343: Can You Make Your Own Software?

Welcome to the Automotive Diagnostic Podcast.

We're going to explore ways to sharpen our diagnostic skills, find learning resources,

and hear from experts in the automotive field.

This episode is brought to you by L1 Automotive Training and Keith Perkins. If you're looking

for education on module programming, J2534, eProm work, key and immobilizer, electrical

over 60 hours of training videos on all those subjects and more.

When I first started out doing mobile, I utilized Keith's videos on module programming and J2534

in order to get my head wrapped around what I would need for the tooling, the computers,

the software setups, what kind of obstacles I would be up against when I'm out there programming

modules on cars, and it was a huge benefit to me. And I continue to use the training videos

that he has on his website. So I strongly recommend checking out L1Training.com. The link is in the

show notes. Hey, what's going on Automotive World? Welcome to another episode of the Automotive

Diagnostic Podcast. My name is Sean Tipping. I'll be your host once again for this week's episode.

Thank you so much for joining me. Just me on the show this week. And I'm going to be bringing you

some of my most recent thoughts working in the automotive space, which is what I do primarily

on this show. And, you know, a lot of the times that is a case study on a vehicle or a particular

topic that we're dealing with on vehicles while we're out in the field. Today's a little bit

different, although if you've been listening recently, you know, this is a recurring topic for

me. And I don't mean to turn the Automotive Diagnostic Podcast into every other episode is about AI

tools. But I would be doing everyone who's listening a disservice if I wasn't talking about

literally the coolest things that are actually making a difference in, you know, fixing cars and

my business that I have going on. And that is these tools that I may be talking about today. Now,

specifically what I want to dial in on, because obviously I've talked about a lot of various

things with AI systems and, you know, some of the agentic opportunities that are out there right

now. I talked about that a few episodes ago, but today I want to dial in specifically on something

which can be really useful to you as a technician or potentially as a business owner,

is actually creating your own software. And I've talked to several people about this because

I've been experimenting with this and developing some stuff for myself and for the technicians

and my business as a whole. And, you know, not everything's a hit. You know, sometimes you'll

make something and okay, it doesn't quite, it's not quite as useful or doesn't do the thing that

I was hoping it would. But then other times, if you keep experimenting with this, you can,

when we have built things, they're actually really useful and really move things along

with the productivity of the business or just, you know, something that helps us or not wasting

time. You know, that's usually the idea of a software is to help you complete or automate

some sort of process, speed something up or a repetitive task, like we've talked about before,

but that you can make your own for whatever pain point you have in your daily operations,

fixing cars, running a business, whatever it might be. Like, you know that pain point and

you can potentially build a software to help you with that. Now, when I talk to people about this

and I tell them what I've built with it, and I'll share some examples here as well,

the most common response I've gotten from just about everybody, you know, especially my personal

life is like, I would have no idea where to begin with this. Like, I don't understand any like

software coding, building software, I wouldn't even know where to start. And the fact of the

matter is, hey, me either, I was talking about this with somebody just a couple of days ago, I'm

like, I should not be able to build software. Like, I don't have that skill set. I have never

taken a course on software coding. I've never built any of my own software. I don't know how to

operate in the terminal, which is like the black command screen that comes up on Windows.

Like, I don't know any of that stuff. I've never taken any time. I've never taken any classes.

Just not something that I know how to do. Can I operate a computer? Can I operate software

programs? Sure. And I think the amount of work that I have had with computers, with Windows,

with laptops has helped me a little bit. But the tools that are now available to you with some of

these AI models, jump jumps things ahead, like crazy, it, it's like, like, I've learned a lot of

stuff. And it's like, I've gotten to sit shotgun with an expert coder. And that's really what this

is. You can interact with some of these AI tools. And all you need to know is, okay, what is the

pain point? Right? What is the thing that is eating up my time that is repetitive, that is

wasting time in my day? But it is something that, you know, I need to get done, right? I need to

find this calibration within this 1500 page document. I need to reference this hardware number

and figure out if it's compatible with the vehicle it's trying to go in. I need to find this file

that needs to go into this Nissan from, you know, 15,000 files that I have stored between

a Google Drive and a laptop folder, these sort of things that, yes, we've built, you know,

shortcuts and, you know, okay, we figured out how to use control F and the ways to search for

things to speed things up. But if there is something like that, we're like, this is a clunky

part of the process, right? This, on a regular basis, slows things down, right? And here's the

thing, you know that best, I can't tell you what that is. Like, again, I'll give you some suggestions

of what's been good for me. But you know, the things that are hanging you up, and that are

repetitive, and that if you could imagine just, you know, doing what that is with a click of a

button, it would be great. Those are the types of things. So, and that's actually a really important

part because, you know, the people who build software, traditionally, right before these systems,

aren't necessarily aren't necessarily going to be the people who know what the pain point is,

right? That's you, you're in it every day, you're experiencing it. And it might be something that

is very unique to what you are doing. And we find that within mobile, right? That there's, yes,

there's other mobile companies out there, there's other programmers out there. But if you take the

general workforce as a whole, even just did within the automotive space, our problems are

very unique compared to everybody as a whole. And so, understandably, there might just not be a whole

lot of effort put into building something that is going to relieve the pain point that I have

particularly with my job. That is not true anymore, because you can build something that is uniquely

specific, as specific as you want it to be to solve your specific problem. You just need to

know exactly what that is. And of course, it's probably going to have something to do with a

computer. Obviously, if it's outside of the computer, it's in the physical world, not to say

that you couldn't build something that would help you. But this is generally going to have

things to do with stuff that is on a computer on a laptop connected to some sort of drive,

some sort of digital information. But obviously, fixing cars nowadays, there is a ton of that.

So, again, finding what that pain point is, what the problem that you need to solve is,

that's going to be you. But I have a feeling you could come up with a few things where like,

this sucks, I hate doing this, it eats up so much time, it's clunky, it's really repetitive,

it wastes my time, I'd be willing to pay somebody else to do this instead of me doing it,

those are the types of things. So, you get your problem, you identify your pain point.

And now, you just need to articulate that problem to one of these AI models. Now,

I do want to pause there and touch on that there are obviously various types of AI,

large language models out there. And at the moment, there are a number of companies that

are competing for that top spot in a number of different categories. But obviously, there's

a lot of money up for grabs for whoever can be the best. And if you follow it all,

what's out there right now, as far as the American companies go, which are generally perceived to be

at the top overall, there are some Chinese companies that are doing impressive things,

but from the US, it's Anthropic, OpenAI, and then Google with Gemini. And ChatGPT or OpenAI,

that was kind of the first one that came around in the fall of 22. And obviously,

I've talked about it and I've used it plenty and I still do. Google's Gemini has also been

along for the ride and they've come and gone as far as the usefulness. I actually think they've

made some big strides in the last six months. So if you haven't tried out Gemini, give it a try.

And Gemini does work really well within the Google workspace. So if you're already using

the G Suite or the workspace for Google and you have Google Docs and Sheets and all that stuff,

it does work really well within that ecosystem, obviously. It's all Google. I would say its overall

capabilities compared to the other two still might be a little bit behind. But the one personally

that I feel is ahead of everyone right now, and this is March 15th, 26th, that could change in

six months easily, probably will. But right now, it's anthropic and Claude. And especially for

building software systems, it is 100% the best. And the reason that I can say this,

and again, it's personal opinion, so maybe you disagree or there might be specific

things where the other models might be ahead or might be better. But when I have something new

that I'm going to attempt to do that maybe I haven't solved yet with any AI model is I will

put the prompt into all three and just hit go on all of them. And then I will review the output from

all three. And then I'll see, okay, you know, which one is getting me to where I need to go.

The smoothest and the most accurately, of course, it does depend on what I'm what I am attempting

to do. But honestly, over the last six months, that has routinely been Claude that ends up being

the best and the most accurate and moves me along better than the other models. So it's the one that

I've been using for this. You know, I talked about using open claw and that you do have a choice of

what model that runs on. But the one I have runs on Claude's opus 4.6, which is their most advanced

operating system that they have right now. But it seems to be the best, right? And when I say the

best, like, the answer that you get, it's, I get that like, Oh, wow, that's a really good answer

for either a question I'm asking research, I'm trying to perform or thing I'm trying to build.

It most often gets me that like, Oh, damn, that's awesome. I'm blown away by the output from it and

what it's able to do. And it's, it seems to be more intuitive to understanding what I'm trying to

accomplish. Without without all the extra fluff, that is one thing about chat GPT, I'm not the only

one who's observed this is it's just like, sometimes it gives you so much and you can regulate that,

right? Don't get me wrong, you can instruct it to, Hey, keep your answer brief and to the point and

you can adjust it. But that's just seems to be built into Claude at the moment. Anyways, pick

your model that you like the best experiment to because again, tomorrow, one could be better,

right? There's new model comes out and then all of a sudden, Gemini is ahead or open AI is ahead.

The one thing about this is if you're not experimenting with these things on a regular

basis, and let's say you haven't really done anything with it within six months, the technology

and the outputs have completely changed in that timeframe, the thing it's moving so fast, it's

getting better so fast. Again, kind of scary if you extrapolate it out, because I don't think

that's going to slow down at all. But for right now, as far as the capability of what these have,

if you haven't tried it out in a while, give it a try, because it's significantly better than it

was six months ago, a year ago, whatever timeframe. So anyways, right now, again, I'm using Claude,

but pick your AI, identify your pain point, and then you want to articulate that pain point and

what you'd like to accomplish to whatever model that you pick. Okay, so let's just use an example

here. We use the Fed World Report. If you've never Googled this, just Google Fed World Report for

Chrysler, Jeep, Ram, Dodge, Stellantis vehicles, when they need an update for either a PCM or TCM

it's a powertrain module. Now, there's other ways that you can find whether there's a software update

for a particular module. You can go to the Stellantis website and figure this out as well.

Obviously, you could be hooked up with YTech, but away from the car, we need a method. But one of the

reasons that I utilize the Fed World Report is that it also includes this show is brought to you by

Auto Rescue Tools and Isaac Rodel. If you've been looking for a programming laptop, you're not sure

which one to buy or how to set it up, especially if you want to program multiple brands.

You've got some domestic vehicles, you've got European vehicles, can the same software go on

the same laptop? What size hard drive do I need? All those questions. Isaac's your guy. He can custom

set up programming laptops that are ready to tackle any make or model. I've got one of these

laptops myself and I can say that it is outstanding and it really streamlines the process by having

everything you need in one device. So if you're looking for something like that, I highly recommend

checking out AutoRescueTools.com. You'll also find scan tools, diagnostic equipment, key cutting

equipment and much more. Check out the link in the show notes. I highly recommend it.

The applicable hardware for PCMs and TCMs. It is only powertrain modules and then a couple,

it does include like a couple other ones on some models, but primarily it's PCM and TCM,

but it gives you the applicable hardware. And so one of the more common things that we get is

used replacement ECMs, PCMs on these Chrysler Dodge vehicles, and sometimes the hardware is

different. And we can correct the software part of it, but only if the hardware is correct for

that application. Now you might say, well, just look up the part number for the car. And if it's

not the part number, it's not right. Not exactly. There are hardware applications or there are vehicles

that can utilize multiple hardware applications, right? And some of the hardware applications work

across a lot of different models within the various Chryslers that have been manufactured.

And so just because it's not an exact match for the hardware doesn't necessarily mean that it's

wrong. It might work. And here's the other thing. Sometimes we get them where we don't have the

original. And so again, we could look up the hardware number, but that's how I utilize this

sheet, right? To figure out, okay, do I have the most current software update? Like what software

does this car, is this car supposed to have? Because it tells you to make in the model.

And then is the hardware applicable or what hardware would I need? And we just use a series of

Control F and screenshots to be able to figure this out for this report. But what I did was,

because it's a little bit of a tedious process, and I have it down with some Control F,

I guess I wouldn't even really call it tedious. But I've done this a ton. I've done this hundreds

of times looking up various modules and softwares here. My employees, especially maybe a new employee,

has not done that as often. And so I have to, why have I made a video on how to do it,

giving an example of a Jeep that had a TCM that was thrown in it. And it turned out,

I think it was the wrong hardware. And I went through and figured out, okay, this is the wrong

hardware in this Jeep. We don't have the original one, but we can tell like, it's not going to work

in this application. And using that Fed World Report to show that these are the ones that would be

acceptable. But this one isn't. And here's how we know based off the software, that sort of stuff.

And so that's my example to any employee that kind of wants to train on this stuff. And they've

got to do it a few times and it's clunky at first. But I explained all of this in plain English.

And there's another thing I want to pause on when you're doing this, you are literally just talking

to one of these AI models, you can use the transcription or the dictation feature, the

little microphone, once you're in the application, and just talk to it and explain, here's the thing

that's hanging me up. Here's the tedious process, the repetitive process, the thing that

wastes a bunch of time. And here's what I want out of it, right? Here's the reason why I'm doing it.

And here's what I would be nice if a software could do this for me and take care of the control

F and search and then go to the next one and search and then explain that to it, which is exactly what

I did, right? I explained how I use the document, what's in the document, where I look and how I

use it to compare. And then it gets to work with you on building something that is going to do

that for you, right? It's just a big PDF. So you're going to feed it the PDF. It's a 1500 page

document, explain what's what, it's going to go through it. And that's where this gets really cool

is you may not know exactly how to convert the data within a PDF into a software, but it does.

And that's where I'm going along, riding shotgun on an expert coder that is now building a software

because it knows exactly how to take this data and then condense it down or put it into a format,

which then it's going to be able to reference and then do what you want. So now what we have

from this tool is something where we can either put in the software number and then it tells us,

here's what it's for. Here's if there are any updates and here's the compatible hardware,

or we can put in a hardware number and then it tells us what that hardware applies to. And it's

just one software application that we load from the desktop. We can even do it offline. You can do

it on a computer or you can do it on a phone. And you punch in those numbers and then it gives you

the output based on that fed world report. Now, if there's updates to it, we'll have to update

its reference point, but it's not very difficult. But it built this out for me in I had it done

in a matter of a couple hours while I was doing other stuff. That's the other thing is the building,

the actual like heavy lifting and the coding and stuff. This is happening behind the scenes,

you know, with the AI tool. And you can actually pull down and look at the coding

as it's building it. And it'll have probably and it should have questions for you along the way.

And then you'll test it out and you'll try to, oh, okay, maybe this doesn't work exactly like I

thought. Or I forgot about doing this one particular thing. Can we go back and can we change this?

And then it can edit it and can add things and it can adapt. And that's the other cool thing. Like

I mentioned, you may have a specific problem and you may want things specifically your way.

You can ask it for that. You can build it so that it takes care of whatever your problem is,

the exact way that you want it to take care of. You literally just have to know what the problem

is and then explain what you want fixed or what you want it to do to it. And it's going to be able

to build something for you or suggests a way to build something for you that's going to take care

of that. And that's it. And that's where this is so cool. Because like I said, I don't know how to

code stuff. I don't know how to build software. But I know what problems we have. I know where

the sticking points are in the day to day and the things that hang me up and the things that I get

calls about maybe for my employees. And then I have to, you know, make a video on or explain.

And now I can just point to, hey, just punch it into this software, right? Here's the tool.

You put this number in, it gets you the output. Boom, you're done. Okay. Now, again, maybe we

have to update, maybe we have to tweak. But the fact that we can actually make these things

and then make them however we want with whatever details or features that we want to add, we can

make it better whenever I have the time and that, you know, the desire to go in and work on these

things and tweak them. And we've been doing this with a number of different things within the business

that are speeding things up, making things more accurate and automating a lot of tasks that we

would normally take a lot of time on. The other one I mentioned was the Nissan file lookup. For

anybody that's programmed Nissan's, you have probably had an instance where you purchase the

file, because that's how you program a Nissan either a blank or an update is you purchase the

file and then you put it into the repro programming folder that consult lives on. And when you go to

put it in that folder, it says, do you want to replace this folder or this file? And you're like,

ah, I already, I already had that file. I just spent another $55, which they just keep bumping

the price up on those files all the time, because you can reuse those files. So after you buy it,

you can use it as many times as you want if you're programming that same application.

So you don't need to spend the $55 on the next one. But searching for those files is not the most

straightforward thing. And I had figured out how to go into the CSV lookup table or the HTML,

and then actually find the file name that I need to search for within the drive. And then you

can find it. But again, that's clunky. It's time consuming. It's a thing I've actually had a tough

time, like really getting clear across, you know, to my employees all the time, they struggle with

the details on it, we end up buying files that we don't need. This software does all of that for

us and not only tells you, like we put in the name of the file as it appears on the website,

because that's the problem is the website, the Nissan website doesn't tell you the file name

until you buy it and download it, then you have it, of course. But you can get what file you need

off the Nissan website, put it into this tool, it not only has the capability of checking our

drive and our laptops, and that's a little extra, that's the extra thing I wanted to add to this.

But it tells you the file name. So then you can actually search the file name if you wanted to,

but I took it a step further and the software just tells us if we have it and where it is.

And if it's not there, okay, then let's buy it. But that saves us an enormous amount of time,

number one of going back and forth and using control f within these spreadsheets. And that's

good to know how it works. And I would say if you're going to solve a problem, you should already

know how to solve it, you just want an automated way to solve it a better, faster, less time consuming

way to solve it. But you should probably know how to actually solve the problem. But you've

realized that man, this eats up a lot of my time every time I do it, or it's difficult to explain

to someone else how to do it. That's where this comes in handy. And then you, you can articulate

that even better to your tool that's building the software is like, Hey, here's what I do to

solve the problem. Can you build something that does that? And then I'll say, yeah, okay, let's

give this a try. Let's do this. And then you can add things and you can adjust. So those are just

a couple examples of what we've done. And again, I just want to stress as much as possible.

These aren't tools that I'm building because I'm an expert coder, I understand any of that stuff.

Again, just going to what I've had people say to me is like, I have no, I would have no idea

where to start. It's literally just identifying the problem, articulating it well, and then iterating

with it until it does what you want it to do. And you'll be amazed the output that you can get out.

One final note that I should mention is when you are building these tools with these AI systems,

it is going to burn tokens, right? And there's various levels of free and paid plans. And so

depending on where you're at, you might hit your limit on a particular application. And then,

of course, you have to weigh is the money I'm spending on this subscription worth the output

that I'm getting, right? And that's for you to decide. For me personally, in my business,

what I'm able to do with these things is well worth the money that I'm spending on it.

But I do just want to throw that out there so that nobody is mistakenly thinking that you

could do this on the free plan for one of these applications, you might be able to get started,

but it's going to hit its limit pretty quick. And then you'll be like, Oh, okay, I can't finish it.

Just be aware that the tokens do cost money. And so you'd have to be willing to put in

a little bit of cash to get some of these systems going, depending on what you're trying to do.

Maybe your project is straightforward and simple and you don't need much, but the more complex

and the more iterations of it that you build, the more data that's going into it,

that's all going to consume resources. And I think with time that price will come down,

right, especially as different companies are being more competitive and they build in

efficiencies within the AI systems themselves. But it's not an infinite resource at the moment.

And so you may run up against those limits as you're experimenting and trying to build with

this stuff. But really what I wanted to get across to everybody listening and again,

the prime candidates for something like this or listeners of this show is to find something in

your day to day that is like I mentioned, that's a repetitive pain point and go to one of these

systems and just see, Hey, can you build me something that is going to solve or help me solve

this problem or automate this task or get past this thing a little bit quicker and more efficiently

than I would be able to otherwise. And those are the type of things that really move the needle

throughout your day. It's cool, but they're not just toys, right? They're actually helping you

in production, they're moving things along, they're making you more productive. And like I talked about

with the AI agents is they're freeing you up for things that actually move the needle, the things

that you do best that only you can do that there's no way an AI could do one particular task or this

particular task makes you the most amount of money in your business or your job. Now you don't have to

be, you know, searching through PDFs, wasting your time doing that, right? It's a means to an end

that has to happen, but you can speed that up, you can automate that part. So I encourage everybody

to look for something like that in your day to day, and then experiment with one of these tools

and see what you can build. And please let me know, right? Shoot it to me and Messenger if you

don't want to share it public or put it in the Facebook group and show like what have you built,

what have you done with one of these tools because it's pretty cool stuff. But that's where I'm going

to wrap this one up today. I really appreciate everybody listening, all the feedback I get about

the show. If I don't get back to you immediately, don't think I'm ignoring you. I try to get back

to everybody I can. I know some of them do slip through the cracks, but I appreciate everybody

that listens to the show. So with that out of the way, let's get out there, start fixing the world,

one car at a time.

343: Can You Make Your Own Software?

About this episode

L1 Automotive Training

key and immobilizer

module programming

J2534

eProm work

Fed World Report

Stellantis

software update for vehicle modules

Dodge Ram Dodge

TCM

PCM

Auto Rescue Tools

ECM

hardware application

Nissan file lookup

Request an Explanation

Want to learn more?

Help improve this episode

More from Automotive Diagnostic Podcast