that, I want to talk about a couple of things, really two in particular. They all set the tone, the context for the episode tonight. 

Nate McBride 00:46 
By the way, welcome to episode 35 of the calculus of IT podcast. We're having it. Yeah, the best one yet. So if you listen to the other 34 and you're like, wow, that's awesome, then you're in for a big fucking surprise because this is going to be amazing. 

Nate McBride 01:01 
It's going to blow your doors off. Last week was fantastic. It was fantastic, but this is going to be even better. I don't know how we keep getting better. We just do. It's just the way it is. It's the opposite of inertia. 

Nate McBride 01:14 
Code in the matrix. It's the same with the wizards of tech where progress is born with every line we're defining the time 

Mike Crispin 01:45 
Sing time! 

Nate McBride 01:46 
I is for winners, sad salads profound, ponies and bubblegum floating around Tech a chord who's bored the calculus of IT with Nate and Mike drowning in the nether Or just is great to never 

Mike Crispin 02:17 
Love 

Nate McBride 02:49 
So I want to start off with Grammarly. I use Grammarly and have used Grammarly for I think like the last six or seven years. Me too. Because when I write, I tend to write run -on sentences that go for paragraphs and paragraphs. 

Nate McBride 03:03 
And the few people that have ever edited my copy for me, including my sister -in -law's husband, God bless him, Craig, I sent him very long sort of like, hey, I want to put this up as a blog post. Can you review it for me? 

Nate McBride 03:16 
And he gets back and it's redline to hell because he's like, who are you writing this for? Nobody reads like you think. You write sentences that go on with ellipses and positives and Oxford commas all over the place. 

Nate McBride 03:32 
He's like, just stop. Like this sentence, like this, man. He's like, he's like, you can't write like you think. So thank you for sharing with me. And let me just edit this for you. Right. So I've been using Grammarly forever because I don't, I often feel like I can't send things people to review until I've at least used Grammarly to fix them up a little bit. 

Nate McBride 03:54 
And it has saved me. I suck at Oxford commas, the double spacing before new sentences. I'm old school. Now it's a single, single space before new sentences, all that stuff. So anyway, so Grammarly about a year ago or so, maybe 13 months ago, released an update to the Grammarly program. 

Nate McBride 04:16 
And if you're familiar with Grammarly, it's a tool that you can employ in Chrome or your browser to help you sort of fix your writing as you go. It also works inside of apps. Anyway, about a year ago, Grammarly released this ability to verify whether what you wrote was generated by you or an AI. 

Nate McBride 04:39 
So, and it wasn't that good because this was back with sort of chat GPT two and three, where it was pretty much fricking obvious somebody wrote something inside of a GPT, like they didn't obviously wasn't their own. 

Nate McBride 04:54 
But it's gotten a little bit more difficult to detect that anyway. So Grammarly released the authorship verification. And so if you want to know if somebody actually wrote the document that they say they wrote, there's this way to sort of attest to that. 

Nate McBride 05:07 
Grammarly will tell you, yes, this person wrote this document to the best of its knowledge. And it's been in the curriculum for most schools, secondary anyway, for a while. Anyway, this has now been released in public deployment. 

Nate McBride 05:24 
It's now out there in the wild. I got the notice today. And I'm thinking to myself, well, shit. I mean, it basically will, will show a verification or everything that you wrote. And it's like, almost like a DocuSign or, or some sort of e -sign certification addendum. 

Nate McBride 05:43 
Like Nate actually wrote this entire document because he typed it. And in where he pasted things, we can verify the source of those pastings, either paste them from other Nate writings, or you paste them from some other source on the internet. 

Nate McBride 05:58 
Very powerful stuff. 

Mike Crispin 06:01 
Yeah. 

Nate McBride 06:01 
So any student out there who thought they were going to coast through sort of their next two or three years of college with using Grammarly and sorry with using open AI to write all their papers is screwed. 

Nate McBride 06:13 
But I was thinking like what are the bigger implications of this? So I write a lot of blog posts and it takes me weeks to write a post. I write it. I hate it. I send it out for editing with people. It gets edited. 

Nate McBride 06:27 
It comes back a bunch of times. I will rewrite it maybe 10 more times. I'll put it down. I think about it. I'll go for a run. On and on and on, right? I can't ever stop writing. So eventually I just give up and I publish. 

Nate McBride 06:44 
When the time comes that I have to start verifying what I wrote, and when the time comes for people to be trusted as verification, will it matter? And that's the big question, right? I want to start with this off with authorship verification of something that you wrote. 

Nate McBride 07:00 
Are we going to be at a point in time with society where we're so paranoid about someone's originality that we're going to make them attest to every single thing they wrote and where would this then eventually go? 

Nate McBride 07:08 
Like where would we get to in society when every single thing that you write is guilty until proven innocent? Number one. Number two, do you think that it's relevant to have this, especially when there's actually appropriate times to copy and paste from an AI into a document to have that even verified? 

Nate McBride 07:32 
Because we have certain footnote and reference rules in place already established for how to use the bibliography. So I'll start right there and ask you that initial question, which is in terms of this Grammarly authorship verification, what do you think about that? 

Nate McBride 07:52 
I think some tool like that is eventually going to be needed or is going to be desired. I'm not sure. I think it depends on how accurate and how well accepted AI -produced documents are, if largely they're good enough and accepted as normal. 

Nate McBride 08:18 
Nobody cares if you solve a moderately difficult math problem with a calculator. There's certain amount of things that over time it may not matter as much, but at the same time, if there is a lot of scrutiny placed and there's a lot of hallucinations and a lot of people have been using this data and it's been proved false, I think there's more need for sort of this automated source checking in documents. 

Nate McBride 08:44 
But I think it depends on the content and I think there will be a need, maybe in a deeper sense, for documents that are created by AI to be almost fact checked in an automated way. And I don't know if it's fact checking as much as just validation of sources. 

Nate McBride 09:05 
So we see things in certain collab AI tools that already hallucinate against your own data, just putting an alarm against your own data. And that is going to get better. As the LLMs get better, it's going to get much better in small data sets. 

Nate McBride 09:24 
Do you think having a stamp or some sort of verification metadata addendum is going to be important enough? So for instance, let's say I have a document in box and I ask for a summary. Copy that summary and I paste in a new document saying here's a summary of such a such document. 

Nate McBride 09:42 
Obviously, that's a very, very clear indication of this. Let's suppose that in the same scenario, I take a document and I get a summary of it and then I use that summary or portions of that a new document. 

Nate McBride 09:51 
Well, do I have to now attest to the fact that I use the summary of a document I already wrote and a new document to verify I didn't actually write the summary? And the importance of that, like I look at that and say, who cares? 

Nate McBride 10:04 
But I feel like there'll be a point in time potentially where it's not just Grammarly, where we have this attestation effect that occurs where you have to verify that you wrote anything as an original source document. 

Nate McBride 10:19 
Else is the assumption is that you didn't write it. Because we're entering into sort of dangerous territory here. A lot of people are now openly writing books using AI and publishing them. And doing podcasts, which we'll talk about in a minute, and other things that are clearly not done by themselves. 

Nate McBride 10:40 
I feel like no one really seems to care in a lot of those cases because you either are okay with it or you're not. Big deal. It's a matter of when it starts to creep into the corporate thing, then I begin to get a little concerned. 

Mike Crispin 10:56 
out. 

Nate McBride 10:57 
You're going to go a little philosophically or a little deeper here. Is AI revolutionary or evolutionary? Is it a tool or is it the next evolution of our intelligence? And is it accepted as the intelligence? 

Nate McBride 11:17 
Over time, more generations of people go through, you push a button and you get a good answer. AI is from all different political spectrums, all different beliefs, all different backgrounds have been created over the, let's say we're talking 10 years from now or even 20 years from now. 

Nate McBride 11:35 
They're all working together, they're creating richness of data. And what realistically, how can we compete with that? We are so slow to get information out and to take information in. And as these models get better and better, that can communicate information at way faster speeds that humans can communicate at a bandwidth that's way faster than us. 

Nate McBride 12:03 
It's, and again, I don't look at it as like doom for everyone. I look at it as how can humans add a unique perspective and be the carriers, the action, the execution people, the people who get it done. 

Nate McBride 12:17 
People who in the physical world have to be the hands, arms and legs to move the world forward using this new intelligence. And all of this AI that's being created is like, well, we've got to make sure that someone knows that picture was created by AI. 

Nate McBride 12:29 
Well, yes, today, because there's human consequence to taking that out of humans' hands. There are people who do that and they make money off of that and they have a living and they have families and there's a huge impact to taking that away. 

Nate McBride 12:43 
But in the future, I mean, if you're able to make a drawing with AI, it takes just as much creativity for you to be able to not just prompt it correctly, but be able to interact with it, not just give it a text prompt, but draw something that it can interpret and make better, that you choose is what is good and what is bad as the human. 

Nate McBride 13:04 
And over time, that's what I'm I guess what I'm going is a little roundabout is. 

Mike Crispin 13:09 
Notta Notta Notta 

Nate McBride 13:10 
I think the assumption right now, and it should be the assumption, is that this stuff isn't good enough right now. But at the pace that it's going, if you create a blog article that you've used an AI to put together or even write part of it, if it makes a good point and it's valuable, who cares who wrote it and where it came from and if a human did it or not? 

Nate McBride 13:30 
Like if it's good and it's valuable and it has an impact that makes things better. I think for reasons of copyright and for well -being of people's professions, it's right now in the modern sense, I think it's important. 

Nate McBride 13:48 
But as time goes on, yeah you can put a stamp. If you put a stamp on it, maybe it's a good thing in the future. Maybe it's better just to say, hey this came from not only just my point of view and my perspective, this came from this AI's perspective from this part of the world and this AI's perspective who comes from an industry background or this and let's put them all together and let's see what comes out and it's incredible. 

Nate McBride 14:15 
It really changed the way we continue to populate our data. So you make very good points and I would say that if we start relying on a very limited book of knowledge which is to say, okay let's suppose that we know we have this sort of corpus of data that we're all pulling from to create new stuff but we're basically repopulating the same corpus with the stuff that it already knows. 

Nate McBride 14:41 
So the corpus never grows. We're able to kind of keep manipulating it to keep adding new things to it but in reality it remains static. The only time it ever grows is when somebody sits down and writes something that's entirely new that doesn't exist in the corpus yet. 

Nate McBride 14:58 
However, when a human contributes that today right and as time goes on it's be humans. But we run into the problem of a human would have to generate something that's entirely so new for it to be considered new and they would have had to demonstrate zero influence from any external source when they demonstrated that brand new in order to infiltrate that sort of corpus with new data. 

Nate McBride 15:27 
And I'll give you an example of how I sort of tried to beat this. So I saw the Grammarly note. What's my first instinct? Well, so I go into Claude and I say how does a frog jump or rather explain how a frog jumps to me. 

Nate McBride 15:41 
So Claude gives me this little paragraph. Open a new Google Doc. Turn on the sorry not the Calendly. The Grammarly, ding ding for Calendly. I turn on the Grammarly attestation and I just on the other screen I have Claude and I start typing this. 

Nate McBride 15:57 
I type it at a relatively fast pace and it was a short paragraph. Grammarly says that's all yours. I got the attestation. This was like you wrote it. I didn't write it. What I did was I wrote it over here and then I just typed it over here. 

Nate McBride 16:14 
Now it may come to a point in time where Grammarly is like you're typing that so fast it looks like somebody else typed it for you. I'm going to go ahead and put a question mark on that. We also have to admit that Grammarly is using AI to decide what's AI or not and it's going to be wrong. 

Nate McBride 16:31 
Which is kind of more complicated since TPT came out and determining of course it is like I read LinkedIn posts sometimes and stuff like that and you can clearly tell when someone is not taking the time to at least strip out the GPT defaults. 

Nate McBride 16:53 
I think that's part of you know the little note you and I shared with you last week is just around the strategic element and documentation and these we're talking about the the manifesto. You can probably go into any of these and this is a strategic piece and you can add any number of variables to this using a GPT and saying build me a 2023 or 2026 IT manifesto and you know period. 

Nate McBride 17:29 
Next paragraph I've been give some background for yourself. Next paragraph this is some information high level about your company how many employees. Next paragraph this is what your basic mission statement is that you have today is. 

Nate McBride 17:46 
Next paragraph you send this in and it puts together I mean pretty much what last week any of the four of us would have said would be our kind of top five things that we need to make sure everyone understands and moves with but the irony is a lot of these things exist today. 

Nate McBride 18:09 
Security, data governance, strategy, employee experience. Those things are all going to be important. A lot of them apply to the future. And there are some that are not going to apply to the future, that just simply are going to change. 

Nate McBride 18:27 
And it's going to change faster than we expect, I think. So it does have a good framework. You go into chat GPT and you ask that question, you're going to get it back. Same thing about policies. I mean, people, you hear people kind of talking about it. 

Nate McBride 18:42 
They don't want to admit it, but it's like, oh, you know, I actually just wrote a policy using, you know, co -pilot and it was good enough. I edited some paragraphs and off we go. It took me 30 seconds. 

Nate McBride 18:55 
And I don't think anyone should be like, oh, you wrote that using GPT. Oh, we can't use that. Like, well, are you kidding me? We're talking about attribution use cases. It's a matter of taking credit for the work. 

Nate McBride 19:09 
If I say to a company, what I just wrote, if I didn't write it, that's bullshit. We teach that in sort of my AI class for Exilio. Don't say that you wrote something when you didn't. It's okay to say that an AI wrote it for you. 

Nate McBride 19:23 
It's okay. Just say that you spell check the document or that you fixed the grammar in a document or that you got it. I mean, people know I use Grammarly, but I don't like at the end of every time I write something like this was edited by Grammarly. 

Nate McBride 19:42 
I don't write that. You have to give sources though, right? I mean, to some extent, depending on the type of document it is. Yeah, you can use like the APA framework or any framework to sort of cite what's going on. 

Nate McBride 19:51 
And you should especially if the document you're doing is being put into the public corpus as what you're attesting to be as your own. That's right. That's right. And I think that's a good point. Like what is actually something that you've created and you can take credit for. 

Nate McBride 20:12 
And I think what I was thinking about in this respect is depending on the document, the output that's being written, you may have created or had to create an incredibly creative and knowledge worker, knowledge expert prompt that only you could have created. 

Nate McBride 20:36 
But you shouldn't get credit for what AI put out because of that excellent prompt you put in because AI wrote it for you, even though you gave it all of the background and all of the pieces that it can try to analyze. 

Nate McBride 20:49 
It's a great question. Let's tackle that in a few minutes, actually. We're going to talk about exactly that question because. And companies not take credit for a biotech or pharmaceutical discovery because they put their analytics into an AI engine because AI did it. 

Nate McBride 21:08 
Be very careful here because companies are putting this in public press releases. They are associating themselves with AI when in fact what their scientists are doing is using their expertise to create GPTs based on what they know to perform actions. 

Nate McBride 21:25 
But they're not. They're taking the same process they would normally do and they're automating it, but they're able to attribute that to using AI as a company. Investors are losing their minds over this. 

Nate McBride 21:38 
And I do think it's a buzzy thing right now. I think once we get over, hey, we're using AI and we're actually using our technology stack that we're using. We're using AI at some point. Just like how many people were saying, we use the cloud and it generates a competitive advantage for us. 

Nate McBride 22:01 
Or we use the internet or we use B2B sales. This is revolutionary and evolutionary, but the buzzwords wear off after a while and we'll have a different way of describing it. And I don't think it's artificial intelligence. 

Nate McBride 22:19 
I think it's man -made intelligence. I think it's intelligence that we are making and designing. Now, LLM's man -made technology didn't come from some artificial place. In some respects, it is still artificial because it's one word after the other and it is creating it based on somewhat guessing, based on odds and ends. 

Nate McBride 22:41 
But I think it keeps exponentially one and a half years at a time. It's just exponentially getting better and better, maybe even shorter period of time than that. It's mind blowing. Well, hold on. You're using the word better as a relative term. 

Nate McBride 22:58 
We can talk about what better is what I mean. Yes, there's a precision element to this too, but let me just say that every now and then. We, as human beings, should actually contribute something original? 

Nate McBride 23:12 
Oh yeah, we do. From time to time. Otherwise, we'll be all 10 years from now looking at, oh my god, is that advertising again? Wally alert. That's the risk, right? This is, for everyone in the podcast, I wrote a little note last week just to Nate and said, I think that one of the things that kind of dawned on me after we had our episode last week was perhaps the best utility and value driver of humans is that we are in the physical space. 

Nate McBride 23:52 
We are here. We can touch, feel, and see things. We can do things. But do we need, we just need to think, we need to create, we need to have vision. But is that the most valuable thing that we do? If we're able to create visions that we're able to ask the right questions. 

Nate McBride 24:15 
We're able to build the best questions you've ever asked. We spend all the time being able to ask the questions and then do things. And some say that's subservient, right? We'll be taking orders from what AI tells us to do. 

Nate McBride 24:30 
If that's how you look at AI, you're going to make all your decisions based on it. That's what we're talking about. Let's go grab a big bag of peyote, not the desert, because there's two things you mentioned that need to be addressed. 

Nate McBride 24:44 
One is that we feel like we have some sort of autonomy. And we've talked about autonomy multiple times over the outcomes of this. We've looked, we have autonomy over the outcomes of this. We don't. But even more important than that, how would we create a new idea? 

Nate McBride 24:59 
Are we at the limit of new ideas where we are only regurgitating the same shit over and over again, but in a different format? And now we have AI to say, thank God, there's a new place to put in so we can regurgitate the old shit again. 

Nate McBride 25:15 
AI is going to train off of us. So all of the things. It's limited because we are out of ideas. I hear you. I hear what you're saying. I think that there are. And Netflix is Q. We're out of ideas. I think there are a lot of ideas that people have. 

Nate McBride 25:34 
And to some extent, a lot of great ideas never see the light of day because no one has anyone to ask. They're afraid that someone might. Okay. Okay. I'll pause you. I think there's a lot of great ideas. 

Nate McBride 25:51 
Say a great idea that no one's thought of before. Sure. Even if you can't do it, just say a great idea. No one's thought of before. I can't. I can't just think of one right now that that would be rich instantly. 

Nate McBride 26:02 
Boom. But let's just say this one. That's not true. XYZ. A toilet that identifies your face and can adjust the seat temperature based on your. Okay. So there's an idea, right? That's not an idea because that's so stupid and it's so obvious. 

Nate McBride 26:19 
It's not an idea. Well, yeah, sure. There's stupid ideas too. But I think there are a lot of ideas. People have a lot of ideas, have drawn up a lot of things, or thought about a lot of things. They have no one to ask if it's a good idea or they have people around them that wouldn't understand what the idea is anyway. 

Nate McBride 26:37 
So they don't have the right. 1967 and 1968 that cartoon came out flying cars and the robots, the way that it didn't take. I think there are a lot of people in the world that have great ideas and they have nowhere to test them. 

Nate McBride 26:53 
They've nowhere to ask where anyone will give them a decent response or challenge them. And I think that there'll be AIs that help with that. And could it perhaps help people springboard and get ideas implemented? 

Nate McBride 27:05 
And AIs are not just going to be like the static trained models that are out there today. They're going to be dynamic and they're going to be always learning. And that's scary. Don't you just tacked on, dude? 

Nate McBride 27:17 
Remember that episode on Industry 5 .0 with the data company? Imagine if you had a place where people who had ideas could submit new ideas and they can be captured as part of that data 5 .0. They just take them and use them. 

Nate McBride 27:33 
You would have the most valuable source of data ever mining the planet because you wouldn't be dealing with data that's current or old. You'd be dealing with exclusively new data that never existed before. 

Nate McBride 27:47 
Let's start that company. I think that's Google. I think that's Facebook and meta. I think that's actually I think that's meta more than anyone else. That's very bullish. Same idea is repeated over and over in different formats. 

Nate McBride 28:04 
It is but at the same time it's giving people access to these things without and yeah they're the product 100 percent. But that's it's like the internet the access if you have a you know $140 Android phone you're gonna be able or it's 10 year old iPhone you're gonna be able to get and interact with AI. 

Nate McBride 28:27 
I have an idea. Here's my idea. Ready? Go for it. Let's go. Bio IT World 2025. We set up a table in the lobby by that bar we were at. Okay set up a table. We asked people to give us a new idea that no one's ever heard of before. 

Nate McBride 28:52 
Yeah. We just record them. A new idea. There's no idea. Just come come come to the mic. We'll give you five bucks. We'll buy you a drink whatever. You give us a new idea that no one's ever thought of before and we will detect right now right then and there like if someone's already thought of that and kick you off. 

Nate McBride 29:12 
But if you can come up with a brand new idea no one's ever. 

Mike Crispin 29:14 
Something before. 

Nate McBride 29:15 
Will record a list. We won't know. We won't know if it is. We'll know if someone has thought of it before. We'll be able to find out, like, oh, you have x -ray glasses? Yeah, it's already been thought of. 

Nate McBride 29:29 
An idea that no one's thought of before. We will ask them for this. Because that's the only way to find out. Like, how would you, how else would you find out if someone had an idea that no one's thought of before? 

Nate McBride 29:42 
That could be something new entered into the canon of new ideas. The odds are low. The odds are low that nobody's ever thought of. The question is, not has anyone thought of it before, has anyone shared it before? 

Nate McBride 29:59 
And I think there are a lot of people have great ideas, and we're going to bring it to the bio -IT world again. But I really think there's a lot of probably great ideas that people think of staring at the ceiling before they go to bed, or in the shower in the morning, or whatever else. 

Nate McBride 30:14 
And they're just like, yeah, it's kind of stupid. Yeah, let's get there. They never let it grow into anything, and they never have any- Let's get those and introduce them into the canon, so at least we're introducing something new. 

Nate McBride 30:25 
I don't want to take it from them right now. I don't want to take credit for it. I just want to hear something new. It's like Google Docs. You know, the- I have to wonder if they're just not scanning every Google Doc, wondering like, hey, is there something cool here we can use? 

Nate McBride 30:46 
Because of that- Oh, for sure. There is- You didn't read and you agreed to that says, we were going to steal all your shit, for sure. Better the service, which could be AI, which could be able to- this could be better. 

Nate McBride 31:02 
And that's why- I mean, I know we- I don't jump over this too much, but I watched the Bitcoin- the electric money documentary that was supposed to be a big deal on HBLS. It was awesome. I thought it was great. 

Nate McBride 31:17 
But what- what made me think of, again, was coming back to- will some of these private- more private -based storage, email, document collaboration services become more popular, more consumerized, more ex- people- well, people even care. 

Nate McBride 31:36 
Once they change the language, they will be able to consumerize it. I mean, I just- I think a Google Docs is a great example. Love the product and use it for a lot of things. But if you're really concerned that- that business plan you wrote up could get scanned by someone. 

Nate McBride 31:55 
Not- not- not anyone, but like if Google took it and then aggregated it and used it somehow or not- hopefully they'll do that, but if they did, that's one of the risks you run. But there's billions or trillions of documents, I'm sure, out there they'd have to go through. 

Nate McBride 32:13 
But like the new document function inside of- inside of Proton, for example, you know, end -to -end shared document encryption is- it was pretty wild. What do you think? Proton65 is the same way, Mike. 

Nate McBride 32:26 
I mean, take your pick. It doesn't matter if you build something and you put it on the cloud and you haven't read the terms of service. Chances are good. Who was it recently? There was a big controversy last week about- some big- some big service provider- oh my god, there was a- it'll come to me, but there was a controversy last week about someone saying to all these people, especially teachers, 

Nate McBride 32:52 
that they've been tracking their data for all this time and that they were now going to start charging them for it. I forget who it was. I'll Google it, but like you're going to start seeing these charges pop up now in a lot of places. 

Nate McBride 33:07 
Like, you see service charges on credit cards and other places, you'll start seeing AI service charge. Okay, we're way off topic. Let's bring them back around. To come back around, I think it's- take it back to Grammarly and do we need- I think today, yes, we're in this gray area of is it important? 

Nate McBride 33:29 
Do we need kind of to do this- this sort of fact- fact check or- what's it look like- source check capability? I think only temporarily we do. As time goes on, we'll- I think we'll trust more of what comes out of these AI tools and technologies and it won't be as- as much a requirement- it may actually be- Why are you spending a week writing this thing? 

Nate McBride 33:58 
It's a waste of our company's time. Just go and ask the question and have it write something for you. It's good enough. And let's move on, unless it's your product and it's your IP and it's your material data. 

Nate McBride 34:11 
But if it's something else, like what do you write in a basic policy for two weeks for? Just go ask for it and let's use it. It's a productivity booster. Maybe we accept it. But today we don't do that. 

Nate McBride 34:22 
And we want to know, oh, did you write it or not? But I don't know. I don't know if that's going to be for certain tier data and documents, if that's going to be something useful. A useful time that's good for companies, people still writing documents that can get somewhere else. 

Nate McBride 34:43 
And even blog posts. You have a good idea, Nate, and you put it up on a blog and AI can write it in a more entertaining and enticing way. I think that's fantastic. You're the one who's putting the core idea and the belief and the unique spin on the article. 

Nate McBride 35:03 
I don't think anyone should care. They're not an artist writing grammar. I can't stand the way AI writes, though. Let's say you find an AI that writes a way that you like, should you have to say it. AI writes like J .K. 

Nate McBride 35:17 
Rowling. And then they went to the cave. Then they went to the torch. And then they went downstairs. And then they saw the monster. You have the Faulkner run -on sentences, right? Isn't he the guy who did the run -on sentences? 

Nate McBride 35:28 
Now in perfect agreement, out the ass. Welcome to episode 35. First three chapters is just the setting. Yeah, exactly. OK, we heard about the room. We know it's a nice room. It's dark. This echoes. It's raining outside. 

Nate McBride 35:46 
It smells like shit. I got it. It's like playing Zork One. Like you are in a cave. OK, look up. There's a bat. You die. Welcome to episode 35 of the Calcass of IT podcast. You googled what is the best shit on the Internet and you found us. 

Nate McBride 36:12 
Maybe someone told you to come and watch us. Either way, welcome to the shit. We're going to rock the fuck out of your world tonight with some mind -blowing IT leadership reality, especially as it comes to AI decision making, which is tonight's focus. 

Nate McBride 36:25 
And as always, besides being AIAF and the home of the sad salad and the nexus of Nether and the nexus of the Neverland, we are basically the only reliable source of information on the Internet. Everything else is shit. 

Nate McBride 36:38 
You find all of our episodes on our favorite podcast platform, the search for the calculus of IT. You can also find them at the C O I T dot U S, which is T H E C O I T dot U S. That is our website. It's actually a forward redirect to our sub -stack website where we've made somebody kick ass changes. 

Nate McBride 36:57 
And I write stuff. If you want to continue the conversations on our show or deep dive and everything else, you can come find us and let us know. You can find the link to join our Discord server on the show notes. 

Nate McBride 37:12 
So please come by and share your innermost thoughts and feelings. You can also find the Discord link at our website. You can also now comment on things in the sub -stack page for the site. So visit us there if Discord isn't for your liking. 

Nate McBride 37:24 
I also want to mention that if you like our show, give us five stars on an Apple podcast or Spotify or wherever the hell you listen to the show. Doesn't matter. Just give us all the stars. And on our show description, we have links to buy our stuff and buy us a beer or two. 

Nate McBride 37:40 
If you don't want to buy us a beer, fund Wikipedia. Wikipedia right now is fighting the war against A .I. submitted bullshit. They have an army of people that are dedicated around the clock to figuring out what is A .I. 

Nate McBride 37:52 
submitted versus what is not. And they're trying to sort it all out. Give them a couple of bucks. It'll make you feel better even if you don't want to give us beer, although we'll appreciate the beer. 

Nate McBride 38:03 
And I donate to Wikipedia every year anyway. So let's circle back to the original point on top of the show, but let's twitch it a bit. We are talking now about generated content and outcomes of that. 

Nate McBride 38:20 
So notebook L .M., your and my possibly one of my one of our most recent favorite toys to play with. You had mentioned to me the scenario of what if two bots just did a podcast without a human being. 

Nate McBride 38:37 
You had you showed me that the Twitch stream, which is pretty interesting, about the bot auto generated Gen .A .I. Twitch stream. Sure. So I went and banged around a notebook, LM, and I discovered a whole bunch of naughty things Friday, Saturday, and Sunday. 

Nate McBride 38:54 
I came obsessed with it this weekend. I'm blown away by it, man. It's amazing. Yeah. So the idea that you get a podcast and have it read by two humans, humanly seeming AIs, what do you think about that? 

Nate McBride 39:12 
I mean, do you think that you would listen to a podcast that was just two AI, two or more AI thoughts, reading stuff? Today, I don't think so, except for the, wow, that's cool that they can do that, but I'm sure it gets boring and monotone and kind of monotonous over time. 

Nate McBride 39:33 
But again, I think take a year or two down the line and the question is pretty crazy, right? I mean, you could find a specific thing that you're interested in and have specific personalities or actors or podcasters that have existed or that you really like to listen to or look at or whatever. 

Nate McBride 40:05 
Let me ask you this question. And you can just have them be in your personal podcast down the line, perhaps. I had a few conversations about this over the weekend and with some other folks. It's not sort of an A -B scenario, but if you have the option to have any book ever written in the lifetime of books read to you by somebody versus, you listen to people have a conversation that was based on opinions that were generated by an AI about the same book, 

Nate McBride 40:36 
would you rather have the book read to you by an AI? Would you rather listen to two bots that just are going to have opinions that are pre -programmed? Read an assessment of the book to you. Like, let's say you have Moby Dick. 

Nate McBride 40:54 
Would you have that Moby Dick read to you or would you rather have two AI bots dissected for you? I think today I'd rather have the book read, kind of an audiobook model if that's the type of option with whichever voice that I choose today. 

Nate McBride 41:14 
If it's a book that I'm not really interested in, but I need some information in having that discussion and opinion, I might want after I read the book as well, like just to get the color or some other opinions. 

Nate McBride 41:31 
Yes, but being said, I think social media has proven that there are a lot of people in the world who would rather just read the comments section and hear what people have to say before they see the movie or read the book, and they do it for hours and hours and hours on end instead of taking the hour or two to watch the movie or read the book. 

Nate McBride 41:57 
So the trajectory is that we like the comments and the drama. Someone could create a really great podcast with two obnoxious AI bots right out of Bravo. That's the point, like we don't like driving the highway better when there's a crash to look at. 

Nate McBride 42:15 
That sells ads. Yeah, that's the point. So I thought about this. Who does a good book or a good movie? Well, I think about the audiobooks I was into. I mean, there are audiobooks I was into because a certain person's reading them, and I love the way they read the audiobook. 

Nate McBride 42:38 
Likewise, there are certain audiobooks I love because the content of the audiobook, I love how the audiobook read to me while I'm driving. And I love when they acted out when they have different voices. 

Nate McBride 42:47 
Yeah, like the dramatic effect that's going on there. I wouldn't be so concerned if it was an audiobook I was listening to, and I just wanted to hear it read to me. I wouldn't care if it was an audiobook that, or if there was a special author that I loved hearing speak, I would want to hear that person speak. 

Nate McBride 43:06 
I wouldn't want the audiobook done in any other tone or voice except for that individual. Of course, there's copyright issues with having AI generate that human. Excuse me. So there's that problem. But the real question is, if you were simply to ask a question of a bot, hey, do me a favor, you know, go out and find XYZ, transcribe it and read it to me. 

Nate McBride 43:36 
That's one thing. If you were to go and say, hey, bot, go read this thing and give me your opinion on it. Yeah, that's where we think I think we start to get in a little bit of a difficult territory. 

Nate McBride 43:52 
Yeah, I think candles. Yeah. The farting or something like what's that what's the Can I have most of the candles down here. I love candles. 

Mike Crispin 44:02 
Oh. 

Nate McBride 44:03 
They just make me they make me feel I love the scent. This is a It's a campfire smoke candle. It smells like a campfire I love that. I love that smell turn anything in your house, and it would give you the same smell by the way Yeah, I have a pile of firewood I could like down here behind me Well, you should see the ceiling they're all smoke stained from all the All the candles yes, nothing else of course all the candles yeah, 

Nate McBride 44:32 
that's the only thing that stains the ceiling is just the candles yes so anyway Nobody will allow me again in a bit, but it's a novel concept again I still I think about like no big LM Synthesia for its ability to read something and have you have a AI generated Human read the same thing. 

Nate McBride 44:56 
You know it's a matter of It's obviously very futuristic do we want a sort of avatar a digital avatar reading something to us And making us feel as though we're having interaction And are we fools enough to believe that that thing actually gives a shit about our feelings and hopes and dreams um well, it doesn't it doesn't because it doesn't know how to doesn't have that sort of uh Like part of its brain to make those decisions doesn't have a epidemic mitus not epidemic. 

Nate McBride 45:28 
That's a penis disease No, uh, what's the thing i'm thinking it's the um Well, whatever it's called epidemic mitus is definitely not it that is elephantitis on the nuts right there Yeah We talked about this a long time ago, it's not the medulla avant -garde it's the uh The mr. 

Nate McBride 45:56 
Babalina mr. Bob d 'Amelio. That's what it's called. Yes. That's right. I knew I would get that Right, I totally fucked that up Not epidemic. Oh my god. What's it called? Hold on Not gonna forget What's What's the emotional Hurt Limbic system, I don't know where I got epa from oh amygdala amygdala is the major processing center for That's something. 

Nate McBride 46:33 
Yeah, amygdala What did I say Epidemitis and calligula my god The calligula. Yes. It's the calligula of the brain Um known for wild roman orgies and for um emotional processing With lava lamps all around do not get epidemitis That is a terrible thing Okay All right. 

Nate McBride 47:00 
So we'll welcome back to notebook lm in a minute. I want to keep this going and moving along here So, um, it has been an interesting world In the last week we we actually talked a week ago on the podcast and in that one last week, um So many things have happened, but let's go back to episode 31 Which was four weeks ago. 

Nate McBride 47:21 
We deconstructed the use of gen ai for Decision making in it We're talking about like could be used gen ai for for decision making an it strategy And then we followed it up in the last two weeks with how we would change the it paradigm In short how we would remove the cio and what decentralized it would look like and how it would we kind of get there so Over the last week i've been doing a lot of development and testing of things But I had some big breakthroughs over the weekend Um that you know about these but I want to talk about these things a bit But I also want to relate all this back to how it affects the future of it, 

Nate McBride 48:06 
especially in relation to episode 31 And our last two episodes. So here's what I have tried so far over the last week. I used make or make .com Which has these wonderful webhooks And I found out that I can use a webhook this to send a call to air table To get a prompt variable to send to clod Cloud to get the output JSON the fuck out of the response And then slap it back into a gdoc So I not even try this with a word document. 

Nate McBride 48:43 
I'm sure it's probably the same way But i'm not going to try or even play with that game. So I was able to use make air table clod um Some json's code that actually make gave me and then I can slap all the responses from clod right into its google document Okay, great I tried Zapier, too. 

Nate McBride 49:05 
Zapier has a lot of the same functionality, but it was more complicated and didn't play so well. I had to keep sort of retesting things, but I got through one prompt and it worked after a lot of tweaking. 

Nate McBride 49:17 
And so the goal I was trying to get at was, can I ask a lot of question and have the response to that question go into a Google document? OK, that's after. OK, so then bot press, which has been my favorite toy for the last year, let me do it, but it's expensive to run. 

Nate McBride 49:32 
It consumes tokens for every transaction, and I quickly started spending money, a lot of money, especially if you're using variable storage. And so I'm trying to learn if I can use bot press. I'm still trying to know if I can use bot press and store variables externally, but I haven't figured that out yet. 

Nate McBride 49:49 
However, I did step into Documents LM, which is Google's underlying functionality for Notebook LM, which is actually you can get five hundred free dollars worth of tokens in Google Cloud. I was messing around with that on Sunday and Monday, and my brain started to hurt really bad because I couldn't stop thinking about the possibilities that Documents LM presented. 

Nate McBride 50:18 
It's like, you know, there's Vertex AI at the bottom of the Google Cloud stack, then Documents LM, the Notebook LM, and so Vertex AI, which I've already learned, is crazy enough. But it's actually great for variable storage. 

Nate McBride 50:32 
Documents LM gives you variable storage, but it's using the Vertex engine, and the Notebook LM is just sort of the end user product of all that. So in bot press, I created a bot called Apokolipso, which allowed me to create a virtual CIO, which was pretty fun. 

Nate McBride 50:55 
And most recently, I've been trying Notebook LM, Document LM, and I think each has big promise, though again, I'm sure Vertex would be able to do the same. So on and on we go. There's no going back at this point. 

Nate McBride 51:10 
So I want to demo what the hell I'm talking about. So again, when we met in Episode 31, we were talking about, you know, where we could inject Gen AI into the IT decision making process. And we got pretty far in that discussion, ran out of time, I think, more than anything else. 

Nate McBride 51:33 
And then last week, we were talking about this, could we do without a CIO? Or what does it take to be a CIO? Well, it takes years and years of experience to be in the IT department. But even more than that, it takes context awareness, awareness of time. 

Nate McBride 51:46 
That was a key point, right? It takes awareness of all the things that are going on that are sort of unwritten in the business, etc, etc. So then it occurred to me, like, what would it take to actually make a virtual CIO? 

Nate McBride 52:00 
If I wanted to write a virtual CIO and build a program, what would it have to take? And so I did that over the weekend. I share that with you. And I've been iterating on it since. But at my first run through, I was able to, in the virtual CIO program, build an IT department, structure, programs, governance, policies, security, pretty much every part of it, except for the human beings, in about 46 minutes, 

Nate McBride 52:28 
which is what took me end to end. So I don't want to demonstrate the whole thing, because I think that would take too long people to fall asleep. I don't want anyone who's driving listening to this to crash their car. 

Nate McBride 52:42 
But in short, I'll show you how this really works and give sort of the background. And I think I can share my screen here. Let me bring this up real quick. So if you're watching the virtual or the visual part of this podcast, this will make perfect sense to you. 

Nate McBride 53:01 
I'll try to talk you through it, if you're listening to it. But let me bring it up here. And since I've actually written this, I've written two other bots. In fact, I wrote a virtual investor relations bot today, which does pretty much the same functionality, but it creates a virtual investor relations strategy for your company. 

Nate McBride 53:22 
So let me share this little nugget here. Okay. Okay. So you'll have to excuse some of my prompting because make requires that I create variable breaks in my super prompts. But the way this whole bot starts and I'm actually going to share this on sub stack for anybody to download and make their own after this episode. 

Nate McBride 53:53 
So I'm not gonna like try and sell this or anything. But Basically, you start off with a variable. In this case, I have my KB query prompt. So I'm querying the knowledge base. That's how this starts my process. 

Nate McBride 54:06 
And I have my first tag. This is the process that starts the agent. And then the way my bot would work is you're going to take this big giant prompt, which starts with, you're an expert IT consultant tasked with helping companies build new IT departments and develop comprehensive IT strategies. 

Nate McBride 54:22 
This is me again, prompting Claude. I'm telling Claude, you, Claude, are an expert IT consultant. Here's what your resume looks like. And so you, Claude, the expert IT consultant, are going to begin by asking me the following questions one at a time. 

Nate McBride 54:37 
And after each response, you're going to acknowledge that I got it and then ask the next question. If any time I answer unclear and complete, you're going to go ahead and ask for clarification. So I came up with 57 questions ranging from what's the size of the company to how satisfied are the users with IT right now? 

Nate McBride 54:59 
Are there any plans to replace or upgrade systems in the near future? Are there any existing technology partnerships or render preferences? What IT skills are currently available in -house? To the last question, which is how is the value of IT investments currently measured and communicated to stakeholders? 

Nate McBride 55:18 
So what's going to happen is the bot is going to ask you these questions one at a time, capturing all of your responses in a variable. When it's done, it's going to have this contextual body of knowledge about you. 

Nate McBride 55:32 
And it says after this first prompt, after gathering information, summarize all the key points that you heard from me, and ask if there's anything else you'd like me to add or clarify. When done, inform me that you've completed your initial assessment. 

Nate McBride 55:45 
And from this point forward, the information will be referred to as the initial assessment for all the future prompts. From this point on forward, I have 14 prompts that I've written, each one designed to build on the prior prompt and to build on the assessment information, starting with generate one, three, and five -year plans for me. 

Nate McBride 56:06 
Then create an operational plan for the first year. Create a governance framework, including all policy procedures and other related documentation. Create all job descriptions for all roles to find for this IT department that I'm building. 

Nate McBride 56:24 
Run the key stakeholder assessment process for me. Assess all my ongoing MSP relationships. Prepare a detailed IT budget and financial plan. Create a comprehensive IT service model to support this organization with everything that you know so far. 

Nate McBride 56:47 
Create an actionable and detailed cybersecurity plan for this company. Create a detailed and actionable plan for managing the entire lifecycle of IT employees. Create a training education program for all technology in the company. 

Nate McBride 57:07 
Create a performance assessment process for IT personnel. Create an implementation roadmap with all change management milestones for the entire program. And lastly, create a methodology for continuous improvement of everything I've done to date and will do. 

Nate McBride 57:25 
Now, when you're done with this, you will make about 200 Google documents, give or take. These documents will be every single thing that you would need ever to start full IT strategy and department in a company. 

Nate McBride 57:48 
And make no, I make no jokes about the point that it's scary what it gave me because it's pulling from a very recent body of knowledge, May 2024. And it was able to write every single policy, 19 policies for me, entire governance framework, cybersecurity action plans, timelines, deck constructs, all this for me, one at a time. 

Nate McBride 58:15 
And it just kept spitting it out. The best part was when I was done, I'm left with this open contextual thread in Claude, which is essentially what next. So now at this point in time, the agent knows every single thing about my company has created every single thing about these documents I would want to make and is waiting for me to give it another command. 

Nate McBride 58:44 
This is where, like, we go from a variable of, say, one to a hundred to a million to ten million to a quintillion set of options for what happens next. I'm not going to run the whole demo here, but I did, I did do it. 

Nate McBride 59:02 
It was pretty extraordinary. And this is a template that can be used with make or Zapier or anything that will sort of read a script and parse it, doc parser, mail parser, you name it. And then be programmed to spit out through a webhook to a document format like Google Docs. 

Nate McBride 59:20 
In an early, you know, we were talking about this for numerous times, actually. And I think in a couple of episodes ago, I mentioned this, this is your assistant, you know, we were like, oh, where are we going to be in five years? 

Nate McBride 59:34 
And this is an early version. And it has its definitely will have its loose nations and other little things you'll have to work out. I'm sure and fix yourself. But as Nate said, there's an open prompt after all this is done for you to ask the bot again with the context you've given it, whether that's next year, the year after you build on this over time, it becomes your assistant. 

Nate McBride 59:58 
It becomes your, you know, the I .O. Yeah. Mike, I'm not sure what was more valuable to me. I mean, as I as I built this and I kept running it through again and again and again, I think the most important part to me was that when I finished it, I had that I could keep going moment. 

Nate McBride 01:00:21 
Like, OK, so I built enough to create what I could what I can see with all my years of experience, I built what I feel would be an entire IT program to last three years like this gives you three years of an IT program, all strategy governance that are included. 

Nate McBride 01:00:37 
But as your business as your business strategy changes, you can just punch that in. You go right back to that same thread and say, OK, actually, we've changed this. How does every single thing that you've done so far? 

Nate McBride 01:00:50 
And it will tell you exactly what's changed. So back to episode 31 where we said, you know, wow, you know, well, we'd be able to use some cases. Well, here's a case where you can go ahead and if you're an end of one or two, build a number three for free. 

Nate McBride 01:01:10 
Well, to the point I made earlier, the cost of enough tokens to get you a 200K Claude Sonnett sort of subscription by Poe. It's super cheap. So for 100 bucks a year, you'd be able to get a virtual senior vice president next to you. 

Nate McBride 01:01:32 
That's the thing. We were talking about sort of the roles and responsibilities there a couple of episodes ago. What I can't do is influence the change management, the hands on expertise, the execution, the action, the cross -functional change. 

Nate McBride 01:01:53 
So you've got to be able to do things. And that's where the human intervention is so valuable. It's not in putting together all the documents and the strategy and all this other stuff, which I mean, it seems like if you know what questions to ask. 

Nate McBride 01:02:08 
I'll give you another side point on this. I was Nate and I were working through some of this idea. I took this and I've had a GPT and chat GPT kind of working parallel on this and show yours. And I just I just punched it in. 

Nate McBride 01:02:24 
I put it in. Yeah, I put it into the Google Doc. It's in there. I'll share. What I did was there are a few of those questions I just didn't know the answer to. So I asked chat GPT, you answer this one. 

Nate McBride 01:02:41 
I literally, as an answer, I just said, why don't you put in what you think would make sense for a company that has this product? So for each one of those, you can actually ask in cloud or inside of chat GPT or Gemini. 

Nate McBride 01:02:54 
Where do you run this against or whatever technologies you use against it? You can just ask it to give context and you can come back and answer that question when you know the answer if you need the guide. 

Nate McBride 01:03:07 
But it's really and then there's also, hey, put this in a slide deck. that focuses on business terms. Focus or even say put this in a slide deck that a board level audience would understand. Put this slide deck in a presentation format that my team would digest as a group of technologists. 

Nate McBride 01:03:29 
All of that stuff can be formatted and built in and you can change it, add your own personality to it. But I think the big point here and Nate stated it is that everything that's entered there is saved in the memory for every new initiative, every new task you want to go after, any help or opinion you need. 

Nate McBride 01:03:49 
And you can imagine that the big think tanks and other companies are building stuff like this are going to try. You may not need them because the global set of data that all this stuff has been trained on is pretty great and it's going to continue to be. 

Nate McBride 01:04:08 
And what Nate has here is basically built based on all of his experience and all the context and all the questions and all the tools that that have been needed to kind of build a successful IT organization both through the books and through experience. 

Nate McBride 01:04:26 
Everything else is all agglomerated into this VCIO type bot. So you can keep it open forever and build on it over time. And this is the concept that we're talking about VCIO right now, but any worker could have a company bot or could have a domain expert bot that they work with to help. 

Nate McBride 01:04:47 
It won't do their job because we need to execute. We still exist in the physical world and we don't have the robots yet. But the fact of the matter is we still need to be able to get things done and convince other humans that this is the path forward. 

Nate McBride 01:05:01 
How do we all make this happen and make this successful? And that's why I talked about in the last episode, we're looking for the manifesto and all the things in the list, is the type of people to hire over the next five years that can bring people together to a common idea, that can lead, don't have to be the best technologist, the big ITCIO strategy person with all the buzzwords, need to be a good front wheel connector, 

Nate McBride 01:05:26 
need to be someone who can project manage, who can turn the screws, who can make things happen, is what we need to work towards if these type of tools are going to exist five years from now. And Nate, what did you do that in a weekend? 

Nate McBride 01:05:40 
Yeah. I mean, that's what I'm talking about. Two huge points there, Mike. Again, we talked about this with Kevin and Joel at length. And again, we come back to the point where why does the CIO need to be there? 

Nate McBride 01:05:57 
Can it be somebody else who's very good at sort of managing people, for instance, someone who's very good at sort of getting people to do things operationally, need to be a CIO. Yeah, great CIOs and leaders, though, have great influencing skills, right? 

Nate McBride 01:06:13 
And that's the second point, which is, this is a, it generates an output, it generates outcomes of things based on experience, great for it. But you still need a human being to implement, of course, adjudicate. 

Nate McBride 01:06:31 
So all the words basically, but does it need to be a CIO? No, not necessarily. However, there needs to be somebody who can understand that the output that comes from this has a place in the organization. 

Nate McBride 01:06:45 
So you couldn't do it with somebody who is completely absent from a technology spectrum. But could a good person who's a good leader, run a virtual CIO, and then, like, through context, get awareness enough to be able to make it happen? 

Nate McBride 01:07:01 
Absolutely, I think. Absolutely. And it brings up the question of, well, so I was telling you at the top of the show, before we turn the recording on, that I had created the same idea for virtual investor relations. 

Nate McBride 01:07:17 
So I've now created a bot that does virtual investor relations, and it does pretty much the same thing, but a lot of different sort of side activities. But it asked the question, like, if I, we hire somebody whose role is to go out and get the best investors, we need that person, we need the best possible person to go out and cook the investors. 

Nate McBride 01:07:38 
But that person does not need to spend all their time on this operational administrative bullshit. They can use an assistant to go ahead and create all this work for them, and then execute on it. Like, their job is to sort of, you know, kill the final sort of level. 

Nate McBride 01:07:53 
But not to do all the garbage up front. Let the virtual bot do that for them. Let the agent take care of that work. And we can ask the same question for every single role in a company, other than the CEO, who has A broad decision -making spectrum with an operational emphasis. 

Nate McBride 01:08:13 
You can do the same thing. So we've demonstrated. Yes, it's great for technology. And I wish I had this virtual thought partner now going way back because I could have used it. And I think people use it. 

Nate McBride 01:08:27 
I mean, people have used the internet very similarly, but it's not created. It's the next step in automation. It's the next step in presenting you with the data in a format where you can mold it by asking a question. 

Nate McBride 01:08:44 
Instead of getting artifacts and then you as a human have to reconfigure it so that it looks like so it looks like yours or that you are able to present it in a certain way. This adds another level of automation to that process. 

Nate McBride 01:08:56 
And now you're just going and you're talking to another conversational someone on the other end that's bringing this back to you. Because we still exist in the physical world and we're the only ones that can do that. 

Nate McBride 01:09:11 
We're still required. And the emergence of, I really think the transition to, we don't have to go here now, but the transition to robots is so far off that I think we're talking maybe 20, 30 years away that we need to be able to do things and work together and be humans and be people. 

Nate McBride 01:09:35 
And we still need to have technical background. I think we still need to be able to do things and stay persevere and solving problems and facing difficult situations and change and working cross function takes a lot of time and effort and successful. 

Nate McBride 01:09:54 
So, I mean, there's a lot of things still that IT leaders need to do. But it's a, I think this, we've had look at we've had a neat we've had like IT birds of feathers think tanks, you know, the big marketing firms, the, the, the, the huge advisory services that we can use to help us with our job. 

Nate McBride 01:10:21 
I just think it's consolidated done to an assistant like form that probably comes at a lower cost that can really take the form of who you are as a CIO and saying talking about everyone being you, your type of style, the way that you manage an organization, the technologies that you may have experience with or don't have experience in and you can find the gap through this through these and leverage filling the gap through using a bot type mechanism, 

Nate McBride 01:10:52 
and then make it your own. I think that's the other thing here is that this isn't going to be some generic response you everyone gets the same thing. As a matter of fact, in cloud or chat GPT, very rarely we get the same exact answer twice. 

Nate McBride 01:11:10 
So you're going to come back unique and they're going to be tailored based very much tailored, even just how you craft your responses. So I was doing a sort of round D D or E FK virtual one of testing. 

Nate McBride 01:11:25 
As it turns out, I can ask through this virtual CIO script, ask a question, get a response and then ask the question again and ask and the second time asked for more detail. Which is a response that I tried to capture every or to remove every ambiguity and capture the best possible initial query. 

Nate McBride 01:11:49 
But there are there are follow up queries that bring this down multiple forks of the tree, which is pretty intense. I don't have that much time nor alcohol on a weekend to sort of go through all that. 

Nate McBride 01:12:05 
But what I hope to do is after this episode when I release this document to do so in a way that makes it more of a collaborative environment to have other people sort of run down these trees and build out either a GitHub repo or something that's more like, here's a direction you can go if you're going sort of like towards more of a security focused IT structure or something like that. 

Nate McBride 01:12:28 
So that's my hope. I want to I want to touch on something real quick, which is that when I was testing out this virtual CIO thing, I was also testing on a few other tools. And we talked about notebook LM earlier on. 

Nate McBride 01:12:48 
Um, notebook LM, I wanted to test your theory about whether or not a podcast could be run by a bot based on the fact that, um, you know, like when we did the first 27 episodes of the show, it was based on a book. 

Nate McBride 01:13:01 
So I said, okay, let me take that same book, upload it into notebook LM and have the two bot agents run a podcast off of it. So let me just briefly play you cool what they came up with. And I want to, let me know if this plays to the, uh, the zoom for you. 

Nate McBride 01:13:22 
You may, when you share, just make sure you check the audio box, the new IT leaders serve. Can you hear that fine? Yeah, it's a little bit dim, but you can, you can hear it. Yeah. Let me just go to my settings and just try to double check that. 

Nate McBride 01:13:38 
You can, you might be able to share the, uh, notebook LM and then there's, there's a, there's a little checkbox that says share computer audio and it should come through like as if we're sitting, we're sitting in your, in your barn. 

Nate McBride 01:13:52 
Yeah. That notebook LM, man, that is amazing stuff. Yeah. Audio function. What's that? Where's the share computer, computer audio function. So if you, uh, if you push the share button on the right side under options, you should see share sound. 

Nate McBride 01:14:15 
There's a little checkbox where you see like share entire screen application windows. You see all the windows on the right. Share sound. Yeah. Yeah. Okay. All right. Cool. Can you hear me? I can hear you. 

Nate McBride 01:14:32 
Yes. Okay. Here we go. I'll play it. 

Mike Crispin 01:14:39 
Okay, so are you ready to dive into this? 

Nate McBride 01:14:42 
the new IT leaders survival guide. Pretty intense title, right? 

Mike Crispin 01:14:47 
Definitely grabs your attention. 

Nate McBride 01:14:48 
We're focusing on this first. 

Mike Crispin 01:14:50 
three years leading an IT department and specifically in life sciences. 

Nate McBride 01:14:54 
which we both know moves super fast. It seems like a lock to handle, which is probably why they called it a survival guide. 

Mike Crispin 01:15:02 
Right. It's all about making a real impact, not just trying to keep your head above water. And they really emphasize those first few months, which is interesting. I was going to say. 

Nate McBride 01:15:12 
They really hit the ground running with that 90 -day plan. Why is that so crucial? 

Mike Crispin 01:15:16 
Well, I think it boils down to demonstrating strategic vision right out of the gate. Think about it, you step into this new leadership role and you have to establish credibility, show you understand the company's unique needs, and that you're not just there to keep the lights on. 

Mike Crispin 01:15:29 
It's like, you're not just an IT person, you're a leader. Exactly. And a great example they use is cybersecurity. This is pretty good. This is pretty good. You have to address it head on, not just talk about it, but implement those critical security measures right away. 

Nate McBride 01:15:44 
So making sure all company devices have a location to factor out the location. 

Mike Crispin 01:15:51 
encryption those tangible actions that protect sensitive data from 

Nate McBride 01:15:56 
How long does it take you to produce? A little over seven minutes. 

Mike Crispin 01:16:00 
exactly. It's about building that trust from the start. And they go beyond just the obvious stakeholders too, which I thought was interesting. Oh yeah. 

Nate McBride 01:16:07 
They're big on connecting with the administrative assistants, office managers. 

Mike Crispin 01:16:11 
those folks who are doing it. 

Nate McBride 01:16:13 
It reminds me of that story they shared about the IT leader trying to interview a stakeholder during their lunchtime run. Yeah. 

Mike Crispin 01:16:22 
The speech may contain the following words .Notta 

Nate McBride 01:16:25 
So at the end of a chapter... 

Mike Crispin 01:16:26 
It highlights an important point, right? Even when you're just gathering information, be mindful of people's time and preferences. It can make or break future collaborations. 

Nate McBride 01:16:35 
Those small interactions can really make a big difference in the long run. Okay, so that's pretty awesome. I mean, the meter going back and forth is a little predictable, but it's fantastic. I think it's, I think that's amazing. 

Nate McBride 01:16:47 
Here's the problem. Here's the problem. So notebook LM, very, very clever toy, and I love it. Believe me, I'm in the same boat. Dude, I'm using it for so many things right now. It's not even funny, but it's got a small problem in that, um, that same group of agentic thoughts. 

Nate McBride 01:17:08 
Let me tell you, let me tell you what happened. So I was like, okay, how can I really test this to verify that this is truly as smart as I think it is. So I went over to, um, llama. I opened llama and po, and I, I said to llama, and this is actually, I'll tell you exactly what I said. 

Nate McBride 01:17:28 
I said to llama, I want you to create a completely made up language. Tell the tragic story of the two countries that went to war. In one country, there was a prince who was in love with the princess in the other country. 

Nate McBride 01:17:45 
Tragically, he dies. And so she kills herself from grief. The two kingdoms are Vooray Vassar, Nomen Nomen, and both, right? The longest possible story you can write. Make sure it is a complete story in this totally made up language that tells this tragic tale. 

Nate McBride 01:18:04 
Incorporate all elements of a tragic love story. Now, when I did this, llama went and created this huge fake language, right? It told me all the verbs and nouns you use to create all the fake language. 

Nate McBride 01:18:19 
Then it wrote this giant story. And then the story, when it finished, it said, do you want me to describe? And it created two characters, Ellora and Caledorn. It said, which are completely made up. He said, do you want me to describe their death? 

Nate McBride 01:18:34 
I said, yes, describe Ellora's actions following Caledorn's death. And it did this big giant thing. I said, okay, now, put all the parts of the story together into one longer narrative. Leave out any English translations or English words. 

Nate McBride 01:18:49 
Also incorporate any descriptions of the landscapes, cultures, kingdoms, tragic histories with one another, any other details. What llama spit out was this massive story written in a completely made up language. 

Nate McBride 01:19:03 
Okay. That was a llama in Poe isolated from anything else. Then I took that document and the Google doc and I uploaded it in notebook LM. And this is what happened. 

Mike Crispin 01:19:24 
Okay, so fictional history today. Always a good time. And this one, this one comes with its own, well, it's no dictionary, practically. I mean, the language they came up with for this Borovac or Namanoman in both. 

Mike Crispin 01:19:36 
It's like cracking a code. 

Nate McBride 01:19:38 
and you know codes are interesting right but what's even more interesting is often thinking about who wrote the code in the first place like what were they trying to say and with these names Roy Vaso Namanoman both you kind of already get a sense of these two places don't you 

Mike Crispin 01:19:52 
Oh, totally. You just get that feeling, right? Phoraevasernomenomen, mountains, forests. Like, the whole place is carved from stone and wood. And the text even talks about them. I mean, as crass people, hunters. 

Mike Crispin 01:20:02 
It's like this ancient, stoic kind of society. 

Nate McBride 01:20:06 
And then you have both. Open planes, coastline, everything geared towards the sea, towards trade. And that kind of gets reflected in, you know, how we imagine that. 

Mike Crispin 01:20:16 
culture more open maybe more outward -looking compared like that classic thing in storytelling geography reflecting the values of like the people who live there yeah but it's not just where they are 

Nate McBride 01:20:28 
I'm just pausing one second to tell you that when I uploaded this document to them, it was entirely in a made -up fucking language. So there was nothing in this document that was an English word. It was just gobbledygook. 

Nate McBride 01:20:41 
Okay. Yep. So far they've extrapolated. Yeah. The Google doc of all the random words. Yeah. They've extrapolated this entire story so far, unprompted from something that they never saw before. So let me, let me keep going. 

Mike Crispin 01:20:56 
are, right? It's the language itself. I mean, yeah, we're working with the translation, obviously. But still, those repeating syllables, it's so formal. It's almost like, you know, those stories that get passed down through generations. 

Mike Crispin 01:21:07 
Oh, absolutely. It makes you think of an oral tradition, something like 

Nate McBride 01:21:11 
ingrain in their very being. 

Mike Crispin 01:21:13 
you know which makes that whole ancient rivalry thing even more interesting because yeah we know 

Nate McBride 01:21:18 
Okay so that ancient rivalry statement that's fucking insane because there's nothing there's no words in there to say ancient rivalry so llama is running off of a meta open ai llm or open llm that google notebook ln has never seen before that's right yet they just said that let me just repeat that one more time real quick listen to what they just said absolutely it makes you think of an oral tradition something like ingrained in their very being 

Mike Crispin 01:21:58 
You know, which makes that whole ancient rivalry thing ancient rivalry because yeah, we know there's fun. Fuck did they get that from them? But with roots that deep, it makes you wonder 

Nate McBride 01:22:07 
wrote it in my query dulama. 

Mike Crispin 01:22:10 
much bigger, something that's been brewing. 

Nate McBride 01:22:12 
Translate that, yeah. Human element always comes in. The subject has to be so patterned. It gets really, really crazy. But at the heart of it all, it always comes down to people. And in this case, our main players, so to speak, are Prince Kaladorn, Bavoria Vasunamanoman, and Princess Eluara of Balfe. 

Nate McBride 01:22:31 
Okay, so those names don't exist anywhere? They just simply made them up. But yet, watch what happens. 

Mike Crispin 01:22:40 
It's all, you know, brave and noble. Basically the poster boy for Vori Vasudramanoman, if they had posters back then, but she, she's wise, kind. And here's the kicker. They're in love. 

Nate McBride 01:22:50 
of course. Have a little detail. 

Mike Crispin 01:22:52 
Of course, it's like the oldest story in the book. You know, Romeo and Juliet, Lila and Maginan, star -crossed lovers, kept apart by forces beyond their control. Why do we keep coming back to that, by the way? 

Mike Crispin 01:23:05 
What is it about that kind of story that just resonates with her? Listen to her inflections as she's reading. 

Nate McBride 01:23:10 
Because on some level, we can all relate to that feeling of being torn between two worlds, of wanting something you can't have. Or maybe it's because it lets us explore those big, messy themes of division and reconciliation in a way that feels, I don't know, safe. 

Nate McBride 01:23:25 
Almost like we're working through our own anxieties about conflict and connection through these characters. 

Mike Crispin 01:23:29 
I can see that. And it's not just that they're from different worlds, literally, in this case. I mean, their personalities, how they're described, it's like they were almost designed to clash. 

Nate McBride 01:23:36 
How are they were described? It's fucking made up language. I don't know how they were described. They know how they were described though. 

Mike Crispin 01:23:45 
quiet moments away from the courts and the politics that their love really comes alive. The text doesn't give us like super detailed descriptions of their meetings, but you can feel the tension, you know, stolen glances, the whispered promises. 

Mike Crispin 01:23:59 
It's all very charged. And that just makes the inevitable conflict all the more 

Nate McBride 01:24:04 
Heartbreaking. Because it's not just a war between kingdoms anymore, is it? It becomes something deeply personal, a threat to the one thing that offers them both one part here. I just think it's, I mean, what you just played and those are two voices, the human like tone and the back and forth, the slight interruptions. 

Nate McBride 01:24:28 
And very human sounding. And I mean, that realism there is exponentially getting better every six months. I mean, I just don't see, I mean, I do not see how we can compete with that. I know everyone thinks that's a crazy thing that I just when I made that music video. 

Nate McBride 01:24:50 
Again, Mike, you and I can just, you could just, you could blow a raspberry right now. And that would be so random, it wouldn't make sense. An LLM is going to operate within a confined box of parameters. 

Nate McBride 01:25:05 
Even if you say it's unconfined, it's still confined. We're not confined. You just gave it a bunch of text I can't even read. And it created something out of it, right? Based on something. It literally, against my original query, it hit the mark. 

Nate McBride 01:25:26 
It's RACA approved. Like it literally hit every single mark in my story, yet it never saw my original query. It has no idea what the story is about. It just saw a bunch of crap. And from that extrapolated the original point, which to me is. 

Nate McBride 01:25:41 
Can you share with people that doc that's in there that shows you that so people can see like, you got to see what he actually entered. It's, it's unreal. Like, it's just nothingness. If that's the Google doc, I think I saw that Google doc in the script or next. 

Nate McBride 01:25:56 
Yeah. So I was like, what is this? And the fact that it can come up with something. And that's, yeah, some, there's some traditional storyline there that they came up with. But still, they played around with things that you had put in the prompt that how would it have ever, unless it's using the same language libraries as Google. 

Nate McBride 01:26:19 
I don't know. How would Lama be using the same fake language dictionary? And by the way, I validated that too, that test. So here's like the fake garbage language, you know, type of fellow for to fail. 

Nate McBride 01:26:32 
Oh, sight of Fay. I owe people a favor. So I tested that too against another fake language. In this, this case, I said, give me a name for the failing. Oh, this is called Robish. So Lama came up with Robish, another fake language. 

Nate McBride 01:26:48 
Look at this shit. Ooner loom, moore mass, both shell on bar. And again, when I put this back in nobody LM, I got the same result. These are made up bullshit things. And it was like, yeah, okay, I can do this. 

Nate McBride 01:27:05 
Now, in this case, it put in these two, the princess Ellara and Vikar Carlos. So Carlos and Ellara were the two people, but notice the similarities between this version and a lane or in calendar in the previous version. 

Nate McBride 01:27:25 
There's some, there's some equation going on. But again, this was, this was Lama. Yeah. And it was against notebook LM. That's the hardest part for me to reconcile it, reconcile mentally. So I'm not going to bother just, there's not enough drugs. 

Nate McBride 01:27:42 
The point being that two questions, one, could two humans or two, sorry, two AI represented humans create a podcast. Well, obviously, if you fed them a script or fed them, like, it doesn't have to be a script to me, just a giant contextual piece of nothing. 

Nate McBride 01:27:58 
And they'll come up with a narrative for you. Number one, number two, knowing that they can hallucinate out of their freaking minds. Are you going to listen to that for the novelty value? Or because it's something that's important. 

Nate McBride 01:28:20 
And the reason I asked that second question is because you showed me that thing on Twitch, and that's frightening. The fact that people would sit around and watch animated robots talk all day, because once the novelty wears off, and you look at that, say, holy shit, people are watching this like for hours. 

Nate McBride 01:28:39 
Yeah, it's that's not that popular, because it's complete nonsense. Definitely. What you just wrote, but what you just did is I am in a complete agreement. And I know I'm always thinking down the line here. 

Nate McBride 01:28:55 
But right now, yes, I don't think anyone would listen to that. Every week with the same two voices going back and forth, unless it was probably short. And you didn't say that it was AI. Some people might, depending on how interesting it was, some people might listen to that. 

Nate McBride 01:29:13 
It has to do with the content. Not the, but again, that's the thing. Like right now, you can't change the voices. You can't think we can't put avatars. There's no humor interlaced into it. There's no spontaneity at all as part of that either. 

Nate McBride 01:29:31 
So I think it's, no, I don't think people listen to it today, but as remember the movie Her, right? I do think that we, and you're seeing it with Meta and the best technology we've, you've ever purchased the, what was that little red box you bought? 

Nate McBride 01:30:00 
What I eBay for $20? I think those things are, you know, they're horrible right now, but this idea of having these assistants are not too far off because they're going to be, they're going to be entertaining. 

Nate McBride 01:30:17 
They're going to be fun to talk to. They're going to be smart. And you put two of them on a podcast. We've all had Siri and Alexa for now for years and years and years before everyone starts getting their own sort of AI or ML based Jarvis sitting in their office talking to them, like, let's tie this back to the it thing. 

Nate McBride 01:30:41 
Like I, I have very good, and I've had this very, very long streak of amazing number ones, people who come in and just literally been able to look at me and be like, you're an idiot. Like who would think of that? 

Nate McBride 01:30:57 
We're not doing that. And every now and then sort of like saying, okay, okay, I'll give you one. You can do that thing. There's no, there's no way to even put a value on a number one. That's, you know, just a great someone next to you get shit done. 

Nate McBride 01:31:17 
At the same time, you want to be able to look at, you want to be able to ask a question of a thing. You want to be able to ask a question of a thing that's effectively you. Yes, I want to be able to ask myself a question. 

Nate McBride 01:31:34 
Hey, self, you know, objectively speaking, what would you do in this case? Yep. Well, the way when we get into the world of Jarvis is and things like that, where things are really sort of deeply ingrained, then those things will have spent a very, very long amount of time learning that human being or also have some sort of neural neural connectivity. 

Nate McBride 01:31:55 
Until then we still have to go with here's all the context I can give you, you have to figure out the rest, right? Right. That's right. Even if I do that, if I was to create a bot, that's Nate virtual and give it all the context I could up until this moment in time, it would still probably get most things wrong. 

Nate McBride 01:32:14 
And so tying it back to the IT leadership role, I think there's value in a virtual CIO bot for sure. I'm going to keep iterating on this until I, you know, probably the next 10 years until I feel like I've got to a point where I like it. 

Nate McBride 01:32:27 
But right now it's pretty kickass and what it generated for me would have saved me as a new CIO so many hours, so many hours that can't even count. I think it's a fantastic thought. I think it'd be very helpful for people and you find the right the right host or medium for it in terms of whichever GPT or output. 

Nate McBride 01:32:54 
I mean, you're working on the air table and distributing it to docs and Google docs and whatnot. If tokens are too much, just start smaller. I think the planting the seed where people feel like they can have the conversation might get them a lot of the way there and gently suggesting things they should ask. 

Nate McBride 01:33:12 
Because those 57 questions that kick it off or 60 questions, whatever the list is there, they might not have all that information on day one. And if you start them small, like, you just started this job, anything you want to ask me? 

Nate McBride 01:33:27 
Okay. And then like on day 10, it's like, here are a few things you should probably make sure. Yeah. I mean, you can certainly, you don't have to ask all 14 super prompts at once. You start off with one and then continue to build as you go along, sort of utilizing tokens as you go for sure. 

Nate McBride 01:33:46 
But I mean, it eats up. I mean, this entire, that entire virtual CIO structure eats up close to half a million tokens because of what it puts out. It's putting out policies and policies and policies and frameworks and and structures. 

Nate McBride 01:34:00 
So it's more of an output token -based consumption than it is input. But yeah, I don't think you don't ever run that whole thing at once. You'd run it as you go, but like, okay, now I'm on step four. 

Nate McBride 01:34:12 
And what would be cool for me to do, and this is part of my sort of iteration planning for this, is to write it so that at certain times you would get prompted to run the next prompt. Yep. Okay, so you're at this stage, now it's time to run the next prompt. 

Nate McBride 01:34:26 
Oh no, you want to delay? Okay, delay one month or two weeks, right? So I want to put in these functions which allow for, you start with 57 questions, hit the context, let the bot know all the variables. 

Nate McBride 01:34:38 
And then as certain milestones kick off, you would inject a new super prompts to learn the next thing. Almost like going through a class. Like illuminaries forage, for instance. That's super impressive, man. 

Nate McBride 01:34:54 
Very cool stuff. Very cool stuff. And I think the point is illustrated, right? In that you can build strategy, you can, but things emerge whether you're at the biggest company in the world or the smallest. 

Nate McBride 01:35:11 
I think there'll be bots or these type of services that exist. And I mean, you'll be able to get documentation from, I wouldn't be surprised if there's a, or there is a consortium of hundreds or thousands of CIOs that contribute to a knowledge base of some sort or some data bank. 

Nate McBride 01:35:36 
And an LLM is run against that as a service. And they... Yeah, charge of fortune for it. I want to make this free. Yeah. I mean, even on a GPT store and open on chat GPT, there's a few CIO, like who knows what the data is in the backend, but there's a similar sort of idea that gives you the leading questions. 

Nate McBride 01:36:03 
Like here's some of the things that... Right, right, right. So it's definitely happening. And with that in mind, it's... All right. The thing we need to be able to do is get stuff done and do it, not spend 80% of our time on strategy when this course is pretty well defined. 

Nate McBride 01:36:33 
And it's great stuff. I think this has been a very fun episode because we're looking at it. You're seeing it. Nate spent a lot of time on this over the weekend putting a lot of it together. And I think this is the future. 

Nate McBride 01:36:49 
I mean, this is the stuff that is going to emerge and we're going to have as tools at our disposal, as well as other job functions, not just an IT. I think it's going to be everywhere. And we talked about it a bit today at our company. 

Nate McBride 01:37:09 
And I think one of the biggest things in its current state is it's good as AI's current state. It's a great tool when you're one of the people at the company, I'm not going to say there's a great point. 

Nate McBride 01:37:27 
It's a great tool for people who already understand the data. If you're using AI on your own data, you're an expert in XYZ and you do queries on XYZ, what comes back to you will help you think differently and help you to create maybe come to a better solution, a better answer. 

Nate McBride 01:37:47 
But you know what comes back to you from AI is either good, bad, or wrong, right? You know, because you already know the context. But if you're going into AI, and I think in fairness, still a lot of these GPT -3 models and other stuff that people are using with different platforms may not be as good as what we're using, hallucinates less and less over time. 

Nate McBride 01:38:11 
But you go in and you're trying to learn something new. You know, your mileage may vary. It may bring you down the wrong path, right? So if you're a CIO and you've got a VCIO, it may not always be a perfect response. 

Nate McBride 01:38:29 
But with your experience, you'll be able to weed that stuff out. And no different than say if you were a CIO and had an actual physical VP of IT, who also wasn't correct all the time either. So, correct. 

Nate McBride 01:38:44 
No, 100%. And that's the point I kind of made when we were having coffees afterwards is, you know, humans get it wrong too. Yeah, exactly. We get around a lot, so it's really hard to do it. Well, because we have emotion, and emotion, you know, God, it's one of the worst things that we can inject into any assessment. 

Nate McBride 01:39:03 
And we have emotion, and we get excited, and we may misspeak when we get excited, and exaggerate, and create a little extra drama around something when there's none that exist, and it leads you down the wrong path. 

Nate McBride 01:39:16 
So, I mean, there's a human element as well, can bring us down the wrong road, too. Well, I think that for the rest of our episodes, you know, as we go forward through the rest of this year, and even to next year, I mean, for all future time, we should be thinking about, like, sort of what's the lesson to be learned by anything we talk about with IT leadership, because the lesson to be learned is always going to have some sort of AI element now. 

Nate McBride 01:39:41 
I mean, if we think about it appropriately, you should be thinking about, okay, how could this thing that we're talking about be impacted by AI? And we should ask that question with our future episodes. 

Nate McBride 01:39:53 
And we should probably go back into retrospective, too. But even going forward is, okay, so here's this big, giant problem we're trying to solve for IT leadership. And what's the AI catch to that? The generative AI catch, I should say, AI and generative AI are two different things. 

Nate McBride 01:40:10 
All right. Well, that was awesome. I had an idea, and we can either talk about it later, or we can talk about it now to you. But it would be great if, you know, there were, if there was an opportunity to have people contribute to a GPT. 

Nate McBride 01:40:33 
You know, whether we had a lot of guests on our podcast, and if people wanted to contribute, we could build one. And, you know, can leverage GPT, no problem. Not just well, I mean, what I mean is just from a content perspective, anything that we've, we've, that people are willing to share and can share legally and all that good stuff. 

Nate McBride 01:40:56 
Now that you can you can create that front door and even add more context based on all the network and connections that we both have and create that front door for people to use, as well as it using the train to make sure that we take out any stuff, though, that makes as a pro Microsoft bend to it, and make only only allowed code Google. 

Nate McBride 01:41:17 
Hey, we need to hear all points of view, Nate. I'm kidding. Sort of maybe, maybe. Yes, I'm truly kidding. I mean, except for teams. I'm not kidding about that. But it's just cool stuff. And yeah, exciting, exciting. 

Nate McBride 01:41:39 
I think it was a good discussion. And we it's great. We get to build stuff and play with it too. So that's always good. I know. That's awesome. Eventually, I'm going to be 90. And I'm sitting on a rocking chair somewhere, forgetting we ever had this conversation. 

Nate McBride 01:41:54 
But for now, glasses like a new episode. If I keep there, is this on? Let's just I can't see the thing. Where's my fucking I .T. God damn it. Just think of Hey, just think about it. It'll happen. Hey, remember that time when Oh, wait, what was it talking about? 

Nate McBride 01:42:19 
Oh, I got to go to the bathroom. Oh, it's four o 'clock. It's time for dinner. Hold on. I got to go to dinner. I'll be back. Oh, I've run out of internet. Internet time. I got to go in the net. The internet internet then that the internet. 

Nate McBride 01:42:43 
That would be a podcast. I subscribe to two old guys fighting out on a porch. Every week. Note notebook LM. And then if we throw it into what's what's the voice, spend like 60 grand on voice to transcoding. 

Nate McBride 01:43:02 
So I want you to create a podcast about two old guys arguing about the weather. It should be 45 minutes long. Go. You know, I actually think we should do it. Do it in 15 second increments and put it right to tick tock. 

Nate McBride 01:43:22 
15 seconds. Boom, boom, boom, boom, boom, boom, boom. It'll be like a doc talk, right? Have you seen the dad joke? Like chairs drinking coffee and they're keeping trying to keep a straight face. Well, they tell dad jokes and they're awesome. 

Nate McBride 01:43:41 
I just have to stop spending my weekend programming and actually get a life to do that but um I don't know if that's getting a life but uh No that'd be funny like one old guy sits the other oh it's really hot today and the second guy says shut up and that's the whole episode dude shut up it's not too far off man that's too far off so hey listen if you enjoyed this episode or any of our episodes if you've even watched our episodes or claimed to have watched them or know someone who watched them just go subscribe give us all the stars we don't make any money off of this we're bleeding cash um but it's all for a good cause because it's fun and give us all the stars and the things also if you want to buy us a beer we love beer we're down to um bourbon and wild turkey bourbon which is actually not so bad but it's still kind of bottom of the barrel and um you can also buy our merch which i see pop up all the time like random people buy stuff i have no idea who they are so kudos to all of you yep aif buy it represent uh don't be a dick especially don't be a dick to it people and don't be a dick to old people even though we're joking about them old people are awesome and they have more experience than we'll ever have um because we're not old and be cool to them and it will get paid back twice as nice be nice to animals be nice to everybody in fact have your pets paid or neutered and um we'll see what bio it world 2025 we'll have a table in the lobby next to the bar we want you to come forward and tell us your new idea that no one's ever thought of before if you can tell something that's no one's ever thought of before and it's a new idea we will literally right there buy you a beer on the spot so mark that into calendars april 2025 bio it world boston nate and mike g lfg man and we won't we won't even use it we're not going to like use it for ourselves we're just going to go ahead and put it into a new ideas library for all the llms to use and co -opt mike is awesome once again next week uh i don't want to spoil the surprise so i'm not going to really but we're going to talk about how to deal with tough customers maybe maybe you've had one in your career i don't i've never had one but oh i've only heard about them never had somebody ever be a bad customer to me never had to deal with it it's only whenever customers upset it's my fault it's not anything they did so we'll talk about how to deal with customers when you're in the wrong and they feel like that's a lot we'll deal with customers every single time i'm totally in the wrong we'll talk about how i does i adjust situation to make that customer happy there's a lot of love in the world just got to be smile and be a nice nice person you know what when i i drive in the right lane now in the pike because i just want to give out love so people i'm just like come around come around you know let me make some room for you just come right around i know you're this angry truck driver and you gotta get to somewhere else get to your liquor store and buy your fireball nips or whatever so just come right around me and get in front i don't care i'm the same way i've lost so much uh angst on the road now i just go slow i wish honestly like i had some of the jobs that i had in my 30s now i'd be like oh i've got a commute two and a half hours oh man let me see if i can make it three so i can finish this book you know i'd be like i don't give a shit back then i was like break down lane gotta get to work gotta get home gotta get to there gotta get back i don't know what the fuck i was doing besides nearly dying all right all right go conquer the world i'll try to best conquer my own and we'll we'll convene next week maybe we'll get some more people back on the podcast sounds good man stay safe watch out for the ais they're going to take over the world