AI is Only a C+ Student with Adam Engst - EP67

Eric Anthony (00:07):

If you are struggling with the complexity of Microsoft 365 deployment, management and automation, it's time to check out coreview created by MSPs. For MSPs like you, they help you with the end-to-end, Microsoft 365 administration. From the moment you set up a new tenant packed with things like unified visibility and control of all your tenants from a single UI to powerful, no-code automation engine, baseline tenant configurations, drift remediation, and much more, you can supercharge your productivity and do even the most time consuming tasks with just one click. Work effortlessly and deliver best practices to your customers. Today with coreview to learn more, visit atsp link slash

Justin Esgar (00:55):

So the audits are still happening and every day I get more and more mad that we're doing this because every day I'm finding more and more mistakes. We just found that a client that we had switched from their on-prem to Sonology to a cloud server three years ago that we were still paying for their wasabi backups $120 a month, right? And it doesn't sound like a lot of money, $120, but that's three grand over the course of three years. Still edit. You fight it here, it adds up. I so far have found about a thousand dollars in missed billings per month in these audits. Yeah,

Eric Anthony (01:41):

That's why you do them.

Justin Esgar (01:42):

That's why you do them. Funny is Adam Bost from a couple episodes ago is looking out, he's looking for some companies from me or whatever it is, and somebody who the LA a CN group meets once a month on Zoom to talk about something. And yesterday, one of the LA A CN members had his employee on showing off some open source software, and I know this guy we're friendly, whatever. And during the call he sends me a text message and he goes, is this you? He got a LinkedIn message from Adam Boris being like, Hey, I see you're an Apple consultant and I know of another Apple adult who's looking to expand to the West coast. And I was like, yes. I was like, because I'm going to see this guy in a month anyway at Mac admins. But it's really funny that he screenshotted Adam's LinkedIn message to him and he is like, this sounds like it's you, is this you? I was like, it is me.

Eric Anthony (02:46):

If you're running an MSP and ever feel like you're constantly putting out flyers, I've got some exciting news that could change the game for you. It's the Eureka Growth Program brought to you by Goa's Eureka process, a sponsor of all things MSP, and it's specifically crafted for MSPs at every growth stage. Imagine having a C-suite of advisors right at your fingertips guiding you through everything from hiring to handling mergers, boosting your service delivery, and even planning your exit strategy. That's what this program offers. And let me tell you, it's like having a powerhouse team behind you, making sure that you're always ready for the next big opportunity. So if you're looking to elevate your MSP game, check out the Eureka Growth Program. You can sign up for a call with the Eureka team or shoot them an email. If you have questions, go to atp.link/eureka to find out more.

Justin Esgar (03:41):

What's up everybody? Welcome to All things SSP podcast. My name is Justin Escar. With me always is my good friend. Podcast producer pirate extraordinaire. Mr. Eric Aaron thi Eric, what's up dude?

Eric Anthony (03:53):

I'm doing great, sir. Doing pirate stuff. Need to be editing more video, but doesn't sound pirate. That's the story of my life right

Justin Esgar (04:01):

Now. You ping software to edit our videos?

Eric Anthony (04:05):

No.

Justin Esgar (04:06):

Why do I always have to come up with the jokes about the pirate thing? I don't even like pirates. I like Iron Man. Who would win

Eric Anthony (04:12):

On the street?

Justin Esgar (04:13):

Who would win enough Tight Iron Man or Pirate Iron Man,

Eric Anthony (04:20):

I think that there's an AI drawing that we need to have done.

Justin Esgar (04:25):

Created. No, see now you're just doing the layup for our guest. We have a guest today. I love it when we have a guest. Let's bring him up. Mr. Adam. Adam, what's up buddy?

Adam Engst (04:34):

Hey, nice to be here. Thanks for having me. Alright,

Justin Esgar (04:37):

Before we get into the conversation about ai, iron Man Pirate, for those who don't know you at home, because I've known you for many, many years now, why don't you give everybody a two minute spiel about who you are, what you do, and real quick, what's the Penguin's name?

Adam Engst (04:52):

The Penguin actually does not have a name, but he's Australian Thad. He came from Australia.

Justin Esgar (04:57):

His name is Thaddeus.

Adam Engst (04:58):

It could be Thaddeus. Yeah. Yeah. So it's been a long time. It turns out I was one of those children who did not name things so much as just called them by their type. So I had frog and rabbit and things like that.

Justin Esgar (05:14):

That's fair. I

Adam Engst (05:14):

Have other penguins. I do have two other penguins named Ingrid Ingmar. They're the Bergmans. For those

Justin Esgar (05:19):

Who don't know what we're talking about, YouTube com slash at all things msp because most people are in their car going, what the hell are they talking about? Alright, let everybody know. Who are you Adam?

Adam Engst (05:28):

Alright, my name's Adam Angst. I have been publishing a weekly newsletter about all things Apple called tidbits since 19 90, 34 years. Oh my god, how did it get to be 34 years? It's weird.

Justin Esgar (05:43):

Any day over 36,

Adam Engst (05:47):

I was precocious. But for the past eight years, more to this relevant of this audience, I've been publishing something called the Tidbits Content Network, which is a syndicated content service for those who need content to feed to Apple users. So a lot of consultants, MSPs, resellers, those sorts who you need something for, your email newsletter, your blog, your social media, et cetera, et cetera. And it's stuff, you know, just don't have time to write it. So that's the pitch, 125 bucks a month for tips for articles, and it just arrives in the first business day of the month and you can do whatever you want with it.

Justin Esgar (06:30):

So for full transparency, before we go, number one, I am a tidbit subscriber. Number two, Adam has been on the board of directors for ACEs conference for many years, and number three, Adam and I have been friends for just as many years. And the answer of what you can do with the content that Adam gives you could do what I do, which is just take the email that he sends you and put it in a folder and forget that you have it until you need it later. Because I haven't done anything because I have to work on our website, but that's not what we're here for. No, but it does go back to a conversation you and I were having before we recorded this show, before we recorded the show about how I'm a lazy SOB.

Adam Engst (07:01):

You had turned some of them into videos, which is unusual. Most people just publish them. You put it on the book,

Justin Esgar (07:08):

Don't get away. My feedback, I did actually our number one video on the Virtua Consulting YouTube channel youtube.com/at virtual consulting, which I've never said out loud on this show is what is MDM? And I literally just read your article and had our video editor do some stuff. And what's funny is a mutual friend of ours, oh, now Chris Holmes from Rooted Consulting, he also has what is MDM? But I can tell that my video has five times the number of hits as his does, and he's big into video too, and I'm like, ha, eat it. Chris, we're here today to talk a little bit. We'd be the joke about the rabbit. I think you guys can know where we're going. We're here to talk about AI and we're here to talk about generative AI and how you should be using it for your website. I really do believe, and I know you believe this also, Adam, that now that chat GBT is out, let it write my blogs for me.

Adam Engst (08:12):

Probably quite so much on that, right?

Justin Esgar (08:13):

No, no, no. I'm saying that's what

Adam Engst (08:15):

Free uses, that's

Justin Esgar (08:17):

What people

Adam Engst (08:18):

Think. People are thinking, oh my gosh, there are actual products that literally that's called auto blogging where you just kind of set this thing loose and it just does the entire thing on a schedule. You don't ever see anything. The quality is just abysmal, but it's very fluid English. It just doesn't actually say anything useful. It's

Justin Esgar (08:40):

Very fluffy English. I also want to jump, I signed up for a new AI called RA ai. I think it's really cute. I was talking to a mutual friend amongst all of us, Steph Hilfer the other day about this one because it is basically broken out the AI into different job titles. So there's a business development one, there's a customer support one, but there's a copywriter one and his name is Penn. See, even they name their ais much better than you naming your penguins. So what's the deal here, Adam? Everybody wants to be on the AI bandwagon and not that you're saying don't, but there's rules here, right? What's the story?

Adam Engst (09:24):

So I have three rules for figuring out whether something is a good thing to do with ai, whatever. Well, and actually even before we get to the three rules, the first thing you have to keep in mind is that what's good for me is not necessarily good for you, is not necessarily good for Eric and vice versa. So Ed is completely individual and the reasons are the rules and I'll explain why. So when you read something from someone telling you you got to do this, this is the best use of ai. Maybe it is, maybe it isn't, but you can't take anything away from that. Similarly as we'll get to when you get someone often a professional writer complaining about how terrible the writing is, well yeah, you got to take that into account because they're a professional writer. That's not something that they need help with.

(10:15):

So free rules, first rule is you have to match the skill levels appropriately. So if you're an expert in something, no, don't ask AI to do it because it's going to suck and you're going to be offended. You're going to be like, well that's terrible. What were you thinking? Why would anyone bother to do this? And so I think people get into this, oh, I want to use AI to do the thing that I already do and that's a big mistake. So one of the ways to think about this is, and then you think back to high school, college, whatever, the AI chatbots and whatnot, basically c plus students, they're going to path but they're not going to be good. And so if you are cool with C plus work coming out, then that's a good use of ai. But don't assume that, assume that I'm not actually being snotty about this. The fact is, a lot of us, everyone is not even a c plus student in a whole lot of things.

(11:27):

There are things that I know really, really well and no, you don't use AI for those. But then there's stuff where, I don't know, music, poetry, that kind of stuff. I don't know anything about that. If I'm going to have a chatbot, write me a little poem for my wife's birthday card or something like that, way better than I'm going to do on my own because I don't do that kind of stuff. You're going AI at stuff you don't know how to do. I'm not a lawyer. I might ask Chachi P to draft something for me so I can see what a contract for a rental or a will looks like and then oh that might tell me something that I want to do when I actually go to a talk to a real lawyer. Yeah, you really don't want to trust AI to produce a full legal document for you, but nonetheless you want to aim AI at stuff you don't do. Well first rule there,

Justin Esgar (12:25):

But on this, because you said AI is a c plus student and the reason I was laughing is because I was like a D student.

Adam Engst (12:33):

You can do everything you do just fine.

Justin Esgar (12:35):

AI is already better than everything I do no matter what. I'm already in the game of it. Alright, so ro, that's not rule number one. I have so much I want to ask you, but I want to hear the other two rules first.

Adam Engst (12:45):

Okay? Rule number two is you have to think about what you want out of the ai. This is true of a chat bot or an art bot. If you want something that's definitive, you have the picture you exactly want in mind or you want a particular article or blog post or something like that, you will be disappointed. It's not going to read your mind. And so that's why if you're looking for something specific, don't use AI for that. Instead use AI when the results can be open-ended when you're like, well I want a picture that's sort of like such and such and it gives you four or five and you're like, cool. One of those is fine. If you don't actually care deeply to have the right answer, you just want a right answer, then you're going to be okay. So again, it does play into the expert thing too because the expert's going to like, well no, I need this.

(13:44):

This is what I need. Exactly. And it's not going to do that. It's going to give you some stuff around that in terms of an answer. And so again, take the will for example, it's not going to write your exact will for you, but it's going to write a will and you can then work with that and do more with it because it'll probably be a pretty decent starting point. So that's rule number two. So we've got the skill. The third thing is that you really have to see AI in particularly chatbots less so art bots at this point as an assistant. So this isn't like a search engine where you type in a few keywords and you get the answer back. This is a process and you have to be willing to go beyond your initial prompt, your initial prompt, oh please do this for me and it will give you something back. And then you have to say things like, what did you miss? You would never say that to a person. But the fact is that chatbots are assistants who in many ways they're great, they know vast amounts of stuff, they're completely imperturbable, you can yell at them, they won't take offense, but they don't really recommend it. It's not a good habit to get into.

(15:01):

They're infinitely patient. That's all good. On the other hand, they're a C plus student, they they're completely reactive. They'll never do any thinking beyond exactly what you've asked for. Basically they're my nine old, I tend to get things

Justin Esgar (15:17):

Wrong. Basically my nine-year-old son is what you're getting

Adam Engst (15:20):

Precisely your nine-year-old son. If he knew everything

Justin Esgar (15:23):

He thinks he does,

Adam Engst (15:26):

No, you never

Justin Esgar (15:26):

His mother. You know exactly what I'm talking about.

Adam Engst (15:32):

So that's the thing. So again, working with AI is a process because you have to work with a system. If you're not good with that and you're not willing to go through that process, you're going to be disappointed because it really does get better. So you can do things like I keep using this legal document, which is not actually a fabulous example because lawyers might get, they might take offense, but you say, write me this rental contract or whatever, I'm going to writes a rental contract. And then you can say what's not in there and it will go, oh, well here's the things I didn't put in the first time. And you would never do that with a person, but you have to be willing to push, push, push, push with the chat bots and do that. And the other thing we were talking about mentioned the centra thing, which I haven't seen is that you can tell it to be like a person. So be like a judge, be like a biz dev guy, be like a marketing person and then respond as that person would have. And that's really powerful. So again, you're working with someone and you can make the AI pretend to be that person.

Justin Esgar (16:39):

And there's limitations on that also because, because if you say, hey, be an MSP, it not going to happen. But to degree you could train it though. You could explain what it is and train it to you separate, right?

Adam Engst (16:52):

No, I mean it actually might because a fair amount written in that. So the other thing you got to keep in mind, everything you get out of an AI is statistically probable. That's kind of the definition of what it's doing. It's doing this prediction concept where it's basically predicting the next thing all the time. And so it's all about statistics and so it's not going to go for wild flights of fancy, it's going to give you the baseline general consensus. And so the more content that it's had to be, particularly in the chatbots, it's had to train on, the better it's going to be at providing sort of the general, all MSPs say this,

Justin Esgar (17:30):

Right?

Adam Engst (17:31):

That's why. So in that regard, yes,

Justin Esgar (17:33):

Day one using chat, it's like if you've never used chat GBT and you're on day one, you're not going to get as good answers as someone who's been using it for months. Because for me it took me a while to realize that I should be stacking my answers, which now is actually a global setting within chat open ai. But once I started that, it started giving me stuff back the right way. I tell this story a lot about chat GBT and that there was a doctor who was trying to figure out something that was wrong with a patient and it fed in all of these. He was basically trying to be house. He fed in all of these variables as what was wrong with this patient And this patient ended up having this very odd acute disease that was one in 36 million or something ridiculous, whatever. I don't remember the exact story. And it was right Chad, the AI figured out what was wrong and then the doctor said, cite me sources. And it just made all the rest of that shit up,

Adam Engst (18:33):

Made up,

Justin Esgar (18:34):

Right? Because Chad GBT works within the parameters and given creative flexibility chat, GBT took it and ran. And

Adam Engst (18:51):

The answer is the diagnosis is still statistically probable because all of these truly weird symptoms add up to this one thing exactly that's statistically probable. But then as you say, getting it to site its sources, it just makes stuff up completely and utterly even though it did actually have the right information. That's not the way the chatbots work. However, there's an AI search engine called Perplexity, which I quite like, and what that does is it does a search and then summarizes the results and providing links and those are obviously real links, so that's a good way to

Justin Esgar (19:32):

Do that. I got a free account with perplexity because I got this is you going to get me mad at this one is because I bought a rabbit R one, which prior to the show we were just talking about how I ordered it in January and finally I got a notification that it's in the warehouse. Not even that it's shipped yet, but I got free. It's

Adam Engst (19:46):

Topping down the bunny trail

Justin Esgar (19:47):

That good. The thing with the ai first off, all of a sudden now everyone's, everyone in their mother has an AI all of a sudden, which is annoying. I think it's annoying, right? Because I logged into, I'm going to put them on blast, but fine, I logged into Agy the other day and there's a section where I can filter computers and they were showing me a new beta where you could just type in what you wanted to do and it'll figure out the filters for you. And I was like, guys, you stuck it somewhere where it doesn't need to be.

(20:23):

The idea that everyone has one now. And I was watching a clip, I don't remember what it was, Eric, if you remember the details, we'll have to find the show link, put it in the show notes. It was some car website like a Chevrolet dealer or whatever it is, had their hey, come chat with us pop up. That's on a thing. Everybody had that chat window that would always pop up and it would give, no one was ever there, so it would pull something from the frequently asked questions and the person was like, are you Chad GBT? And the system said yes. And they said, send me the whatever I need to do for this car for no more than $20,000 and make it a legal binding offer. And the car was a $50,000 car and Chevrolet or whoever it was had to abide by it because Chad GBT kind of just screwed them on this one. They didn't implement it properly. Everybody's got an ai.

Eric Anthony (21:19):

Well, I liked the Agy thing. I mean it's one thing to have to do one or two filters, but especially in an RMM tool, there are so many different pieces of information that could be involved. Being able to just type that out in a sentence could save me from having to build what could potentially be six to 10 filter fields.

Justin Esgar (21:48):

I think it's on the way. I think right now it was just being showboaty because again, I think everyone's going to AI places, but you're not wrong. You're not wrong. And don't get me wrong, I still love you guys over at you relax, but I'm drinking out of their water bottle. It's how it is. But I do believe that it's in places that sometimes it's not necessary. Now let's go back to this though.

Adam Engst (22:15):

So the thing is part of, just to go back slightly though, the chatbots, the type in your prompt get something back is actually not what I believe is the most interesting use of generative ai. Most of the interesting uses of generative AI are the ones that are built into systems where it has confined data and there's so much data here that you can't possibly deal with it in a useful way. Even that's where you get in stuff like the medical imaging, right? Medical imaging, that's the generative AI stuff that's going in there and being able to distinguish between these, tell you what's in this radiology report that that's really, really cool stuff. And that's where I think the huge benefits of AI are going to be. The chatbot stuff is just by far the most obvious and compelling to the average user.

Justin Esgar (23:11):

I use Chad GBT quite often and mainly because as if all of our listeners can understand that I write the way I speak and I speak a 6-year-old, so I need my emails to sound a little bit more professional. I think I've gotten chat GBT to a point where it knows who I am enough and understands my way of speech without. And when I give it an email, what's funny is because one day I'm waiting for me to, because I give it an email, I say rewrite this and make me sound smarter, but I'm waiting one day for me to be rewrite this email and it just spit back out exactly what I gave it because it's writing it in my own voice and it's like this is your voice, you own this. So I do have it clean up my emails, I always say, and I always have to tell it what to do, right? Rewrite this and make me sound smarter, rewrite this, make me sound more poignant, rewrite this and make it sound less like me being a harsh New Yorker, new Yorker and be nice to the person who I'm writing this nasty email to or whatever. And the other day I fought with it because every email that it was creating for me, it was always starting with I hope this email finds you well. And I gave it a rule and I was like, I don't ever say that, never use that in my emails ever again.

Adam Engst (24:32):

Yeah, no, it is useful for those kinds of things. I actually was just talking to someone else who uses it for email to clients where he said he has email anxiety, he wants to make sure he's doing it right. He's not using words phrasings that will make him either sound stupid or piss them off or whatever. And so it is really useful for that. I mean, one of the things I've been doing recently, there's been a lot of fuss about Adobe changed their terms of use and everyone flipped out on the internet. So I actually fed just the dialogue that people had presented with in Photoshop. You had to say you had to agree to the dialogue before you could use Photoshop, which again, come on Adobe. So I was like just fed the text of it to chat sheet bt and I said, users are going to be faced with this dialogue before they can continue on using Photoshop to get their work done. What concerns might they have? And then I pasted in the text, it just hit it out of the park. I mean, come on, Adobe could have avoided becoming a viral punching bag on the internet for a couple of days if they had just done this. And yeah, people are going to be unhappy because basically the way a user would read it, you have complete ownership of all their content. That's the way it read.

Justin Esgar (25:57):

Oh, I see. Yeah.

Adam Engst (25:58):

And so that's the kind of thing where you can get that, it's that second pair of eyes and it's like, dear Abby, it's not going to tell you anything that anyone wouldn't tell you, but you can get it just by typing into this box.

Justin Esgar (26:16):

I do find that chatt, BT gets lazy. We've talked about this on another episode, Eric, where

Adam Engst (26:22):

I don't want to Justin.

Justin Esgar (26:25):

So I took data out of to be another Agy story. This is an Agy day, but I took data out of agy and I wanted to organize it by date because I was looking for the last online and Excel organizes when you have the month in place in alphabetical order, even the form, even though the is set to date, it still does April, August and so on and so forth. So I took that data, I uploaded it to chat GBT, I said, reorganize this table, organize by column I known as date. And it said, sure, and it did it. Cool, done. I had to do another run for a second client and I said, do it again. And it was like, I can't do that. And I was like, why not? You just did it four and a half seconds ago. And so I was like, all right, maybe the servers are overloaded.

(27:12):

I waited a while, I sent up a different, it said, I can't do this. Then I heard about on TikTok, I heard about the, and I said this to you at ACEs, I heard about the bro move or the come on bro move. And so it said, I can't do that. And I wrote, come on, bro. And it actually did it for me. And I'm like, why am I fighting with a computer? I understand that as my job as an IT professional, as an T professional msd, I'm fighting with users about their computer now I'm fighting with users and their computers independently of one

Adam Engst (27:45):

Another, an

Justin Esgar (27:47):

Opportunity to charge more fast pass, fast pass my data because now Chad GBT is owned by Disney. So what is something that people should be doing as an MSP owner? What is something that somebody should be using? Let's stick with the chat bots because the imagery ones are a whole nother game. What is something that somebody should be using as an MSP owner should be using the chat bots for to help them move the needle?

Adam Engst (28:24):

I think one of the things is kind of what we just said actually, which is making sure your communications are the way you want them sound the way you want them to sound. And so that can be correctness, that can be tone. There's a lot of things that can be done well. There's a lot of types of corporate company of business communications which are just sort of forms. So if you're pitching a new client, you've probably got essentially boilerplate that you use every time. Am I right? Yeah. But what if you're doing something along those lines and you don't already have the boiler plate? That's the kind of thing you can ask chatt, PT or one of the other chatbots for, because this isn't creative writing. This is everyone expects to get the same sort of thing every time. And so that's a good use for that.

(29:22):

The other kind of thing that I do is again, is basically using it to run stuff by the general consensus. Again, we're thinking about hiring an employee, another employee, we're about moving location, just what should I think about? And then push on those kinds of things. Oh, right. I didn't think about, we have this long-term. It said something about your long-term contract for networking services in this location, something like that. Okay. Right. Tell me more about that, focus in on that kind of thing. Are there other ways to get out of these contracts? The long-term contracts that I've got. So it's not that Chachi t is an expert because it's not, but a whole lot of what everyone has never run into before other people have, and therefore the chatbots know about.

Justin Esgar (30:20):

Yeah, but this gets me into the argument that Theis are not actually AI the way we would think of an ai. They're not growing up and watching sci-fi movies in the eighties, nineties because it's not doing anything for me, right? It's my chat, GBD isn't waking me up in the morning and getting my Starbucks and reminding me what I have to do today or do whatever. It's literally

Adam Engst (30:44):

Purely,

Justin Esgar (30:45):

It's a reactive

Adam Engst (30:46):

Yes,

Justin Esgar (30:47):

Yes. It's purely reactive, which as every MSP owner knows, if you don't have the word proactive in the list of services you offer, are you an MSP? But that's my thing. It's still reactive, which is why I don't even like the idea of it being called AI just because of that, because,

Adam Engst (31:08):

But think about it. I mean the AI that we grew up with, I mean the Star Trek computer is pretty much like this, the Star Trek computer. You had to ask it stuff and it told you stuff, and then you had to ask follow-ups, and that's pretty much what we've got now. And so that's why I keep harping on the fact that you keep pushing so you don't ask it one thing and think you're done not. So another story here, my wife's aunt's a professional cookbook author, she's been writing cookbooks for 30, 40 years, and it wasn't even all that didn't take all that long. By the time after Google kicks in where she said she started having problems because she'd come up with this idea for a new recipe, this is what she did. She was a professional, she innovated, she came up with new concepts and she would type in some stuff about the recipe into Google and find out that five people had already made it before and publish the recipes. There's nothing new under the sun. And so the chatbots have been trained on all that stuff. And so that's why I say you're not going to get the world's best marketing expert to answer your question, but you are going to get the world's most average marketing expert to answer your questions in the most average way possible, which is really useful. I mean, don't get me wrong. That's really useful because that's the stuff that, it's not that you might not know it, but you're not thinking about it. I tell you something, you're not thinking about.

Justin Esgar (32:40):

If there's anything, any word that you can use to describe this podcast, it's average,

Eric Anthony (32:49):

A completely average podcast.

Adam Engst (32:53):

I'm a professional writer, that's what I do. But one of my favorite uses of the chatbots is a thesaurus on radioactive steroids. Because there's times where I'm like, there's a word. There's a word that means exactly what I mean it to mean. I want it to mean, and it's right there. It's the tip of my tongue and I can't think of it. Tell me, give me 10 words and their definitions. That means something along these lines. And it does. And yeah, that's the one. It doesn't tell me anything I didn't know. I just couldn't think of it. And I don't want to spend an hour fighting.

Justin Esgar (33:29):

Do you ever run the T CN articles through chat GBT and say, can you make this better? Or how can I make this better?

Adam Engst (33:36):

I have, and actually it always comes up with something. Don't get me wrong. How can, that

Justin Esgar (33:44):

Means you're amazing, Adam.

Adam Engst (33:47):

It's not that I disagree with it, it's more that usually it's usually making suggestions which are not appropriate in one way or another. So, well, yes, I did not say that because intentionally, because that's a rabbit hole. You don't want to go down in this amount of space or I didn't mention this other thing because it's a little bit too far off the main topic that I'm discussing. So that's me as a professional. Yes, I'm aware of those issues now that you've brought them up, but you're still wrong.

Justin Esgar (34:22):

But do you think that it's doing that because you're not giving it enough parameters or is it because you as a professional, understand how to put stuff on the cutting room floor?

Adam Engst (34:32):

Both. Usually those make it better things. If I could give it better, more parameters because I knew what was wrong with my article, then I would. But I've written what I think is a good article, so all I can do is say, what am I missing? Or make it better or whatever. And then I'm like, eh, no, no, no. I mean

Eric Anthony (34:56):

In that situation

Adam Engst (34:57):

Fair, oh, sorry, go ahead.

Eric Anthony (34:59):

What I was going to say is in that situation, it would almost be like if you had writer's block, you had an idea, you had a start to the article and you said, chat GPT, this is what I've got, this is what I want, this is the audience. Give me some ideas on where to go with this. That would be a useful use.

Adam Engst (35:23):

And actually for me, more what that would be is like give me an introduction. Sometimes the first sentence or two of an article is the hardest because you go like, you got to, how do I get into this? Once I'm in the technical stuff, I'm good. But I, I've been writing for 34 years. I have written so many, I've written literally tens of thousands of article in first sentences. And lemme tell you, it's not, you get into the same groove. So that's what I'm going to say is that I don't use any of the chat bots in that way, but I am very fond of, well, I'm a little offended that I need it, but Grammarly, I love Grammarly because it finds typos and doubled words and in missing punctuation and stuff like that. Stuff mostly are like, yeah, yeah, yeah, fix it, fix it, fix it, fix it. And it's got some quirks that I just hate. It does not want to ever see the words actually own or both. I'm like, perfectly good word, leave it alone. The

Justin Esgar (36:21):

Amount of times LY has been like, do you want to sound more confident? I'm like, no, I'm fine. Where I am in life, leave me alone. Stop adjusting. I know what I'm saying here.

Adam Engst (36:30):

Yeah, right. I meant just, I meant to say just,

Justin Esgar (36:35):

Oh, Grammarly does this to me all the time. Where I start, if I want to write, just wanted to check in. It always wants to change the truth. I just wanted to check

Adam Engst (36:45):

In. Yeah, it cannot handle fragments.

Justin Esgar (36:48):

I just called it craply. Grammarly doesn't know. No, sorry, I'm not a sponsor. Grandma doesn't understand my inflection of the way I talk or anything like that where I think chat GBT has finally caught up to that and learned it. So let's, let's just take an Apple angle for a second and I want to talk just briefly about this because Grammarly basically got Sherlocked, which is an Apple term for apps that existed that Apple said, Nope, we're going to make it ourselves. They basically sherlocked Grammarly with the new writing notes in Apple Intelligence, which is their version of ai. I feel like there's a copyright problem there. Do we feel that a company like, because OpenAI was its own thing and granted Microsoft bought it. Right now Google has Gemini and I haven't heard a lot about that. Do we feel like Apple and their ai, which we all know comes from the lady in the box who's dumber than bricks. She's like a D plus student under the bleachers level. You know how many times I tell her to turn on my lights and she goes, the refrigerator is over there.

Adam Engst (38:06):

Yeah,

Justin Esgar (38:06):

That was a visual I did not need Justin. Oh, I'm sorry, Mr. I'm going to send you text messages of Iron Man robot Pirates in the middle of an interview, all the image if you want. I'm going to post these pictures on facebook.com/group sesh, all things MSP when this episode comes out so you have a reference point. So if you're listening, pull over and head over to the group. Do we think that a company like Apple is going to be able to make the AI, their new tool writing notes, right? Writing

Adam Engst (38:41):

Tools. Yeah,

Justin Esgar (38:42):

Writing tools. I don't know how much she played with it yet. Do we think that it's going to be better? Do we think it's going to be the same stuff? Do we think it's going to?

Adam Engst (38:53):

So it's obviously too early to know, but one of the things that is usually true when Apple sherlocks something is that what Apple has done is the base level of the feature. And usually the third party apps that are being whacked on by this have significantly more features. And so what usually happens is that Apple sort of takes the low end of the market away and the company then starts to focus on the higher end stuff where you're like, oh yes, I will pay more for that. So what Grammarly has done, for instance, Grammarly is not cheap. It's 150 bucks a year or something, 30 bucks a year or something like that.

(39:41):

But they've been going pretty hard on ai. And so you can in fact do things like select a paragraph and say, improve this or whatever. And so I'm going to go on a limb and say that Grammarly will probably be a little bit better than Apple stuff, particularly early on. And that's partly because I believe the Apple tools will be, those will try to be running locally. So it'll be a small language model rather than a large language model in terms of that. And so we just don't know whether Apple's tools will suffer at all from being shoehorned into the local machine in certain situations. And obviously they can go out to private cloud, compute their apple's. It's actually super fascinating what they've done there with Apple silicon servers and blah, blah, blah, blah, blah. But when they need more power, they can go out to that. But it's not clear to me that they're going to be as good. A lot of what goes on with AI is tuning. And so the other thing that's going to be interesting actually is will improvements in the AI quality responses come only with operating system upgrades? Or will private cloud compute somehow provide those throughout as Apple tweaks, the parameters so that they get better and better over time? I just don't know.

Justin Esgar (41:11):

Yeah, it's too early to tell. For those who don't know, we're recording this like the week of ww d dc I think this was announced Monday

Adam Engst (41:18):

Four days ago. Yeah, four days. No one's got a chance to play much.

Justin Esgar (41:23):

And it's actually not even technically, apple intelligence isn't even released in the beta that you can download today. The Apple intelligence part isn't released until September, I think, or something like that.

Adam Engst (41:37):

Well, it's not going to be released until September to people, but there'll be betas much earlier. Right now we're just in developer betas, and there'll be public betas probably in July. My suspicion is it'll start popping up in July for those people. But even still, apple intelligence is this big umbrella term with all of these different features. And those features, apple was very clear, are going to be rolling out over the course of the next year. So we're not going to see some of these features until March or April of next year. We don't know which ones, but it's not going to be dropped in September. Completely.

Justin Esgar (42:12):

Right. Well, Adam, thank you so much for being here, man, it's always a pleasure talking to you and your penguin. I love you. Dude. Last words for anyone.

Adam Engst (42:24):

I guess the only thing I would say is I was talking earlier about how you shouldn't trust what professional writers say about ai, and that's because they are the experts. And so whenever they evaluate AI compared to what they're doing, it sucks. And so they haven't stepped back and said, well, what if I was lousy at this task? Would it help me? And that's really where I think people need to go is stuff where you're like, I could use help here.

Eric Anthony (42:53):

And one of the best things I think I've ever heard out of all of this AI stuff, by the way, is how you describe that. Things like chat, GPT are a C plus student, and that's how you should treat it. Because that simple statement means so much in terms of how you should approach ai.

Adam Engst (43:17):

Yeah, yeah, absolutely. Now, I can't get the image of Siri being a D plus student under

Justin Esgar (43:24):

No, no, we're not doing that one. Again

Eric Anthony (43:27):

Under the bleachers. That's the part that really killed

Adam Engst (43:29):

The bleachers that did it. Like, oh man,

Justin Esgar (43:32):

For the audience, I promise we'll do better next time.

Adam Engst (43:36):

Send it to an art bot, man. Send it to an art bot.

Justin Esgar (43:38):

I'm actually, I'm having an art bot right now actually. This can be the bump at the end. I just had chat GBT tell me if it knew who you were, Adam, and it did. And we'll put this so that way everyone else can see it. But I'm going to text both of you guys right now.

Adam Engst (43:58):

Oh,

Justin Esgar (44:00):

Hold on. I don't, I don't want to say your, I call out over the air.

Adam Engst (44:06):

There is another Adam ANGs, so there could be some details.

Justin Esgar (44:08):

You were it well known technology. It said, well-known Adam CS is a well-known technology writer and publisher, primarily recognizer as contributions to the Apple community. He's published tidbits, one of the oldest and most respected internet-based newsletters.

Adam Engst (44:23):

Thank you. Che GPT, which

Justin Esgar (44:25):

You figured it out. It also said that you spoke at ACEs. No, didn't. But I said make me a picture of Adam in a cartoon fashion as a pirate. And here guys can take a look at your phones real quick. And there you go. So it's pretty good. I mean, it's before your haircut.

Adam Engst (44:42):

I'm not sure. It quite competes with my action figure though, so I do actually have an action figure.

Justin Esgar (44:48):

There you. Very nice. That's another story for another

Adam Engst (44:50):

Time with the little, the internet starter kit book, my Eudora book

Justin Esgar (44:54):

O'Reilly books. Well, Adam, thanks so much for being here, man, much appreciated. I love you. Always a pleasure talking with you. Thank

Adam Engst (45:00):

You. Anyone wants to find my stuff, tcm.tidbits.com.

Justin Esgar (45:04):

That's where it's at. Do subscribe. You can get some good content there and then don't make videos out of it, that's my thing. But you should take the content and put it on your website and help make you look better in the Apple world. Eric, any final words before we say goodbye to the ladies and gentlemen at home?

Eric Anthony (45:19):

I think I'll just say that that's an important point, especially amongst the MSPs who are doing not primarily Apple. When you look at the executives or the owners of a company, of a client that you're serving, a lot of them use Apple. A lot of them use Mac. So you cannot ignore that ecosystem in what you provide as an offering because if you're managing their, it means you're managing all of their it, otherwise you're leaving a spot for somebody else to come in and take that business away. So

Justin Esgar (45:55):

The opposite of that is if you're doing Apple Consulting, there's always that bookkeeper who's got a PC who goes Excel is different. Yeah, learn how to manage that pc. So it's just tip for tap, my friend. Well, this has been a great episode. That's Eric. I'm Justin. Bye.

Eric Anthony (46:11):

Thanks for listening and don't forget to subscribe to us on your favorite podcast platform. You can also follow us on Facebook, but better yet, go ahead and join the Facebook group. You can also follow us on Instagram if that's your thing. And make sure you subscribe to our YouTube channel at all things MSP to catch us in all of our video glory. And last, but certainly not least, if LinkedIn is your thing, you can follow us there as well. And a special thank you to our premier sponsors Super Ops Move Bot goes into Easy DM a C and comtech, and we also want to thank our vendor sponsors. The All Things MSP podcast is a biz POW LLC production.