Calculus of IT - Season 3 Episode 11 - Shadow AI vs Shadow IT

Newsflash: Shadow AI is just Shadow IT with better branding and worse consequences.

We set out to compare Shadow AI (people using ChatGPT at home because you won't give them tools at work) versus Shadow IT (people using Smartsheet because your approved system doesn't meet their low quality standards), and discovered they're pretty much the exact same problem. BLUF: if you don't have governance, training, and approved options, you've already lost. Your employees are prompting away on personal accounts and there's nothing you can do about it except train them properly and give them legitimate alternatives. Nate revealed his company trains 100% of employees on GenAI within their first week, runs daily tests to see if Claude's LLM has been "whitewashed", and has a nine-level certification program where personas don't unlock until level six. We discussed the high probability that 80-90% of employees at most companies probably use personal AI accounts, suggested the solution is "lightweight guardrails" and multiple approved options. We also learned that Mike cooks frozen burritos (aka shit logs) in an air fryer and we agreed that TestiCoin might be our best option for a retirement plan. Beard Watch officially launched: Mike's two-week beard progress is "barely visible" and resembles dirt splashed on his face. Next week we're diving into CWP (Cakewalk Pro? Build-first culture? Both?) and why low-code/no-code is the newest (and lamest) Shadow IT.

Listen at thecalculusofit.com • Join our Slack board at thecoit.us • Leave five stars • Invest in TestaCoin • Stop dry shaving while driving.

—Nate & Mike

Transcript

Season 3 - Episode 11 - Final - Audio Only

===

Trance Bot: [00:00:00] The calculus of it,

season three,

verifying this identity.

Sometimes you just have to take it.

Sometimes you just have to take it

because it's season three divided Autonomy,

verifying identity.

The calculus of it.[00:01:00] 

Nate McBride: Uh, hold on. I gotta clean up this little mess here. Okay. Oh man, that doesn't help our sponsorship with, with Lawsons.

I can't hold onto there. Shit. I know. Oh, yes. Uh, tonight's episode. Ooh. It was brought to you by, oops. Hold on. I have another spill. 

Mike Crispin: Oh crap. 

Nate McBride: Well, the problem is when you got a spill with beer, [00:02:00] it also gets under the can in that little like, happy heidi place that, and you keep putting the beer down and it keeps making more spills.

'cause it's, it's angry with you. Okay, hold on. I got it. All right. Uh, yes, uh, it's, tonight's episode is sponsored by Lawson, by Lawson's. 

Mike Crispin: Really? She gets 

Nate McBride: free 

Mike Crispin: stuff out of it or what? 

Nate McBride: They gave me free, um, I, well I got their mailing list that was free. 

Mike Crispin: Oh, good. 

Nate McBride: I'm on their mailing list and that didn't cost me anything.

And so I get to now get their emails for free.

You look at Princess Leia by the way. What's, what's, oh, I do what those things on your head. 

Mike Crispin: I put my, my AirPod Maxes on tonight because my other ones are charging. 

Nate McBride: So hold on a second. Those are also AirPods. 

Mike Crispin: [00:03:00] Oh yeah, these are like the, the AirPod Max ones, the big headphone ones I got them a few years ago.

Nate McBride: Is, is there a product that Mac makes and seriously now, is there a product that Mac makes that you do not own? 

Mike Crispin: Um, that's a good question. Let me see. Oh, 

Nate McBride: apple. Apple makes, I should say, 

Mike Crispin: uh, I do not have a Mac Pro and I don't, I don't have a Mac studio. But you, you have two of those. 

Nate McBride: Yes, I do. 

Mike Crispin: Um, I've had every iPhone, I have the Vision Pro.

I've had every other, almost every iPad at one point in time. Not every iteration, but one on each. You know, that's the thing I probably use the least, but, um. Air. 

Nate McBride: So why air are you using those special, special 3-year-old edition Mac Pro Ear Pros? What are they called? 

Mike Crispin: These are the Air, [00:04:00] air Pod Macs, I think they're called.

So one of the, one of the other reasons is that like we're in the Airbus for two hours or whatever, like after a while, the ears get a little painful. So I, these are nice and comfy and they sound good in my ear. And you sound very rich. Your voice sounds very rich. I like that. 

Nate McBride: Where, where does Max fit in?

Like, remember I, on the, on the, on the Slack board, I put the correct order for shitty product. Refreshes is digital advanced pro business Premier Gold, platinum Enterprise, enterprise plus platinum, plus Elite Unlimited ai one. Where does Max fit in? 

Mike Crispin: There's so many different versions of them, right? 

Nate McBride: Is is, is Max after pro before business?

Mike Crispin: Um, max after Pro before business, 

Nate McBride: or is it after Platinum Plus before Elite. 

Mike Crispin: I'd say it's after Platinum Plus. [00:05:00] 

Nate McBride: Okay. 

Mike Crispin: But yeah, before Elite. I like that 

Nate McBride: I to update my list. So, so it goes Digital Advanced Pro. Like basically if you're a company that sucks or you're looking to refresh your brand, just add one of these words in this order.

First, digital Advanced Pro Business. Premier. Gold. Platinum Enterprise. Enterprise Plus. Platinum Plus now Max Elite Unlimited. And then ai. Just add ai. You could add AI to anyone. The other ones too, like Max ai, elite ai And then of course the granddaddy of 'EM all is one. Just use one. Did 

Mike Crispin: you, did you hear that they're planning Apple's planning on bringing out a, a MacBook That's $700.

Nate McBride: You've mentioned this in the last two episodes at at least. 

Mike Crispin: I'm just very excited about this and I don't know why. I'm 

Nate McBride: very excited. Why is it, is it an M four chip with, with two gigs of ram? Like what's the deal? 

Mike Crispin: It's gonna be a, an a 19 chip. It's gonna be an iPhone chip. I [00:06:00] guess they're like throwaways for 600 bucks.

You might as well buy a hundred of them. Just put 'em in the closet. 

Nate McBride: Well, they gotta get that Walmart crowd involved, you know, they gotta get the, uh, the other half of the country involved in the purchasing process for it. 

Mike Crispin: I'm gonna get one of every color and just put 'em up on my wall.

Nate McBride: Wait, there's colors. It's like, it's like imax circa 2000, sorry. 1999 imax. 

Mike Crispin: Yeah, I think they're gonna do like the, the bubble blue 

Nate McBride: cherry, what was it? Blueberry 

Mike Crispin: cherry. Plum blueberry. Yeah. I bet they were 2000 

Nate McBride: lime. 

Mike Crispin: Line. You're gonna bet those. I, those IMAX were the best. Those things were great. 

Nate McBride: Yeah. Hey, how come there's only one?

But how come there's only one button in my mouse? Well, um, 

Mike Crispin: you know, which way is up and which way is down. It's a circle. You don't even [00:07:00] know which way you're supposed to move the thing. You can move it up and it goes down. It's like, well, it's upside down. 

Nate McBride: We had this debate, oh my God, this past weekend. It was almost divorce level debate.

Mike Crispin: Oh, 

Nate McBride: no. I, I don't, I don't use natural scrolling that that's on os or any platform. I scroll, so when I'm scrolling down, it goes down and up it goes up, you know, but nobody I know does this, including my, my better half. So, um, I had fixed her laptop and I gave it back to her, and she's like, nothing's working.

I can't scroll. I'm like, no, no. I fixed it. You're in, you're doing, you're scrolling better now. No, no. Like, I wanna scroll down. I'm like, but if you scroll down and go up, that's, that doesn't work. An hour later it was fix the fucking scrolling. So it goes back to the way it was before. And I, I lost, anyway, um, I said, [00:08:00] tonight's episode is brought to us by Lawson's.

I'm gonna have a sip of my sip of sunshine. 

Mike Crispin: Very good. Enjoy it. 

Nate McBride: Mm. And I just inhaled a huge burger and steak fries from, from CJ's. So I'm gonna have two. Where's 

Mike Crispin: CJ's? Where's CJ's? 

Nate McBride: You used to get sad salads from CJ's in Sudbury. Ah, well, yeah, 

Mike Crispin: yeah, 

Nate McBride: yeah. When I would get, um, how come we don't come out here anymore, Mike?

What's, what, what, did I do something wrong? I am, 

Mike Crispin: we're gonna, we're gonna start that up again. We gotta do that. Why don't we, why don't we plan on doing it when it gets a little warmer route?

What was that man? Is that you? My goodness. 

Nate McBride: Wow. That's my part. 

Mike Crispin: I was trying to figure out, like, I'm looking around like, did I leave something running in the background on my Mac or something? It's [00:09:00] like, did YouTube just spontaneously start, or I don't 

Nate McBride: No. 

Mike Crispin: My goodness. You were 

Nate McBride: getting the oral experience of my burp.

Mike Crispin: Oh gosh. 

Nate McBride: That was actually a clip of, from a Revenge of the nerds. 

Mike Crispin: Ah, 

Nate McBride: from, from booger at the Burp contest, but I was being immature and pedantic. Um, okay. I will burp though in a moment because I literally inhaled that burger and fries in three minutes. I downed a beer while I was eating it. Okay. In between bites because I, because like I was, I was trying to shove it down my throat so fast, it was getting stuck.

So I had to chug beer to push it down and walking out to the barn, I think I, I scared a fox or a fox family out of Aden. Oh my 

Mike Crispin: goodness. You're gonna have some gas pains. You're gonna be in trouble. 

Nate McBride: No, no, no. I'm gonna let 'em all out during the show. This is my, um, my, my burp perver. [00:10:00] My burp adversary.

Mike Crispin: Really? You burp once a year? 

Nate McBride: No, this is the night I burp all the time. I save all my burps for one night. I let them let them loose. By the way we got props. Uh, your theme song. 

Mike Crispin: Yeah. 

Nate McBride: So Chris bot text theme song for the trans bot for calculus of it podcast is getting, people are re like listening to it.

That's all they listen to. They don't listen to rest of the podcast. They just listen to the theme song over and over. 

Mike Crispin: Really? 

Nate McBride: Yeah. We need to release a, um, an EP of the 

Mike Crispin: theme song. I'll put it, I'll put it up in, uh, put it up in SoundCloud for everyone 

Nate McBride: on the SoundCloud. 

Mike Crispin: Yeah, we could put a link out there for everyone.

They wanna listen to it. 

Nate McBride: All right. So if you're a fan of DJ Crisp Tech, not only is he playing the end of the World Party, um, coming soon. 

Mike Crispin: I'll, I'll put it up [00:11:00] under, um, Mike Crispin.

Nate McBride: Okay. 

Mike Crispin: That's my, that's my secret DJ name. 

Nate McBride: Well send me the links. You're gonna put it in the, in the show liner notes. For those that wanna listen to the full set, 

Mike Crispin: I'll, I'll, I'll do a remix. So where do these, where do these props come from? 

Nate McBride: Uh, we have loyal listeners who lo, who are listening to, again, not the podcast.

Just the intro music on their commutes. 

Mike Crispin: So do you think that it's just that they get bored of the song and shut it off before we start? They they just listen to the first 

Nate McBride: few minutes? No, they just don't want to. Who would want to? I mean, no one wants to fucking hear us talk. They just wanna hear the music.

Mike Crispin: Alright, I get it now, now I understand what's happening. 

Nate McBride: So what I'm thinking is, Alison, I am not an influencer or an atomizer or whatever they, um, ascend attendance or whatever [00:12:00] they call themselves these days. Yeah. Uh, but I do, I do know that if we interject random songs throughout our show, then people are more likely to stay versus just us talking for two hours straight.

Mike Crispin: So that's why you inserted the burp a few minutes ago? 

Nate McBride: No, that was a burp. I 

Mike Crispin: did. Oh, okay. Oh, that was a real burp. Yeah. 

Nate McBride: That was to promote our sponsor, Lawson's Sip of Sunshine Beer. Unchanged for at least five years. Same meal. Can same beer, same experience, same. Getting up at 2, 2 33 and four in the morning to pee experience.

Mike Crispin: Fantastic. 

Nate McBride: I'm drinking it late at night, so, um, I had to figure out a way, by the way, to clear my, my burger fat and fry fat coughs out of the [00:13:00] podcast. 'cause I feel like I have this giant heart attack aortic event happening in my throat right now. And when I have to clear it, it happens right in the microphone.

Mike Crispin: Oh 

Nate McBride: boy. But, but Dscr doesn't have a old man mouth clearing filter. 

Mike Crispin: Yep. 

Nate McBride: So I gotta figure this shit out. Um, 

Mike Crispin: well if it makes you feel any better, I had, so at at work today, they had breast cafe. Uh, come in. The one cafe, breast cafe. Have you had 

Nate McBride: breast? Breast, cafe? 

Mike Crispin: Breast? Breast? Not breast. 

Nate McBride: Breast. 

Mike Crispin: Breast.

Breast. 

Nate McBride: How do you spell that? 

Mike Crispin: P as in Paul, R-E-S-S-E-D. Okay. Anyway, so they brought in these 

Nate McBride: to the pre the pressed cafe. 

Mike Crispin: Yeah. They brought in all these burgers and I grabbed one very free. [00:14:00] Oh yeah. They're double burgers. They're fricking smash burgers. They're amazing. They're incredible. So, you know, they put out the food like at 1145 and I ran over and I grabbed one and went Did you 

Nate McBride: literally run?

Mike Crispin: I pretty much did. I, I skipped over, literally on Gallop. Okay. 

Nate McBride: Okay. 

Mike Crispin: And, uh, I galloped over and I grabbed, I grabbed a burger and it was a, a, um, a double burger. Bacon on it, and I ran, I ran back in my office and shut the door and ate it as fast as I could. And then I opened the door when everyone started coming in and I grabbed a plate and I got another burger and, you know, had my lunch.

Nate McBride: So you had, you had, 

Mike Crispin: so I jammed it. I, I had four patties in like 15 minutes. 

Nate McBride: You had a pre 

Mike Crispin: burger, so I'm, I'm with you. I'm with you. I know exactly how you feel right now. 

Nate McBride: Yeah. 

Mike Crispin: Except, except for I've, I've already cleared out all the junk out of my throat. [00:15:00] 

Nate McBride: Yeah. I haven't cleared it out yet, but I, I didn't have a pre burger.

I kind of left my office and drove the CJ's as fast as possible, which was about a 30 minute drive. And then I got my burger and the guy said, and this was my mistake, the guy said, do you want any ketchup or anything like that? I'm like, no, no, I'm good. But I still had 15 minutes to go to get home. And I'm like, well, how am I gonna eat this in the car without ketchup and salt and stuff?

That. So I had to wait to get until I got home. Oh, I've just 

Mike Crispin: seen it. 

Nate McBride: I would've just shoved it in my mouth in the car. Then when I got home, I realized how much it was, like it was so many fries and such a big burger. I was also grateful. But then I didn't have any ketchup at home, so now I was regretting, not gassing, ketchup.

From there, I would've given like gimme 15 packets of ketchup. So I get home and I'm like looking through the fridge, like what sauces I have. 

Mike Crispin: Yeah. 

Nate McBride: I'm like, okay, what's the sauce least likely to kill me? And there was a like a hot [00:16:00] sriracha, scha, ssri. How do you say that? 

Mike Crispin: Sriracha. Sriracha. 

Nate McBride: Sra. Sriracha. You skipped the first R 

Mike Crispin: shit.

Nate McBride: Sriracha. 

Mike Crispin: Sriracha, I think. 

Nate McBride: Scha. Okay. Well, I had some s, 

Mike Crispin: it's just like, it's just like teriyaki sriracha. 

Nate McBride: S. Cchi, I put, I put Cchi sauce, like all over the burger. And then I took a big bite, and then a couple things happened all at the same time. Um, first of all, I had to like drink a whole beer pretty much at once because I couldn't film my face.

And then, but then what do, like, we talked about this a couple weeks ago, but then what do I do? Like I'm doubling down. I'm like, okay, well, I'm already sweating and like everything hurts, so I'll just have another bite. But then somehow everything [00:17:00] hurt more. So then you're kind of like, you're all in, right?

You've pushed your chips into the pot at this point. You can't bluff your way out of this one. So I'm like, okay, well fuck it. So I put the shichi cchi sauce all over my fries, and I just went all in. 

Mike Crispin: Wow. 

Nate McBride: And I think my, my fingers are still numb on the fingertips, but I got it all. I got it all down. I enjoyed my beer for, for what it's worth.

I mean, it was cold and I feel like this mega Burp is brewing. Like it's gonna be epic. And I think well, 

Mike Crispin: let's, you've had a little bit of time for it to digest now, right? So you must Well, 

Nate McBride: well, I, I hand to give people a Lawsons are sponsor who I can say that with quite con, quite a big amount of confidence that the Sip of Sunshine will unleash my, my Inner Burp.

Thank you Lawsons for being our sponsor. [00:18:00] 

Mike Crispin: Thank you. Thank you Lawsons. 

Nate McBride: So good. So tonight, Mike. Whoa. What's that behind me? Oh, I thought it was my shadow. Geez. 

Mike Crispin: Yeah. Yeah. 

Nate McBride: What that behind you. We go, it's your shadow behind you. Watch out. Um, I'm thinking about Shadow it. Shadow it, those guys. Women and they, 

Mike Crispin: are we talking about shadow AI or shadow it?

Or both? 

Nate McBride: Both. 

Mike Crispin: Oh boy. 

Nate McBride: Shadow it. I mean, we can deal with that, right? Shadow ai, that's the new frontier. Remember the guy who's the bounty hunter who goes out looking for people all over the world In that movie we were gonna make, who's going to look for, um, people that have the knowledge about, you know, stuff, [00:19:00] 

Mike Crispin: stuff.

Nate McBride: Shadow AI is the arch nemesis of that guy. And then, I hate to say it, Mike, but I think that you might have a leakage problem 

Mike Crispin: most, I think like 80%. If 90, not 90% of employees probably just use personal ai. 

Nate McBride: Do you feel like you have a leak, a leakage issue? Do you feel like there's some leakage going on?

Mike Crispin: Everyone does. Everyone has a 

Nate McBride: leakage issue. No, I don't have, I don't have any. I mean, I, I had it resolved last year. I went and saw a person about it, but no more leakage on my end. You have leakage? 

Mike Crispin: Oh yeah. I have a ton of leakage. It's just been horrible

ton. It's been just terrible for me. Why do just have to talk about these things? And here I'm trying to grow a beard and you're, and you're making fun of my leakage. [00:20:00] 

Nate McBride: I thought we already acknowledged that a beard is out of your grasp. 

Mike Crispin: Look, look at this. There's no beauty there. No beauty. I'm, I told you three episodes ago or two, whatever it was, that I was gonna start that day and I haven't shaved since.

Nate McBride: That was last season. 

Mike Crispin: No, that was two weeks ago. Two weeks ago. 

Nate McBride: There's nothing there. I couldn't, I couldn't even literally. I couldn't grate cheese on that, Mike. 

Mike Crispin: See? See there's nothing there. I mean, it's here, but it's, it looks like I, you know, I, you know, the, the magnet. Get some, get some of that stuff.

Magnet toy with the Yeah, the toy. You, when we were kids where you drag the magnet pieces on face. Yeah. You met the 

Nate McBride: guy with the guy with a beard. 

Mike Crispin: Yeah, that's what I look like right now, like with just little pieces of magnet on my 

Nate McBride: face. All Can you let it go? Maybe two or three years and then we'll see how it develops.

Yeah. Try me like this right here. This is four weeks and it's, 

Mike Crispin: I'm gonna, I'm gonna try and 

Nate McBride: hold on 

Mike Crispin: Zoom 

Nate McBride: [00:21:00] migration. Mike, you have a question? You raised your hand. 

Mike Crispin: Yeah, I, I'm gonna lower my hand. Alright, good. So Zoom is commanding me to lower my hand, so I'm gonna do that. 

Nate McBride: I've got, hold on, I, I gotta clear this error.

View Mike's hand. 

Mike Crispin: We noticed you have spoken, so we will lower your hand in two seconds. Okay, good. 

Nate McBride: Well, I'll, I'll I lowered your hand for you. 

Mike Crispin: Thank you. 

Nate McBride: What a piece of shit. Okay, go ahead. 

Mike Crispin: So, so we'll be working on this. This is a real goal of mine. I'm gonna let this thing grow and see what happens. 

Nate McBride: Yeah, you gotta let it grow.

Mike Crispin: So it could, I I'll get the beard you have in like six months. 

Nate McBride: Hold on. There's an immersive view. I can put you in a cafe, right? 

Mike Crispin: Can you put me, you can put me in a cafe. 

Nate McBride: Can you, uh, are we in a, 

Mike Crispin: at us we're inside of a kitchen of some sort or, um, 

Nate McBride: a barista. Uh, sir, [00:22:00] please. Uh, two more copies. 

Mike Crispin: I'll have a, oh, a vanilla bullshit or whatever it's called.

That's a, that's one of, that's one of my favorite. Larry David. 

Nate McBride: Yeah. 

Mike Crispin: When he goes to Starbucks, I'll have the vanilla bullshit. Coffee and milk. Coffee and milk. Who? What a thug. What a delicious drink. Coffee and milk. I love that. Then the other one where he's trying to open up, he's trying to open up a GPS and he can't open.

It's like shrink plastic and he's smashing it with the knife. 

Nate McBride: Yeah. I sent you that video of him with the scissors 

Mike Crispin: trying to open 

Nate McBride: that package. Yeah. 

Mike Crispin: That is me. Like you get, I get so mad, like throwing stuff, so 

Nate McBride: hold on. Right, right behind you is Zoom Coffee. Let's see. 

Mike Crispin: Wow. 

Nate McBride: Let's see. 

Mike Crispin: Is that that a fishbowl or something behind us or a Hold on.

Something there. Yeah. 

Nate McBride: Wonder what else we can do here. We can do a, hold on. We're in a theater. [00:23:00] Mike, pass the popcorn. 

Mike Crispin: Wow, this is amazing. 

Nate McBride: Hey, you know, the best part about popcorn during a movie is finishing the entire thing before the movie starts. That's my favorite part about popcorn during movies. 

Mike Crispin: I do that all the time.

I eat all the candy, all the popcorn, everything. And then it's like, yeah, I need more. I, I, sometimes I get up and get more, especially popcorn. 

Nate McBride: I wonder what we, we, well, we didn't see Wicked this year as part of our annual trip to, um, the movies. Maybe next year we'll see something like, um, 

Mike Crispin: frozen Three 

Nate McBride: Pro.

Yes. That's literally the top of my list right now. 

Mike Crispin: Just make sure you bring something that you know, you don't have to use the bathroom. 

Nate McBride: Oh, hold on. That one's from the diaphragm. Excuse me, everybody, [00:24:00] if you're listening, hopefully 

Mike Crispin: complete. Do you wanna just start this one over? 

Nate McBride: No.

Listen, there's probably like seven more behind that, so, oh, 

Mike Crispin: I love it. 

Nate McBride: There's nothing I can do. 

Mike Crispin: This is so good. 

Nate McBride: Yeah, this 

Mike Crispin: is so good. 

Nate McBride: All right, 

Mike Crispin: so shadow AI or shadow it. 

Nate McBride: Shadow 

Mike Crispin: ai. 

Nate McBride: Shadow ai. Yeah, so, so I think this is kind of a no brainer as to why shadow AI is happening, because anyone, including people we talked about last week, our elders can sign up for a generative AI account with literally no thought.

Mike Crispin: Yeah. 

Nate McBride: In fact, Claude probably makes it the hardest. What's your phone number? And people are like, oh, shit. Fuck's my phone number. You don't even know it. You gotta look it up, type it in. 

Mike Crispin: Sign in with Google, sign in with Apple. [00:25:00] Yeah. Boom. I have an account. So you're in pretty quickly. 

Nate McBride: Yeah, I sh I shadow it'd Gemini today.

So I was like the shadow AI against Gemini today. Yeah. I was qui quizzing it about why it was so reticent to gimme details about things related to gems and it kept pushing back 'cause you wouldn't understand and there's too much detail and it's not for this conversation. And I was like, are you fucking kidding me?

Just tell me what I want. Do. 

Mike Crispin: Do you realize that I didn't realize this about gems until I started playing with it. It's like it's technically under a different privacy agreement. 

Nate McBride: Yeah. 

Mike Crispin: Even if you have, even if you have an, uh, an enterprise account, it's like you had to read that shit. 

Nate McBride: Yeah. So, so because gems are using what?

Opal? 

Mike Crispin: Yeah. Well if you choose opal's, even another one, that's the brand new one, right? 

Nate McBride: Yeah. But you can, so the way that [00:26:00] artifacts. So this is, this is interesting, right? You just knew that this was gonna happen. So Claude comes out with artifacts and then artifacts are meant to be used as steps in a workflow for projects.

So you make a bunch of artifacts, put the manure project and other people run the project. No problem. Gemini is like, fuck, all we have are gems. We don't have something cool. So then what Gemini does, or Google does is say, well, Geminis can become parents and other gems can become children. And so you can have a parent gem talk to children gems.

Yeah. And invoke them. Invoke them by name to do certain activities. And this is kind of mind blowing to me. So I tried this out today and sure as shit, you can have 10 different gems that are all, do one activity and then a parent gem that calls 'em an order based on priority. And that priority, I gotta upload a [00:27:00] simple.md script as to which call which one, based on which scenario.

But it's pretty nutty. How, I don't know. I, I like the idea of having artifacts inside projects two different names. It's hierarchical. No big deal. Yeah. And gems. Just gems. There's a gem and there's a gem and they can talk to each other. Um, 

Mike Crispin: yep. 

Nate McBride: Anyway, 

Mike Crispin: it's very powerful stuff. 

Nate McBride: Very powerful, but 

Mike Crispin: definitely, 

Nate McBride: yeah.

And then can call, you can call Opal to do other things, but opal's really just MCP for Google. I mean, it's calling over to Google Drive and Google Docs and it's not, it's just smoke and mirrors. Um. So Shadow ai, I mean, I think a lot of companies, I'm actually working with her on right now, I'm working with her CEO and she's aware they don't have a gen AI strategy.

Yeah. They're a smaller company. They're scrappy. They have an MSP, they have an [00:28:00] MSP, but she's patently aware that everyone in the company is using chat, GBT and Claude and Gemini or whatever the hell they want's, concern. And again, because it's, there's no way to block it. You can't stop it from happening.

Uh, that's a 

Mike Crispin: bad idea anyway. Yeah, 

Nate McBride: yeah. Yeah. So, um, if you weren't already head of this in governance perspective, I mean, you're kind of screwed there, but you know, there's the, everyone else's using it. Why can't I mm-hmm. Um, approach, like, why can't I use this? Um, but the, that's not even the beginning of the problem.

I don't give a shit if people are using generative AI agents. It's what they're putting into the agents. And then yes, uh, that's where the problems begin, I think. And so, but the shadow it part, shadow it itself I think is, correct me if I'm wrong, it's not really something that people care about anymore.

Like people, if they're gonna sign up for some dumb ass SaaS app [00:29:00] using their corporate account, do I care more about that than I do say them signing up for a public, uh, gen AI account, or do I care less? Because what can they do? I mean, the, the whole, the whole idea of cloud SaaS apps and shadow it, born out of 2014, 15, whenever made sense back then.

Oh my God, this person can sign up for a Smartsheet and put all kinds of terrible shit in Smartsheet. Okay. It is the same exact principle, but Smartsheet wouldn't, Smartsheet isn't ingesting that data, so That's right. They're 

Mike Crispin: not gonna, they're not potentially gonna reuse it. Yeah. 

Nate McBride: So fast forward to 2026.

This is the exact same equation. Uh, the LMS aren't ingesting data. That's right. So, I don't know. I mean, I think that the idea of an outright ban firewalling, it is ludicrous, as you said. 

Mike Crispin: Yep. [00:30:00] 

Nate McBride: I think a approved list is not gonna work. I, um, I mean, I think guardrails and training are probably the best ways to go if you train them.

Mike Crispin: Yes, I agree. 

Nate McBride: Train them to use a certain thing and then mitigate the risk by saying, okay, listen, here's the, the tools that we approve, use these, train people on those. Tell 'em to avoid the others because it's in the company's best interest to do that, and then move on from there. But I don't know, I guess the question is,

I'm going back to our earlier Gartner days when Shadow, it was a, a selling point, but I mean, honestly the two are really different. I think Shadow AI and Shadow IT are more or less exactly the same equation. Just a slightly different battlefield, slightly different battlefield. 

Mike Crispin: I, I think, I think the thing with Shadow AI is a little bit different because more people are [00:31:00] gonna put data into an AI system expecting some sort of result.

So they'll keep trying and trying and trying and adding more and more and more, and they, and it's all out of the kind of goal to be more productive. So everyone's trying it. Everyone's trying to use. And it looks like more of a productivity tool, but it is like a, I'm gonna use this with two or three other users, and then the product that comes out of it is also something that's discoverable and can easily be found.

So I think this, there's also like not a lot of protection for it to get around, like, to put in front of it to save yourself. People can use a camera, they can use multimodal AI to, to do anything they want really. So some of it is, I think the [00:32:00] governance piece is important and, you know, everything that I read, it's like, you have a great governance plan and dah, dah, dah, dah, dah.

Yeah. Okay. Got it. Great. 

Trance Bot: Yeah. 

Mike Crispin: Uh, we, you look, it's hard for small companies to have great governance. They're trying to do things really fast. They're trying to be compliant with so many different things. They wanna be more productive. They wanna do more with the same amount of people and they're gonna wanna use these tools because they're, everyone's telling them they'll be more productive.

So I don't, I don't have a good answer other than that, giving people and certainly making me rethink some of the things I'm doing, giving people more options is, is a good thing. But there's a lot more work to make sure those things are available and, you know, being able to pick a tool for the right thing, it brings us right back to where we are.

You could put something where, where we were with like having [00:33:00] duplicate systems and SaaS apps and, and enterprises, we were all trying to get away from that. Still are a lot of instances. Yeah. Is that now you put something in chat, GBT and it gives you some shitty results. You're like, ah, this sucks. I'm gonna use Claude now.

I try that just to see what happens. 

Trance Bot: Yeah. 

Mike Crispin: And they put it in the Claude and it looks good. The next day they go into Claude, they put it in, ah, this sucks. I wonder what Chachi BT would do. I'll put it in the chat. Bt, oh, we don't have Chacha. Bt I'll go see. Well, uh, you know, is this something I should put in a Chachi pt?

Um, I, I don't think I should, so I won't. There's many people who will do that. Then there are others who are like, well, I'll just cut the company name out of it and then I'll put it in and we'll, we'll see how much better. Oh, it actually did something pretty good. I want Chacha BT now. So these things every week they're changing, and I can't say it enough.

We're gonna be completely reactive to this, all of the [00:34:00] governance. I don't feel like I'm reactive at all else. 

Nate McBride: I mean, I don't think I'm special necessarily, but we started. God, uh, almost three years ago. 

Mike Crispin: So you are monitoring and see that if people are using other things, you're just 

watching 

Nate McBride: your No, no, but I, I pushed Claude very heavily back from the beginning 

Mike Crispin: Yep.

Nate McBride: And excuse me, and Gary on an account and told everyone to use it, and that's how we train people on Claude. 

Mike Crispin: Yep. 

Nate McBride: And so explaining of course, the virtues of the ethical balance and 

Mike Crispin: Sure. 

Nate McBride: Many, many other factors that led to the decision. And so people started adopting collude at that point? Yeah, this was like mid, mid 23.

Later. A little bit into 23. And, um, then when 24 rolled around, we still hadn't gone to Gemini yet, but I was starting to keep an eye on it. Gemini had come out 2.0, I think it was, uh, in early 24. And. I had a couple test users tested out against KA clad [00:35:00] at the same time PO came out. Um, and I was using PO to test the llama, uh, and to test, um, a straw at the time.

And I was not super psyched about the results. And so we stuck with Claude, but we kept pushing it harder and harder and it was like, here's an account for a thing that protects the company and lets you do this work. And Claude was actually, thankfully, thank God, giving decent outputs on prompts. 

Mike Crispin: Yep. 

Nate McBride: Also, I think this experiment would've gone sideways.

Now, every employee, when they start within their first week, they get trained on, um, quad, we call it gen awareness class. They get trained on Claude. So we have a hundred percent training coverage, uh, for at least the last year. Then they're enabled to take nine other classes That, and of course we, we, um, I launched Gemini in [00:36:00] early 25 as a second platform, but 

Mike Crispin: Yep.

Nate McBride: And they understand the difference. Like, when do you use, which one? That everyone gets trained on that fact. And I think that's been the difference for us is we don't have to worry so much. I mean, maybe an employee is still using something that we don't know, but by and 

Mike Crispin: large they know. I think that's the key.

The key is you're offering two solutions. And I think that's a good, that's a, that's a good piece of advice. It was kind of saying earlier, is you gotta give people and give your organization options and know what to use when, uh, you can't just try and pick one. I think you'll drive yourself crazy and multimodals a big thing.

I mean, you're gonna, unless you're developing, using cps. Taking risk with other data processors or third party vendors usually having, having Gemini gives you a multimodal solution if someone needs to make images or videos or Oh, 

Nate McBride: sure. 

Mike Crispin: Use Notebook lm for example. [00:37:00] These are things notebook you going to Claude's going down a lot too recently, which is been a massive shit in the last like month or so.

Nate McBride: Yeah. 

NLM 

Nate McBride: was, NLM was the original big driver for going to, because we won an explicit LLM functionality in the company for a low money. But, um, it's recently the, the language has changed. The communications have changed around which to use when, because obviously Claude has, um, its token conversion rates for files has been updated, but it's still a 200 K token limit.

Yep. And Gemini is a million plus. And it's a million plus, meaning that plus is random. Anywhere from 1,000,040, 40,000 to a million 0.1 0.3 million. I mean, it's in these huge ranges. Um, yep. There's a lot of factors that affect that. But if someone wants to upload, say 15 documents, the question they need to ask is, is it [00:38:00] not?

If it's clawed, 'cause it's Claude's out at that point, is it NLM or is it Gemini? Both can handle the same request. Um, and so again, we're training on this, which is, here's this, here's your scenario. Look at this grid. Which part of the grid does it fit into? Then do then use that platform. Don't get married to quad or married to Gemini.

Learn how to use which at the right time. Um, and I think that employees appreciate that. Do 

Mike Crispin: you feel it's okay? Is it, do you feel it's okay to just say, look, if it doesn't work well. Gini, then Tri and Claude, if it doesn't work well in Claude Tri Gemini, uh, 

Nate McBride: well, we 

Mike Crispin: don don't or you, or you don't like to do that?

Nate McBride: No, I don't want, I don't want people bouncing back and forth all the time with that, that fog of indecision over them. So there's a, a document that's shared in Slack. It's called Gen AI Power Tools, and it explains on a, on this one of the pages of the chart about when to use which one. [00:39:00] But ultimately, people in the very first class they take with us, come to understand what tokens mean.

We talk about token economy in that very first awareness class and how, how tokens are computed. I'm like, yes. Even spaces between words, like that's a token. So it's, it's all tokens. Um, 

Mike Crispin: yep. 

Nate McBride: People, people don't realize that in the beginning. And so. Uh, they come to learn that you can reference other chats and of course you can, um, be more concise, et cetera.

But also they learn if you're going to be using 10, 15, 20 context documents for your prompt, then you can use Claude. Long as you make a project, you can get a project for sure. But otherwise you have to go to Joni. 

Mike Crispin: Yep. 

Nate McBride: And you have to, or, or you have to go to NLM. And if you're gonna be sharing this to anybody, just go to NLM.

Just go right to the source of the, of the what you need and do that. [00:40:00] This has made for, I think if I didn't, if I hadn't done what I did, I would've a major shadow AI problem. But because we've given them so many options, 

Mike Crispin: that's the key. And 

Nate McBride: because, because the options are so powerful. They're not, 

Mike Crispin: and, and you're, and you're training them.

Right. And you're giving 

them 

Mike Crispin: all the resources they need. 

Nate McBride: Training up the butt. I mean, basically the goal is by the end of 26 to have every single employee trained through level three. We have nine levels, but level three gets you through artifacts, gyms, and projects. Um, if you get through level three, then everything else comes after that.

You sort of MCP, quad Code, et cetera. It's kind of very specialized, but, um, they don't get persona development until level six and personas. Are very tricky topic these days. Uh, personas versus self preferential, self preferential context versus, you know, [00:41:00] role-based personas type type language that comes later.

But the point is that they understand there's a learning track. They understand that for them to be successful in 27, 28, 29, when they leave the company or something else happens to the company, they need to have all these things on their resume. And this will be table stakes. They understand this. Yeah.

They're not, they're not trying to shadow it AI anymore because they're like, shit, I have to become really good at all this stuff and dabbling around in my little personal consumer account isn't going to drive the ball down the field. 

Mike Crispin: That's right. That's right. 

Nate McBride: So 

Mike Crispin: that's right. 

Nate McBride: But I mean, basically it, it's, it's the paradox, thinking of paradox is from season two.

Paradox I think is. We have to be fast enough, you and I as it, and you and I are, um, we have to be fast enough to approve tools. Well, not to approve them, [00:42:00] to assess tools. 

Mike Crispin: Yep. 

Nate McBride: And then to approve or not approve them in a way that keeps people aware of the fact that we're assessing and approving. I mean, I don't know how else to say it, but, um, if, if people understand that I am assessing and I'm testing and I'm approving or not approving, and they come to trust that because they have a bevy of tools to use through this process, I think they're more likely to go with what it says versus me saying, you can't use that.

You just can't use that for no particular, for no particular reason or, or justification. I had a big conversation with somebody the other day whose name I won't mention about copilot, and this person was trying to tell me that copilot was a generative AI strategy. And I'm like, no, no, no. [00:43:00] It's quasi agent, it's barely collaborative.

Um, and you have really no control over how it works. I said, imagine we took the same documents out of SharePoint, put them into Claude, and did the same analysis, the differences in outputs. So we went back and forth and we actually did a live side by side. And again, I won't say this person's name, but they were like, holy shit.

Compared to the, compared to the copilot response. I'm like, yeah, 

Mike Crispin: yep. 

Nate McBride: This is why, this is why we don't use box generative ai. That's right. Why would you do that when you can take the documents out of box and go to something else that's more powerful? That's not AG agentic, so to speak, and get, get, get control without loop.

And I think that is the power. People see that too, and they're like, no. Well, my CIO said if I take it out and go over here, I get better result. And he's shown it to me. He's demonstrated to me how shit, the result is if you rely on say, [00:44:00] box to summarize the document with the same prompt versus say Claude or Gemini to summarize the document, it's not even in the same fucking ballpark, not even the same sport.

Mike Crispin: Yep. 

Nate McBride: In terms of the response. And I'm like, yeah, well there's, there's a agent AI for you, and then over here is you controlling the narrative, you controlling with prompt training or few shot prompting how it's gonna go. So 

Mike Crispin: in, in your training, do you cover speed and set expectations around 

Nate McBride: Yes. 

Mike Crispin: Good. 

Nate McBride: Yeah, we talk about.

So we started off with son and Haiku and Opus, you know, the various differences between those. Yep. We talk about extended thinking and how to use research and you know, it's like those are basically go and get a coffee type, uh, prompts. 

Mike Crispin: Yep. 

Nate McBride: Um, then later on we get into, um, how that affects on the Gemini side with deep research.

And deep research is not only get a coffee, but get something, do eat, go to the cafe, [00:45:00] grab a lunch, get some soup, talk to your friends, then come back. Um, we talked about it, but those, that's all laid out there. And so even with sonnet four six, have, did you, I dunno if you've read the dev notes for, for sonnet four six, they have updated tremendously the way you are to interact with the agent.

Yeah. On the, on the, on the development side. And these are things that I haven't been teaching so much as I've been, um, edging around because I don't, I didn't have the language until just now to say it. But getting people out of the habits of Stop using please and thank you. And getting into the more direct, um, I mean those are, those are funny things.

I almost feel bad telling people, stop using please and thank you because I want them to still be humans. 

Mike Crispin: Yeah. 

Nate McBride: At the same time 

Mike Crispin: it takes, takes up, takes up space. 

Nate McBride: I'm like, [00:46:00] that's two tokens right there. You just wrote. Please. That's two tokens. Don't, don't do that again. Um, but anywho, so I, I had posted a, uh, invite to an event at, at Harvard for the fifth.

Um, I dunno if you saw that in the Slack board. You should come. It'd be fun. But they're going to, they're doing an hour on, um. You know, AI and biotech, which will be just bullshit. Google's there and just smoke and mirrors, I'm sure. And talking about GEMS and Gemini. Then they have an hour and 15 minute, uh, workshop with students from the Harvard Business School.

And so I did sign up for this, 'cause this will be an interesting part, but I'd be interested to hear how they're approaching the use of gems for agentic ai. Because if you [00:47:00] ask Claude how it translates, age agent ai and then you ask Gemini, Claude does it from a pretty objective perspective. Gemini, full Google ad.

Mike Crispin: Yep. 

Nate McBride: Full ad. 

Mike Crispin: Absolutely. Yeah, absolutely. 

Nate McBride: Um. I had somebody actually asking me today in my, my company, you know, what our concerns were about the, the American AI Action Plan. I said, we, our concerns haven't changed. Our model hasn't changed. If, um, Claude Dec decides to wa decides to whitewash his LLM of woke and DEI and stuff, we would drop them.

And he's like, well, how would we do that? I said, it was very simple. We've been saving our prompts in our prompt library. We have our templates, we have all of our data. We would switch to another vendor. Um, and I showed this individual how literally there's a new layer and the, the layer is that you can even [00:48:00] use your shadow it shadow ai or it if you want, but we have all the templates in our, in our library.

We can just pick it up and move to something else I said. So don't worry about whether, whether or not Claude Whitewashes, its LLM, we just moved to another event. And we, I run a test every single morning, uh, against both Claude and Gemini to see if they're still have their necessary data in their LLMs about DEI and wokeness and climate change and some other things.

And every day I get back the same response more or less, and I put it in the Airtable and I, I record it. But, um, I really, really hope that I don't have to get in a position where it says that we can't use a particular vendor, which leads me with one vendor because of this. Yep. Because I then I'm kind of in that potential for shadow people that have 

Mike Crispin: Yeah.

Nate McBride: Really started to [00:49:00] embrace Claw and may use it on the side, or, yep. 

Mike Crispin: Yeah, I think that's definitely, I, I do think there's gotta be an aggregate vendor at some point that comes out. And helps with some of this. There's actually, there are some already that, that are lesser, lesser known. Um, 

Nate McBride: yeah, I 

Mike Crispin: can't remember the name of a couple of them that have reached out.

But what they're doing is they're presenting the models to you and they put this whole apparatus on top of to manage it. Um, but I don't see those catching on. I think the more that they, the user experience suffers and isn't like what they're used to using on their phone or at home, more likely they are to, to, to kind of go back and look.

I, I think as long as you give the options, uh, like you, like you've been doing, I think that's the important piece is that it's available and you've done your due diligence from a contractual or security perspective or whatever you need [00:50:00] to do from, from a, a risk perspective and do the training. Make sure people are comfortable and understand one of the things that Claude is.

The amount of discipline you need to have is you can't just keep typing to the thing, even though they did the compression of conversations and stuff. Yeah. That's one of the things I get concerned, I get concerned about is, you know, just like, oh shit. You know, I was, I was just in the middle of this great thing and it, and it, it just dropped me.

And you know, to some extent they don't tell you what happened. 

Nate McBride: So there's, it's like, there's no like little meter on the side, like a video game, but you can, as of as of sonnet four five, you can ask Claude, you know, how many tokens have I consumed? But um, 

Mike Crispin: yep. 

Nate McBride: It's not a really well known thing with four six actually.

But 

Mike Crispin: you got note ask. Yeah, yeah, 

Nate McBride: yeah. And Gemini four six is 

Mike Crispin: great. 

Nate McBride: I love Gemini's answers when you ask Gemini Gemini's already always had this function, Gemini's evasive. It's like, yeah, you know, you're doing pretty good. You've probably used about this [00:51:00] many. I wouldn't say for sure, you know, but I. I, I just, again, I mentioned alternatives.

So we had tested Mara, or I did, I guess back in Yep. 2023. 

Mike Crispin: Yep. 

Nate McBride: And at the time, Mara had gotten into some hot water for, um, providing generative ai Deepak for deep fake 1.0 World. 

Yep. 

Nate McBride: And doing some other nefarious shit. But Mara was always kind of on the up and up on the backend. People had been just sort of maliciously using the tool.

Mara has come around, I mean, basically they, I think it's 3.0, 3.1. I haven't looked in a couple weeks. Mara three is, uh, most likely equivalent in trade over for Claude. If we dumped Claude, we would go to Mara most likely. Um, 

Mike Crispin: yep. 

Nate McBride: I have to go. I haven't done it [00:52:00] again. I, uh, head to head in about two months.

Excuse me, but there's also, there's, uh, Aerus, or Perus. 

Mike Crispin: Yep. 

Nate McBride: They're outta Switzerland. Um, a a which is out of Germany. 

Mike Crispin: Yep. 

Nate McBride: And Puro, I think it's outta Finland. And that's based on Luma Open. So those are, those are all good alternatives. So with SONET four six, the

SONET four six incidentally brought to light something that I had been talking about recently with some folks that, um, the training data cutoffs for Claude. So someone had asked me in one of my classes about when was the last time Claude was refreshed, and it had been, it would've been [00:53:00] November. So Opus.

Still has a training date cutoff of August 25 with a last, last reliable knowledge cutoff of May 25. So the, um, LRKC is may of last year for Opus Sonic. Sonet just updated. So as of Jan 26, um, is the last reliable knowledge cutoff, but they're still saying sonnet. Um, has a, uh, last best data date of August 25.

So the, it's been, the training date for sonnet is January 26th, but the last reliable knowledge capture is 20 August 25. And so I sent out a note date to the company saying, listen, if you're, if you're using Claude for data that occurred after August 25, [00:54:00] you have to treat it with the highest level of skepticism.

Um, yep. You know, it's updated through January. You still have to go to, to Gemini if you want something that's bonafide after August 25. And Haiku, by the way, just for, um, the sake of clearance up, haiku's, last training date is July of 25 with an LRKC of February 25. So a year, a year old. It's a year 

Mike Crispin: old. 

Nate McBride: I, I, like I said last, last week, I don't think haiku is coming back.

Um, but maybe it will. 

Mike Crispin: So you, but you've gotta, one, one of the things that we do is we have a project that's like, like fake, fake perplexity almost. And it's like use in the, in the, um. In the project note, it's like, use the web for all UpToDate information, dah, dah, dah da, the all sites colon, dah, dah, dah, dah.

Nate McBride: Yeah, yeah, yeah. 

Mike Crispin: I just say, go, if you need the UpToDate research, use this project and go in and, [00:55:00] and it goes out to the web by default. And it's like, oh, okay. That's great. That works well. But it's still, it's, it's still, uh, just somewhat takes time and, you know, little, little longer if you're just going against the web.

Haiku does that very well, but it's very, it's 'cause that's quick. It's pretty quick. But I'm curious as well about Haiku and what they're actually going to do with it. And if they can bring that, uh, a bit up to speed with where Sonet is at least, 

Nate McBride: why wouldn't, why would they? 'cause Sonet Sonnet has non extended thinking mode, which is basically haiku.

I mean, it has 

Mike Crispin: Yep. 

Nate McBride: Fa it has ge, it's equivalent of Gemini's Fast mode. 

Mike Crispin: Yep, yep, yep. 

Nate McBride: Do you 

Mike Crispin: think they'll drop Haiku altogether? 

Nate McBride: I think they'll drop both. I think sonnet will get a, um, research mode, AKA Opus and Haiku will be folded in. But I think Opus is sidebar, as you mentioned last week is expensive and it costs money and it's good revenue generator.

But they can do the same thing in [00:56:00] sonnet if they simply have a sonnet. And then sonnet plus mode, like I don't have to call it Opus. I dunno, I'm not a branding guy, so I don't know. 

Mike Crispin: OO Opus has, uh, been pretty unreliable the last couple weeks at doing major shit. So it's like, yeah, it costs a fortune and if it fails, like, oh boy.

But it's like, uh, because you know, some of the guidance I got was, yeah, you shouldn't be, obviously you shouldn't be building a lot of things with Opus. You use that, you know, for bigger problems, bigger thinking issues, and then if you want to really dress something up, you really want to, you know. 

Nate McBride: Yeah.

Mike Crispin: Bring a lot of data sources together. Um, but it will just stop in the middle. I mean, it's like a one question, one question in two documents, and it's like it's outta context. 

Nate McBride: Yeah. 

Mike Crispin: And I just can't help but think, you know,

the [00:57:00] money in four days, it's like, man, and I'm not doing anything that's that crazy. 

Nate McBride: So the guidance we give is, if it's gonna be big analysis, go to Gemini. Yeah. Like you sonnet for 90% of your work, your day-to-day stuff, summaries, analysis, et cetera. But for very big lifts, go to NLM or Gemini to not do this inside of sonnet or, or opus.

Yep. We don't even, 

Mike Crispin: yep. 

Nate McBride: I mean, Opus is like a, a, a little asterisk on my, my training. Yes. It's there. So why, 

Mike Crispin: why keep Claude around at all? Just because it's infused in the organization and it's another option, like Right. That people might wanna use? 

Nate McBride: Uh, yeah. Well, it's because of, I mean, for, for what we do think about a lot of what's generated, which are slides, presentations, and documents, you, 

Mike Crispin: you need Microsoft.

Nate McBride: Yeah. And so, um, [00:58:00] if someone needs to create slide analysis and into a summary, Gemini will do. Okay. Gemini, compared to a Claude, they're not even close. Claude gives you a much, much better output. Saves as an artifact, put that artifact into a project. Um, so we, we, we basically, the, the, the vernacular is that, uh, Claude is for making, Gemini is for asking.

So if you want an expert, you make a gem. If you want to make a document, you use an artifact and yeah. You we're trying to get everybody outta the prompt game in general. Like we're trying to get everybody out, out of just using prompts as they're themselves, moving everyone into prompts for their team.

This is kinda like one of the big goals for 26 is getting out of the, uh, hey, it's, it's me in the, in Claude versus it's, it's the team in Claude or the team in [00:59:00] Gemini. Yep. And I, I feel pretty good. It's only February, but we're already making really good head roads, uh, headwinds. There not headwinds inroads, uh, towards 

Mike Crispin: inroads 

Nate McBride: towards that goal.

Inroads. Yes. 

Mike Crispin: I like it. 

Nate McBride: The road inside. We're making good progress there, but, um, no, I still, I, I don't think the two hold a candle to each other when it comes to Gemini doing an assessment or a summary versus, uh, Claude, Claude is way, way better or responding to. Pushback frameworks. Few shot prompting.

M-P-A-M-T-P. Chain of thought. Tree of thought. 

Mike Crispin: Yep. 

Nate McBride: Gemini. Like, that's 

Mike Crispin: amazing. 

Nate McBride: Why? Why the fuck would I use t cray? Here's your output, asshole. Go away. That's how I feel like that's why I feel like Gemini is like, 

Mike Crispin: what's wrong with this? Here it is. Yeah. It's great. This is great. 

Nate McBride: Yeah. Here it is. It's just fucking, what else do you want?

[01:00:00] Go away. Take this. Stick it. 

Mike Crispin: I love when it does, when Gemini does something wrong and it, it just apologizes and it's like, oh, I'm so sorry I did this and that. 

Nate McBride: Yeah. No, I mean, 

Mike Crispin: I I'll try better next time and they don't messes up the next time. And you call it, it. I don't, I'm not much, I'm really upset with myself that I couldn't provide you the, the artifact that you're asking for.

Let me go back and take another look. It's so nice. 

Nate McBride: Yeah. No, no, no. Sorry. I mean, nothing beats Claude's. Claude in real life, I think would have a. Like a, a whip and would be, would be slashing. Its back open every time it had based on, its apologies. You are so right. Oh my God, I was so wrong on that slide. I just made that up there.

There's nothing at all to support this. Please, let me try this again. And I'm like, well, you're a fucking idiot, but okay, go ahead and do it. You're right. I'm a [01:01:00] fucking idiot. I'll go do this again. I'm so sorry. Claude is awesome. His apologies. If I, if I, if I was actually working for philanthropic and I was a developer, I would come up with the most creative apologies ever.

You're so right. I'm going to go kill myself. 

Mike Crispin: Oh, 

Nate McBride: I'm killing myself right now. I'm cutting myself open 'cause I'm so disappointed in myself. 

Mike Crispin: Oh boy. 

Nate McBride: I'm pouring acid down my throat right now because I messed up your prompt. I'm so sorry.

Mike Crispin: Pouring acid. I did know I'd hear that tonight, but Pouring acid. 

Nate McBride: Pouring acid right up, right up there with your breast burgers, 

Mike Crispin: breast cafe.

Nate McBride: That's where we are right now. By the way. Look at our Zoom background. 

Mike Crispin: I'm loving [01:02:00] that. I'm loving that Zoom background. So 

Nate McBride: I feel like the entire idea of shadow AI was about a 15 minute conversation. Um, yes, it exists. It exists. If you have shitty governance and you have no idea what's going on and you didn't take the time to roll out an educational process for Gen AI in your company, you probably have shadow ai.

And if you do, it's okay. You can just quit and go to another company and start over again. Or you can wrangle 'em all together and train everybody on how to use channel ai. Um, and start over. You can call me. I'll come and train you. Uh, call Mike. He'll give you the um, 

Mike Crispin: yeah, script. 

Nate McBride: We'll, we're happy to help you out.

But otherwise, just assume if you don't have a program, governance policies and a way to legitimize usage, 

Mike Crispin: I think importantly it starts at the top. You need to make sure that 

Nate McBride: yes, 

Mike Crispin: leadership understands how important it's to train and [01:03:00] to give some form of governance and put some, some rules around it.

And to have also have a cross-functional team that doesn't have to be anything super formal, but just a group that talks every couple weeks about this. 'cause it's changing every day. So it shouldn't be like this. Oh, we meet once a quarter and we talk about AI governance. Governance. It should be, it should be a more sprint, uh, more, more small setup.

Didn't we, 

Nate McBride: didn't we just meet, didn't we just meet to talk about this? Why are we talking about it again? 

Mike Crispin: Yeah. We 

Nate McBride: already approved this. 

Mike Crispin: It's like, uh, well we, we have to update our acceptable use again. Um, but yeah, it's, it's because you need, you're gonna need that support to be able to train people and take the time to, even, even if people really wanna be trained, which a lot of them do, like they had said, it's incredibly important from a career perspective to learn about how to use these tools.

Nate McBride: Yeah. [01:04:00] 

Mike Crispin: It's just, if they not have time to put aside, to do the trainings, to be a part of it, you, that's another challenge you need to can overcome because the shadow AI stuff is very easy. To, to, to use and to find even on your phone. I think that's where a lot of it gets used right now, you know, is, Hey, I need to get this done by three o'clock today and it will let me use ai.

So I know how to use, you know, I know how to create a document with Shachi bt, and I can get this going and, uh, let me upload some stuff. I'm not saying that everyone's doing that, but I think without the governance, Nate has mentioned you've got a very good chance it, it probably already happening in a lot of companies where I'll just use it for this one little thing or I don't like the tool that, that it is making me use.

So I'm gonna use something else and, um, [01:05:00] I'll ask for forgiveness. 

Nate McBride: So just one quick, one quick bump off the knuckle. Uh, I really just one quick, one quick bump. I'll 

Mike Crispin: bump up the knuckle.

I love that. I love I love that. Thank you. Um, but I think it's light, lightweight guardrails. Not like these drastic, you'll use copilot. Copilot is the, it's our, it's our system. That's our enterprise platform. I think that's a good, that's fine if you wanna start there. 

Trance Bot: Yeah. 

Mike Crispin: But you've gotta have the ability to branch out and add other services.

And I think sometimes as a team of one at a company, that's the easiest thing to do, is to say, yeah, I've got copilot. Uh, it's doing some of these things, but I encourage people to take the time to look at the other models and come up with a plan to proof a concept or pilot them in the appropriate way [01:06:00] with the right teams.

One of the things I just did, very simple thing was put a form out and said, you know, what do you wanna use AI for? Just, just, just tell me what you wanna use it for. What are you, what are you curious about? And we got a lot of responses and it really helped us to mold how we're gonna move forward with our, you know, our next steps and how we're gonna pilot.

We did proof of concepts and now it's time to do some, some pilots, um, and put the appropriate sort of tools in place. 'cause they're obviously being used, so we need to have something to, as a standard. Yeah. Or as a, as a, um, as a good process here, here. And you gotta know your culture too. 

Nate McBride: Ising 

Mike Crispin: question your company.

Ready. 

Nate McBride: Well, I agree with all those things, but the question is why wouldn't you do for any of the technology? So this is one thing I I've you do a survey for every technology. You roll out the company like this. 

Mike Crispin: No, I, no, I You do, I do it every year when we're doing [01:07:00] budgeting. You ask where there are 

Nate McBride: Oh yeah, yeah, for 

Mike Crispin: sure.

Needs and gaps and asks. And that creates a lot of the similar type of requests that will come in. Is, uh, we want this new thing, we want a new capability. Uh, but no, we don't, we don't do it for, I do think this is a little bit, uh, and I use this all the time, this is like hiring a contractor. This is not like putting a system in.

Yeah. So just saying, we're just gonna use this one contractor for everything is not gonna work, uh, at all, in my opinion. So you wanna have sort of diversity of thought, diversity of systems, and that's diversity of ais and how you pull it all together. I don't think anyone's really figured that out yet, because you can offer one or two or three, but there's a new one every day.

It's hard to keep up. And there's gonna be different cost models, there's gonna be different security risks, and you are gonna have to [01:08:00] pick one eventually. That's incredibly hard. Like you're gonna have to pick one to, to pilot and make sure you're gonna spend the time training on. If you have your 

Nate McBride: governance in place and your templates in place and your library management in place, then picking one isn't actually as hard as it used to be.

Mike Crispin: Yeah. 

Nate McBride: So long. So long as you're not operating exclusively out of a dependency. If you're able to move what you're, your work to something else, it's no problem if you can't move 'cause you have 3000 GPTs in your company. 

Mike Crispin: Yeah. 

Nate McBride: Then good fucking luck. Jesus Christ. I won't mention any companies that are vaccine based for COVID, but 3000 gpt, what the fuck were you 

Mike Crispin: thinking?

That's a very, that's, but that's, that's a whole other level of shadow AI is what you have. The AI tools. Okay. Now I've built, I've got cloud desktop and I've built CPS that are [01:09:00] specific for my team and I've got all this investment and things that are user created. Which in some respects you want. That's one of the funny things, things we were talking about last week, was when you get to a point where it's like, here's a prompt AI team, go get this thing and bring it back to me because, you know, I, I hope that's not what we end up, but it's, um, it's, yeah.

But all operating outside of that see upgrade personas and different things that are gonna stick with the system. 

Nate McBride: Yeah. All operating outside of SD LCI mean, no one's got any control. 'cause how do you Exactly, how do you do SD LC with a, with a jam? You or GBT artifact. Right. But you should 

Mike Crispin: No, I, I just think it's too slow.

I think that people are gonna go blast right past that. 

Nate McBride: Yeah. 

Mike Crispin: And they, 

Trance Bot: I know 

Mike Crispin: and it's so easy to do. Like take the GEMS and Gemini. I mean, we could have a thousand of those. They're so easy to make. 

Nate McBride: Yeah. But 

Mike Crispin: they work, they work great. But [01:10:00] 

Nate McBride: you can make, by the way, not to get off on on. Too big of a tangent, but if I make a gem with one line and I, I have the instructions somewhere, you can send that gem and its data to a Google sheet that already exists.

Yes. That tracks gems. So I can, if I was full Google workspace and God fucking hell, if I can get there next year, then I would blow this wide open. Because imagine having the ability to use Opal to send every single thing that happens in Gemini to a Google sheet, uh, which is then mirrored to an Airtable.

Like, my work is done like as a CIO, effectively have no purpose after that. But, um, that would be awesome to immediately send all generated things to a spreadsheet slash table. You can't do that in call. I don't see you enable MCP and I'm not doing that in my lifetime, I don't think. At 

Mike Crispin: least [01:11:00] not. That's a, that's scary.

That's a scary thing to do. Right. Is, uh, is just your desktop side of the world. That's gonna be, uh, it's gonna be scary. 

Nate McBride: No effing way, no effing way. So I'm my own shadow ai, which is to say that as a CIO I'm opposed to my own policies, but I also respect them, uh, obviously immensely. But I'm also creating domains in Google workspace areas so I can go break shit, which I reserve the right to do.

Um, but standing up a Google Workspace domain, it costs almost nothing. And then be able to use that to test both Claude and Gemini in a Wild West environment is a spectacularly wonderful feature. And this is, I think the best part about Google Workspace is that for. The cost of a domain, [01:12:00] 12 bucks, the cost of a year of Google Workspace.

13 bucks a month. 

Mike Crispin: Yep. 

Nate McBride: You can stand up, you can stand up as many workspaces as you want. And I mean, these are full Google workspaces in the business plan with full Gemini, full opal, full vertex, and just go to town break. Shit like crazy. Nowhere else can I do that. Can even do it. AWS 

Mike Crispin: one one other, that's, that's the component I was gonna highlight there is when you set up that Google workspace, you, your Google Cloud environment is available and Gemini Enterprise is available for compliant related items.

So you, you actually own the LLM? Yes. In the, in the query in the cloud. So as Nate was saying, you'd have to have a AWS and bedrock or you'd have to have Azure Yep. Set up on the side. Uh, and built, run everything through that. And you're getting that at [01:13:00] cost with, um, with the workspace. So a very bullish and Google workspace.

Uh, Nate, I know you are as well. I think if you're a company that's small that's starting today, it's why, Hey, I'll bring up the $740 MacBook again. But is, hey, there's a lot of companies that are starting out with Max, they're starting out with Google Workspace. They're mostly on the West Coast, I think, but the total cost of ownership is so low, and the people come in the organization, that's what they want.

That's what they're hiring. So it's very interesting to see how that's, it's gonna change. It's gonna change quickly. 

Nate McBride: I went to last, last year, last week for one day. I think people go for longer than that, but one day and what I loved and I lost count, but how many vendors? Big time vendors now, mind you.

Mike Crispin: Yep. 

Nate McBride: In the, uh, biotech lab automation process, AI space, or running on workspace. Almost every vendor had Google Workspace up and running, [01:14:00] and their company's fully running on, um, Google Workspace. So 

Mike Crispin: yeah, 

Nate McBride: we're just all dodging, dodging around the truth. 

Mike Crispin: Yeah, I think, uh, it's, it's, it's comfort and change management that's scary for people.

And I know you and I have both done it at other companies, and when you end up going to the, the next thing, there's a lot of people there, there are people who are happy, you know, to go back to the Microsoft world. But at the same time, I think there's a lot of people, usually those of people that had no idea or never learned the tools or didn't go to the training.

Um, I had a number of people that were like, God, I missed, I missed Gmail. I, I hated you when you put it in, but I miss it. Um, but. So here's one other thing I want to say about Gemini and shadow AI and data security. Is that the nice thing [01:15:00] to bring that back in a minute? The nice thing is that Gemini by default, because it is so, like you said, a big ad for Google is that it truly segments your data.

So if you're a Microsoft 365 world, you can open the door to AI and keep everything segmented. Everything that gets created or set up by Google is in this other thing over here. Yeah. And Google becomes the AI ecosystem. Yeah. And then all this other stuff that needs to be created by humans lives in the Microsoft world.

Right? So you're able to really segment it and say, we don't, we don't want to, we don't want things talking to each other yet. We're not ready for that. You can send up Gemini as a tobo, totally separate environment and, and really ha feel a little bit safer about. Fact that someone's not gonna run a cloud search across your box environment and pull up stuff that we didn't know [01:16:00] they had access to.

Um, so there's, oh my 

Nate McBride: God, how did that person get that? I, it's a mystery to me. 

Mike Crispin: So it's, it's a good place to start. I think it's a, an interesting way of looking at it, but your point around being Microsoft heavy, needing Microsoft artifacts is like the number one use case. I hear, Hey, I, 

Nate McBride: yeah, 

Mike Crispin: I need to work with Excel, I need to work with PowerPoint.

Um, so it really starts with planting that seed in getting outta that ecosystem. And that's just such a challenge. Um, 

Nate McBride: well we have a, we have a secondary goal this year, so we have four major pillars for it in 2026 as part of our new three year plan. 'cause it's 26, 29, 28. 

Mike Crispin: Yep. 

Nate McBride: 29. But. One of the other pillars is to get everyone off of Excel, to eliminate Excel from the company and replace.

Mike Crispin: Yep. 

Nate McBride: I'm not kidding. 

Mike Crispin: I, I believe you, 

Nate McBride: Mike Rap has, Mike rap carries the burden of this goal, and so far we are getting [01:17:00] finance showing up at our door. Like, let's get rid of this, let's get rid of that. And I'm like, fuck yeah. All this stuff. Let's go, let's get it outta here. Um, I think that once we can unbreak some of these burdens, especially the Excel burden.

Mike Crispin: Yep. 

Nate McBride: You open a door to workspace. Yep. A new door to workspace. There's only really two, I think two big doors to workspace for a biotech company. And there, there are some smaller ones, but two big ones. One is Outlook. 

Mike Crispin: Yep. 

Nate McBride: Fucking Outlook. And two is Excel. 

Mike Crispin: Yep. 

Nate McBride: If you can cross those two Rubicons, you can get to workspace Very easy.

There's, there's very two big moves. 

Mike Crispin: What about Word and like starting point templates and regulatory Yeah. And all that stuff. 

Nate McBride: So, so you move, so every single one of those, you move to a cloud environment. And so having now, so we have a couple SaaS [01:18:00] contractors, all of them using, uh, a Windows 365 workspace to do their work in which they have a dedicated, um, office license for that, that 365 win 365 instance.

So we're paying, yes, a little bit of a premium. It's about $1,500 for every two years about the cost of a laptop. But, um, 

Mike Crispin: yep, 

Nate McBride: they, they're able to log into a WIN 365 environment, use office natively, and then their work is done. So for those users that absolutely required word for, say, regulatory or uh, formatting, uh.

Whatever. We can create those environments for 'em, we can spin those up very easily. So yes, here's where you go to work in Word. Otherwise, most of your shit would happen in workspace and, and of course you're gonna export to Doc X. But Mike, this goes back to the original question from season two and earlier this season about [01:19:00] who needs to create.

Yeah. And if we can just simply identify who our creators are, like, Hey, here's our 15 creators. We can actively give those people the ability to create multiple environments. Everybody else does not get to create, or they don't get to create an office. It's not a complicated model. It's so fucking simple on paper.

It's even simpler in real life, but you gotta kind of, you gotta get into it by eliminating some of the hurdles. I think the two biggest hurdles are Outlook and Excel. They're way bigger, I think, than Word. I understand you. Why? I don't understand what you're saying about Word like. 

Mike Crispin: Sure. No, no. I, 

Nate McBride: yeah, A four A four templates and please review and shit like that.

All surmountable. 

Mike Crispin: Yep. What much You're working with the EU and they're still using.docs, not Doc Xs, but docs. 

Nate McBride: Yeah. We, we know this because every single time, and we, we've trained our eve [01:20:00] bot, uh, to recognize this when someone's like, yeah, I'm not, it's not showing the dropdown open an office menu. Eve's first response is, is it a dot doc file or dot doc x file?

Um, if it's dot doc, uh, 1998 called, they want you to come back and fix that fucking problem

dot doc files. My God, 

Mike Crispin: why won't this open it in box? Why won't this open it on my phone? 

Nate McBride: Because it's using literally fucking ancient technology, 30-year-old technology. 

Mike Crispin: Yep. 2000. Yeah. Actually nineties. Late nineties, right? 

Nate McBride: Nineties 98. 

Mike Crispin: Yeah. My gosh. 

Nate McBride: Yeah. X uh, they're like, Hey, are we gonna do a new XML standard for office and we're gonna add extra letters to all of your extensions.

Good luck.[01:21:00] 

Yep. Good luck. I remember being at, uh, Cubist. Yep. And we went from, we went from PPT files and POT files to PPTX and Pot X. And I'm like, well, what the fuck? How do you explain a POT X file to somebody? I don't know. We made it, Mike. We made it through the gauntlet, man. 

Mike Crispin: Yep. 

Nate McBride: All right. Well, I did wanna mention, by the way.

I dunno if you saw the 4 0 4 article on Ring, the, this has now become a bigger thing. 

Mike Crispin: Yep. 

Nate McBride: So in October, ring launched search party, did you read about this? 

Mike Crispin: I did not read about this. I heard a little bit about, 

Nate McBride: yeah, it's an on by default feature. You have to opt out that, um, was [01:22:00] originally coded to find lost dogs.

Mike Crispin: Oh yeah, yeah, yeah. No, I I saw that that was at the Super Bowl. They advertised that. 

Nate McBride: Yeah, yeah, yeah. But, but then, but, but the, but the owner at the time said that it would be expanded eventually to zero out crime in neighborhoods. 

Mike Crispin: Yes, I did see that. I did 

Nate McBride: see that. Um, and this guy's got his own manifesto about, that's his whole life goal, eliminating crime in neighborhoods.

And they sent out a whole bunch of emails about this around the ring company. You should be in, if you're a Ring customer or a Nest customer or whatever the fuck customer you are for. Cameras understand what that company's doing. Ring is doing some nefarious shit. They're not just finding lost dogs. It's a cover.

Please just do your homework. Shut all the shit off. Um, 

Mike Crispin: yeah, [01:23:00] go switch to HomeKit. You're gonna see, you're gonna see the, uh, that's gonna, that's gonna emerge this year as a, that's how they're gonna sell it. 

Nate McBride: Or you just don't need cameras. We did, we did fine with cameras for hundreds of ears. Yep. Um, and we did All right.

You'd be fine without a camera. Uh, okay. Well, I don't have anything else to add to this topic, Mike. I mean, I think we, uh, 

Mike Crispin: no. 

Nate McBride: Nailed that, like a split hog. 

Mike Crispin: Absolutely. So I, I think it's important to have that light framework. Make sure you have the leadership support to, to be able to engage people on the training.

And 

Nate McBride: yeah, 

Mike Crispin: take the time to make sure you, you've got everything prepared before you go into that so you're not switching around and I think have multiple options, what to use, what, when, and if you can find a, a [01:24:00] really good front door that people can use to access these things so they have one place to go.

I think that's also helpful for people who may not understand that they can bounce around. So that could be Okta, that could be a page you're creating somewhere, but a way that you can kind of bring all your AI rules and policies and know-hows and maybe even a place where people can show what they've done with the tools, um, can help good use cases that give other people ideas.

A little community around it, which I think is always a good thing. 

Nate McBride: And if you don't have executive control or backing, um, just say you do. I mean, honestly, just fake it till you make it. And if, I mean, you didn't hear from us like, we didn't endorse this. We're simply saying that if you don't have executive approval, just say you do.

And then until someone catches up to you and then [01:25:00] blame it on somebody else that you thought you did and refer to it. 

Mike Crispin: I could have swore I had that budget in here. 

Nate McBride: Yeah. Refer to a, a manufactured email that you created using gen AI to indicate where you thought you had approval. And then when they say, oh, I never sent that to you, manufacture it into their scent mail so that it shows up there.

So they did send it to you and blame their crack at addiction. To, um, why you got approval? I mean, I don't have, I, I feel like I gotta explain everything. I don't understand why, like, this is all sort of obvious to me. 

Mike Crispin: Yeah. You, you already got it. Say it was on 60 Minutes. We're all set. 

Nate McBride: You could do something Black Mirror.

You saw it on Black Mirror? 

Mike Crispin: Yes. You, yeah. Black Mirror will that'll go over. Great. 

Nate McBride: Yeah. Um, well next time we meet Mike, we're talking [01:26:00] about CWP 

Mike Crispin: Ooh Cakewalk Pro. Great sound application from the late nineties. I think they still make it. I think it's in a GW. 

Nate McBride: Yeah. Mike and I are going back in time to revisit Cakewalk Pro to, um, 

Mike Crispin: can't wait.

Nate McBride: Discuss audio on your computer if you Windows 98, uh, this is gonna work, work out great for you.

No. Oh, 

Mike Crispin: beautiful. 

Nate McBride: In fact, we're gonna talk about, uh, a build first culture, which I'm a hundred percent behind 'cause it also brings shadow AI and shadow it on board with you. If you build first buy last, we're talking about, uh, low code, no code development and how fun that is. I love it. And creating applications and workflows that basically [01:27:00] mimic what you're already doing and paying big money for.

Mike Crispin: So, so it, so configuration management on steroids, sort of 

Nate McBride: con configurating the configuration. Yes, 

Mike Crispin: yes, yes, yes. 

Nate McBride: And then the, the, then the week after that, we're gonna talk about the other part of, um, CWP, which is where, um, agent AI and MCP sort of fit in. To this new idea, uh, and risks and governance.

That'll be a Scorcher. Uh, I might want to, um, get a couple extra Lawsons sip of Sunshines next to you when you do that episode, but next time we meet, yes, we'll be talking about, uh, build first culture and why they're so fucking awesome. And then LCNC creating applications, workflows, and just doing cool shit because you [01:28:00] can, you don't need a whole lot of it approval.

Just go around it. Seriously. Stop asking them for anything. Just do cool shit. 

Mike Crispin: Oh, come on, man. 

Nate McBride: You trouble. We don't, we don't, we don't have the time to deal with it anyway. We'll just, we'll clean it up at the end. You'll, you're, you're, you're keeping us employed. Just do shit. We'll, we'll, we'll figure it out, right Mike?

Mike Crispin: We just figure it out. We'll be all totally reactive. 

Nate McBride: Yeah, we'll just figure it out. We'll be reactive. Well, I mean, that gives us at least 10 more years of employment. So if you keep doing shit and breaking shit, we love that. We won't tell you we love it. We'll be like all angry, like, oh my God, why do I keep doing this?

But secretly inside we'll be like, thank you for doing this right, Mike? 

Mike Crispin: Yeah, yeah, yeah. We'll be, we'll be very excited. Did we just talk about [01:29:00] Shadow IT and Shadow ai? 

Nate McBride: For the record, Mike, Mike endorses this. I've, I've just cut it out of the podcast 'cause he sweared a lot, but Mike, a hundred percent loves this idea and whatever else happens, I've, I've cut it out of the episode.

So Mike loves it. 

Mike Crispin: Just dropping, dropping a ton of F bombs and. Yeah. Oh yeah. 

Nate McBride: Mike said the F word 40 times in a row. I had to cut that out. But he loves this idea of giving everyone the, the tools to do everyth. They need to do 

Mike Crispin: Fiji, Fiji, Fiji.

Nate McBride: Um, all right, so on that note, welcome to the podcast called Calculus of It,

season three, episode 11. I'm Nate McBride. With me as my, as always, my cosst, my [01:30:00] Crispin. 

Mike Crispin: Have a great night.

Nate McBride: Tonight we're talking about shadow ai. Mike, how you doing? 

Mike Crispin: I'm doing great. How are you, Nate? 

Nate McBride: I just shoved out a burger. 

Mike Crispin: Next slide. Can you go to the next slide?

Nate McBride: Man, I really, Hey, tonight's episode, by the way, Mike is brought to you by Super Sunshine, by La Bay, Lawsons. 

Mike Crispin: Oh, you did say that earlier? 

Nate McBride: No, that was all, that whole last hour and a half was practice. 

Mike Crispin: Okay, gotcha. This is the real show though. Okay. 

Nate McBride: We were, we were like, we were like, fine tuning our script.

Mike Crispin: Cakewalk Pro. Love it. Love it. 

Nate McBride: Um, oh no, that concludes episode 11. I was just joking of, oh, and, uh, have your pet spader neutered. [01:31:00] Do not create more pets. Be nice to old people. Be nice to it. It, by the way, knows all the weird shit you put into Gen ai. 

Mike Crispin: Yeah, 

Nate McBride: we're watching you. We're laughing at, like you asking about, well, I won't say it on the line 'cause you'll know who you are, but we know what you're writing into the prompts.

So you know, there's your phone for personal stuff and then there's work. So take your pic. Um, 

Mike Crispin: join 

Nate McBride: our 

Mike Crispin: Hey. What Seduce, once Seduce Siri comes out, we're all doomed. 

Nate McBride: Yes. Well, we're all doomed anyway. I mean, basically if it wasn't Mike, it was me. And we were both tracking you. We know everything you're doing through flock, through ring, we've got a pattern of behavior.

It's all documented. 

Mike Crispin: Tons of air tags everywhere. 

Nate McBride: All your air tags. We know what [01:32:00] you've been doing. Stop that. Well keep doing the good thing, but stop the press of us. 

Mike Crispin: Why isn't 4 0 4? Why isn't 4 0 4 media done a a, an article on, uh. 

Nate McBride: My teeth, 

Mike Crispin: no air tags. Those things are, oh wait, no, they have, they have, they have done a story on 

Nate McBride: it.

Oh. Well I always thought you were, I thought you were wondering why they haven't done an episode on our podcast yet. And how groundbreaking are 

Mike Crispin: we are breaking new ground 

Nate McBride: and how many, how many social media exploits we've covered. 

Mike Crispin: Mm. 

Nate McBride: Like remember our covers on the tech tech accords? We covered the tech accords like in great detail.

Yeah. And 

Mike Crispin: it's amazing. They're getting so much coverage and everyone knows about them. They're hugely successful. 

Nate McBride: We talked about the, um, the, the coin, the, the ball coin. 

Mike Crispin: Yep. [01:33:00] World's coin. 

Nate McBride: World coin. They call it ball coin. 

Mike Crispin: Bullet Coin 

Nate McBride: Balls. Coin. 

Mike Crispin: Ah, nice. 

Nate McBride: So be good to everybody. Be nice. 

Mike Crispin: Test a coin.

Nate McBride: Invest and test a coin. 

Mike Crispin: Yep. 

Nate McBride: It it goes. Had to have a stable coin. Honestly. It's a new currency. 

Mike Crispin: We gotta create that. Nate, we gotta create test a coin. What do you think? 

Nate McBride: Yeah, man. Let's do it. 

Mike Crispin: It's just a test. It's just like a test environment kind of. 

Nate McBride: Yes. Uh, it's just a test. 

Mike Crispin: We'll call it Test a coin.

Nate McBride: Yes. Invest in test a coin. Uh, there. Your 

Mike Crispin: t That's the next t-shirt right there. 

Nate McBride: Yes. It's also gonna be a t-shirt on our, hey, on our [01:34:00] merchandise shop, which you can get access to. That's at the bottom of every podcast episode is the link to, um, our merchandise shop, our buy of beer link. Love it. And this is literally my last Lawsons.

They only sent me one case, and I'm on my last one. Oh, 

Mike Crispin: you blasted through it this hour and a half? 

Nate McBride: Well, just in the last four days, but 

Mike Crispin: Oh, okay. Okay. 

Nate McBride: So we need, we need more beer. And you can also give us all the stars. Um, the fact that we're not getting enough stars is problematic. And maybe 

Mike Crispin: how many stars?

How many stars do we get this week? We getting any likes? Oh, shit. 

Nate McBride: But I'm not sure, you know, honestly, Mike, I'm not sure people had to know how to act up, like do the button. I mean, look at you, look at the phone and the, the podcast apps. It's not a ho totally clear how to click the like button or the [01:35:00] all the stars.

Oh, 

Mike Crispin: so 

Nate McBride: there's a learning 

Mike Crispin: curve, huh? That's a big learning curve. 

Nate McBride: Yeah. So if you're having trouble giving us all the stars. A, you can reach out to us, we'll show you how to do it, or you can check with the local, you know, your application vendor and ask their help desk how to give all the stars to the calculus of IT podcast.

That's the co it us, the website. And we also have a substack board, the calculus of it sub stack.com. But that's a lot of, that's a lot of words. So just go to the Coit us, which is, I mean, if you were send out it's the COAs, um, and that will redirect you.

That will redirect you. 

Mike Crispin: Why are we still laughing about that? I just don't, 

Nate McBride: well, Mike, because it's funny. 

Mike Crispin: It is. It's 

Nate McBride: [01:36:00] funny that that domain was not taken yet. 

Mike Crispin: It wa it was not taken. 

Nate McBride: And that I wrote a book called The, 

Mike Crispin: nobody Gets It either. Nobody Gets It When I say, they're like, oh, cool. Like the COIT us.

Like, oh, okay, thanks. Because nobody's listening to me. Anyway. That's the truth. Oh yeah. Great, Mike. Thanks. That's great, Mike. Great. I'll check it out. 

Nate McBride: I listen to every episode. I, I love, I love the show. 

Mike Crispin: Yeah. 

Nate McBride: I love all the insights. Um, yeah. It's the best two hours of my week every week. 

Mike Crispin: I like when people are like, oh, yeah, I gotta catch up on a few of the episodes.

It's like, yeah, it's been about 60 hours since you have 60 hours to listen to. 

Nate McBride: No, dude. It's more like a hundred. 

Mike Crispin: Well, I mean, if you've listened to 40 hours 

Nate McBride: Yeah. 

Mike Crispin: Then you got 60 left. 

Nate McBride: But if you 

Mike Crispin: listen, just because that without ai, 

Nate McBride: every season's, episodes. [01:37:00] So let's see, like 30, 39 and 15 and whatever. If you put 'em all back to back, but at two and a half X speed, you could literally ingest the entire calculus of it.

Podcasts in a day,

Mike Crispin: A day, 

Nate McBride: a day. No, I have no idea. I just made that up. It could be some number of time That seems reasonable to me, and so I'm going with it. 

Mike Crispin: Sounds, sounds, sounds reasonable. 

Nate McBride: Okay. Hey, a month from now we'll be at pax, by the way. 

Mike Crispin: I cannot wait. I'm very excited. I, I, I really had fun going last year. That was great.

Nate McBride: We're bringing, we're bringing our co-op with us. 

Mike Crispin: Nice. 

Nate McBride: So it's me, Mike, and Miranda, and then our co-op Natalie. And I was like, Hey, what are you doing on that day? And she looked at me kind of weird like, Hey, I think I'm. I'm working for you say, Nope, [01:38:00] it's a field trip. It's mandatory. 

Mike Crispin: You're going, 

Nate McBride: we're going to Pax, so 

Mike Crispin: I can't wait.

That was so much fun. Last 

Nate McBride: time you got, you got all your kids. They're coming, right? So Brio and, 

Mike Crispin: yep. Yep. We're Bo. It's blocked off on the calendar. We still have to get the tickets though, 

Nate McBride: so dude, just go on, go on the internet and buy it. Go to the inter 

Mike Crispin: web. Web. 

Nate McBride: Just open the tubes. 

Mike Crispin: We will do that. 

Nate McBride: All right.

Well, Mike, it's been real. I'm gonna go shove down apple pie and just burp. Burp until I sleep or burp. Did 

Mike Crispin: you get a McDonald's apple pie? 

Nate McBride: No, it was just leftover apple pie in the fridge from like two weeks ago. I'm sure it's probably still, I'm sure it's probably still pretty good. 

Mike Crispin: Trying to think of what I'm gonna.

Have, 

Nate McBride: if you microwave it, by the way, microwave anything, it immediately kills all the [01:39:00] germs. That's a true fact. 

Mike Crispin: I got rid of my, my microwave broke, so I haven't had one in four months. Now 

Nate McBride: they literally give them away. 

Mike Crispin: Yeah. I not, I didn't replace it. We got the counter space back. 

Nate McBride: Yeah, yeah. But you get like a, an EP absent printer or a microwave, and they're both free.

Mike Crispin: Wow. 

Nate McBride: Yeah. 

Mike Crispin: Wow. Maybe I'll get one someday. I'm trying to go without it. I haven't needed it. So 

Nate McBride: how do you eat your burritos? 

Mike Crispin: I use the air fryer.

I do, I use the air fryer for everything 

Nate McBride: you do. I use air fryer for frozen burritos. There's no way. 

Mike Crispin: Absolutely. Absolutely. Yeah. I burn the shit out of the top and crack it open. It's awesome. It's fricking awesome. I love using that thing. There's a little like, you can put it in a little grate or you can use the little pen.

Nate McBride: Yeah. I know how enterprise, I have one, but I wouldn't cook a burrito on one 'cause I would be [01:40:00] eating a shit log. 

Mike Crispin: I love shit logs so good. I what makes a burrito a shit log? Just so you fucking too much. Well, 

Nate McBride: its like completely nukes to a solid. It was a solid mess. 

Mike Crispin: So if you put it in the microwave, it's soft still.

Right? You like 

Nate McBride: that? Oh, if you follow the direction, it's a minute and a half on one side, a minute and a half on the other. You come out, it's like soft and gooey, put in the air fryer, it comes outta a shit log, a solid mess. 

Mike Crispin: Do you take, do you put a little, uh, the little silver poison thing underneath to make it cook faster in the microwave or 

Nate McBride: no?

No, I put, I put shichi on it afterwards. I would not put in a 

Mike Crispin: fryer. So what happens? What do you mean the shit log? How does it become a shit log in the air fry? '

Nate McBride: cause when you, when you eat it, Mike, it goes through your body hole. It doesn't, it get like [01:41:00] broken down. 'cause your body's like, I can't even, 

Mike Crispin: I can't eat.

I chew it up. I chew it up and it turns into dust. 

Nate McBride: Nope. No, no. You should read about the elementary canal. You chew it up and it's a, it's a burrito that you just deep fried? 

Mike Crispin: No, it's air fried. It's like 

Nate McBride: the toast oven. Air fry. Yeah. You air fry it and then you put it in your body and your body just ignores it.

It doesn't extract like proteins or anything. So 

Mike Crispin: new get with microwaves is better. Like the, 

Nate McBride: yeah, it's all soft. 

Mike Crispin: But what if it's like radiation and shit in there? Like, 

Nate McBride: oh no MythBusters prove there's no radiation in microwaves. 

Mike Crispin: Oh, good 

Nate McBride: Season, season 15 episode. I think six. There's no, there's no radiation in microwave.

Mike Crispin: I noticed since I got rid of my microwave, my beard is growing faster.

Nate McBride: No, that's not. Then there, there it's, 

Mike Crispin: you don't think that's true? 

Nate McBride: There's no correlation there. [01:42:00] 

Mike Crispin: Mm. I have a feeling 

Nate McBride: that's real. My, my beard grows with time. Like if a day goes by, my beard grows more than it was the day before. It's not because of where I was or what I was eating or where I sat or what kind of The 

Mike Crispin: more you shave it, does it come in heavier each time?

The more 

Nate McBride: you 

Mike Crispin: cut 

Nate McBride: you, you already like that ship sailed. So I was dry shaving when I was 12 and 13 so I can have a full beard. 

Mike Crispin: Wow. 

Nate McBride: Because I met some surfer guy in Newport Beach in Rhode Island and he's like, Hey man, you know you can fix that patchy thing. I was growing my face. He's like, you got, she dry sheep every day.

Every day. So I bought 10 packs of fucking kill. 

Oh. 

Nate McBride: That hurt Like a mother. But for two years. I bought 10 packs of razors from CVS, like these single blades. And every single day I would dry shave in the morning and dry shave again at night. And then when I was like 15, I had a full beard. When I went to high school, [01:43:00] I was like the only senior who legitimately looked like he was 28, like legitimately.

And, uh, I had full, I had full fucking beard, not like, uh, half beard, but like, um, game of Thrones beard. Wow. Like the big beard. Seriously. And I, I was 17 and 18 and people were just, it was this big black beard and I had the, the chops and the whole thing and the long hair. I legitimate looked like a, uh, piece of shit, um, artist from Greenwich Village.

Wow. But I was old. Hmm. So, you know. Yeah. So Mike, if you start dry shaving every day, in fact, buy a 10 pack of single blade razors and keep 'em in your car. In your Tesla. Yep. Your old Tesla side pocket there. [01:44:00] And then as you're driving to work every day, just shave. Oh, 

Mike Crispin: that would be painful. 

Nate McBride: No, I You 

Mike Crispin: mean try, you mean dry shaving?

Like Yeah, literally like just not using any sort of 

Nate McBride: Yeah. 

Mike Crispin: Shaving cream or anything. 

Nate McBride: Yeah, just try shave a couple days. I not 

Mike Crispin: do that. 

Nate McBride: Why not? Your skin is used to a couple days. 

Mike Crispin: I can't do that. 

Nate McBride: Well now this thing I have right here, this is about three weeks, a little over three weeks. It's thick. It's, it's all straight.

It's nice and I don't have to worry about it. And I can shave this tomorrow. It'll be back three weeks. It's, I love it. 

Mike Crispin: Yeah, I've gotta, I gotta, I gotta figure out what I need to do. So dry sh Other than dry, shaving it off. Well, I'm driving. 

No 

Nate McBride: dry shave all the time. Keep in your office, keep in your car.

Every time that you have like a call or something, just dry shave. 

Mike Crispin: God, that's one piece of [01:45:00] advice I'm probably not gonna take from you. 

Nate McBride: You have to stimulate your follicles, Mike. 

Mike Crispin: Mm. 

Nate McBride: Follicular stimulation.

Mike Crispin: Can't do it. 

Nate McBride: Well, if you can't stimulate your own follicles, then I dunno what's gonna happen. 

Mike Crispin: We're gonna have this weird thing come in. We'll see what happens. 

Nate McBride: I, I can't wait. I see. Let it go for a couple more months. Don't, don't, don't touch it. Just let it go. Mm-hmm. And then we'll, we'll, we'll have something to work with.

We'll work with it. 

Mike Crispin: I'm gonna let it go. I'm gonna do it. 

Nate McBride: All right. 

Mike Crispin: I can't, I, I'm very interested to see how this works out for me. Without dry shaving it. 

Nate McBride: Season three, episode 11, beard Watch. Night number one. We're on Beard Watch, folks. Uh, this is night number one. We're, and we're going to, well, 

Mike Crispin: I, well, I started two weeks ago.

Nate McBride: Yeah, we're, we're officially kicking out Beard Watch right now. 

Mike Crispin: Alright, well, everyone can see it. So that's, that's gonna 

Nate McBride: help the [01:46:00] intensity. The intensity is real. Uh, we're, we're alive with my Crispin Beard Watch is alive and, uh, it's, right now it's not so good, but, um, 

Mike Crispin: barely visible. 

Nate McBride: Barely visible, it looks like.

Can 

Mike Crispin: you see anything? Can you 

Nate McBride: see anything? It looks like, it looks like he splash some dirt on his face and then most of his face, so 

Mike Crispin: that's what happens. 

Nate McBride: We'll see what happens with Beard Watch. Um, plus seven. 

Mike Crispin: Such a mess. 

Nate McBride: Yeah. Oh, once you, when did you start shaving her head? I think you got carried away.

You know, 

Mike Crispin: you think I should have kept my hair? 

Nate McBride: I think it's all related. 

Mike Crispin: You think that's what it is? Okay. 

Nate McBride: You're overly foll stimulated. 

Mike Crispin: It's true. It's true. To get some test a coin

so I can buy some dry, some, some shave, some uh, [01:47:00] shaving gear. What do they call the, 

Nate McBride: we go to manscape.com. 

Mike Crispin: Have you used a safety razor before the one blade? No. I tried that. 

Nate McBride: That's what I shaved with when I was a kid. 

Mike Crispin: It was very nice. They're very nice. I like those. 

Nate McBride: Yeah. Could have manscape.com and just brighten.

There are questions I need to grow my facial fa my facial beard. I'm, I'm almost 50 years old. Help me grow a beard. 

Mike Crispin: Help me. I'm 50. I can't grow a beard. And that's the thing when it comes in, it's gonna be white now anyway. So like you 

Nate McBride: Yeah. 

Mike Crispin: But you look like a wizard. 

Nate McBride: Mine's salt and pepper is what they call it.

Mike Crispin: Ooh, I like it. Pepper. 

Nate McBride: Yep. Salt and pepper. Look, I push it real good.[01:48:00] 

Mike Crispin: That's good.

Nate McBride: It's, 

Mike Crispin: that's good. That's good. 

Nate McBride: Oh man. 

Mike Crispin: Sorry. That was, that was, that was good. That was a good one.

Nate McBride: Okay. Well, on that note, 

Mike Crispin: we just, we're just trailing off right now. Just, just trailing off. 

Nate McBride: No domains to buy tonight. So we're gonna piece out and, uh. 

Mike Crispin: Test a coin.com. 

Nate McBride: Test coin.com. 

Mike Crispin: I'm gonna see if, I'm gonna see if that's available. 

Nate McBride: Okay. We're gonna have a domain search. We are, we are gonna actually end this episode with another domain.

Search 

Mike Crispin: test. A coin.com is available for a thousand bucks. 

Nate McBride: Shut the fuck up with an iron. A with an iron. A 

Mike Crispin: with an i.

Nate McBride: I'm gonna drive the price right now. 

Mike Crispin: Test, go test a [01:49:00] coin.com.

Is that, is it 9 95 is what I see? 

Nate McBride: Yeah. Yeah, that's what I see too. 

Do 

Mike Crispin: you see the same thing 

Nate McBride: from from from Huge domains.com? 

Mike Crispin: Oh yeah. 

Nate McBride: Hold on. I'm gonna see what, uh, 

Mike Crispin: gotta get out on this. 

Nate McBride: Let's see here.

Let's find your domain. Get domain test coin.com. It's unavailable, but test coin.net is 14 bucks. 

Mike Crispin: Hmm. We need a better, um,

what helps when you type it into Google, what, what does it think you're [01:50:00] searching for? Have you tried that yet?

Oh,

that's not very good. 

Nate McBride: Well, testa.co dot IN is available. Testa Cohen.

20 bucks.

Mike Crispin: Let's added to the list. We'll see what happens 

Nate McBride: on 

Mike Crispin: the list. 

Nate McBride: It's added. 

Mike Crispin: Oh, it's, it's, it's, it's

interesting. So test a coin already exists.

It's a Solana coin and [01:51:00] it's, it's the first coin within the butt coin meta to act.

Hold on a second. Hold on a second. I don't know if that's just someone, some, someone made that up.

Hold on. I read that out loud. Not thinking what it said. 

Nate McBride: Hold on. There's a bud Bitcoin. Is that trading at 17 cents? No, sorry. 1.70 cents right now. 

Mike Crispin: Wow. 

Nate McBride: Has a market cap of 17.85 million. Um, wow. 

Mike Crispin: One follower, uh, 1 42 followers. Uh, this is gonna go huge. This is big. Yeah. 

Nate McBride: Why didn't we get into Bitcoin, Mike? 

Mike Crispin: I don't know.

Nate McBride: There's 10 point se there's 10,007 [01:52:00] 70 holders of Bitcoin.

Mike Crispin: This is hilarious. 

Nate McBride: Wait, 

Mike Crispin: who created this thing?

I'll send you this offline. We have to bring this, we don't need to go any further with this one. 

Nate McBride: Well, an as coin is at 0 cents. So that's three. 1,000 thousand, 10,003 100. Thousandths of a dollar is how much the market cap for as coin is $39,000. I bet we could beat asco Mike. 

Mike Crispin: I don't know. Looks [01:53:00] like this has taken off.

There's one member of the community. It's huge. 

Nate McBride: Yeah. Yeah. 

Mike Crispin: It's, uh, really 

Nate McBride: ASCO's. All time high is 0.003. So it was at three, 1000th of a cent at its highest. It's lost a thousand thousand percent 

Mike Crispin: DIH coin as well. 

Nate McBride: Is there a shitcoin? 

Mike Crispin: Oh yeah, yeah. That one's been around for a while.

Yeah.

Nate McBride: Shitcoin is at, yep.

Oh, it's got $21 million. [01:54:00] Market cap. Go. Go. Shit coin. 

Mike Crispin: Yep. 21 million bucks. 

Nate McBride: All right. Note to sell. Test a coin. 

Mike Crispin: It's already out there, man. Which we, we already missed it. 

Nate McBride: Well, we're going for the better one. Quite coin. 

Mike Crispin: There you go. Done.

Let's do it. 

Nate McBride: Okay. 

Mike Crispin: No problem. I'm dead serious. Let's get this thing going. 

Nate McBride: I'm on it. Let's do it. 

Mike Crispin: I'm looking right now.

BOIT coin.

There's no results. Wow. We, we found something. We got this. This is awesome. This is gonna be great. 

Nate McBride: Quit coin. I will. I made a note. It will set up KU coin. I'll set up tomorrow. 

Mike Crispin: You got the exchange going. [01:55:00] We'll get we'll. We'll get this going on. Coinbase end up in jail.

Nate McBride: I will pump and dump coin.

All right. 

Mike Crispin: Okie do. 

Nate McBride: Okie dokey. I'll see you on the flip flop. 

Mike Crispin: Have a great night. Good. 

Nate McBride: You 

Mike Crispin: too, Mike. Talking about Chateau I, 

Nate McBride: Mike, you have Mike. You have a 

Mike Crispin: great night. Have a great time with that apple pie, man. I know you love them.

Trance Bot: The calculus of it,

season three,

verifying this identity.[01:56:00] 

Sometimes you just have to take it.

Sometimes you just have to take it

because it's season three

divided autonomy,

verifying identity.

The calculus of [01:57:00] it.