Why Oil & Gas Professionals Are Canceling SaaS and Building Their Own AI Tools

Peter Brecht00:01

Okay, Wade, we hit start on the show because we were going to get off into some wild tangent about Claude code. But today, so I spent the last two hours, two hours, yeah, two hours started as two hours ago, open Claude code on my computer, had a problem that I wanted to solve for our team, put the prompt in, connected the GitHub repo, got all that set up. So it's ready to go and built the first model, get it done. It was a WordPress plugin. So basically.

Wade00:25

Right.

Peter Brecht00:31

We have a website that we use for all of our app tools, et cetera, for our team clients. They all go there and we kind of set that up separately from our normal website. But I said, build this as a WordPress plugin. So it builds the plugin. I upload it. It doesn't work. Like, okay, like I'm a little bit down. I'm a little bummed. Go back into cloud code. I tell it what it's not doing. It rebuilds it. I download it, deploy it, call my account manager up and I'm like,

Hey, I need you to take these pictures, upload them to the website, and then I need you to tell me what the output comes out. And it did it. And she's like, she was over the moon. And I was too excited. Like I emailed her and then I'm like, no, I got to call her and like, get this done. Am, my gosh, I am so impr- I don't know how to put this, but if you told me three months ago that I would be building stuff like this at the-

Wade01:08

Mm-hmm.

Wade01:12

You got to share in the joy. Yeah.

Peter Brecht01:28

At the, with this capability, I think I would have said it's impossible. The reality is that I feel like anything is possible at this point. I'm just, this is just admin stuff, right? So this is just Hey, we've got this problem. It's taking us 15 minutes a day to do it over 30 people. What's the time savings on a weekly basis? And not only just the time savings, the frustration and the tediousness of doing some of these things, it's making it to where.

Wade01:42

Right.

Peter Brecht01:59

It's standardizing everything. It's coming out better, faster, cleaner. Like I'm just, I can't get over it. And then, then yesterday, so this is another part of it. And I posted this on LinkedIn. We have all these applications from insurance carriers that are not fillable, right? So what I mean by that is you have to open up the PDF. You have to get, click the markup tool and then type over it as you're talking with the client.

Wade02:16

Okay.

Right.

Peter Brecht02:25

Some carriers have figured out that they need to be fillable, but not all of them. There's one particular application that we fill out probably two or three times a day across 30 people for EPLI coverage. It's not fillable. No one told me this until I sent an email to our team and I'm like, hey, start sending me stuff that you need that you use on a daily basis that you want improved. So I get one of these and I built a cloud tool and same thing.

I dropped the drop it in. It, it doesn't, it even did this with a protected PDF, which is scary. It took the protected PDF, stripped the wording out, right? Rebuilt it as an HTML page with fillable and then made it a clean PDF. And I'm just like, I can do this all day long. Can take care of your applications and just modify them, merge them, whatever I want. The data is not because none of this is hard anymore.

Wade03:03

Uh-huh.

Wade03:06

Okay.

Yeah, yep.

Peter Brecht03:22

I'm just it's unlocking things I've never seen before. So as giddy as I am, I'm talking to a couple of operators now that are getting ready to go do some drills and like all this fun stuff. I mean, it is providing time for me to like go research, go learn more. I don't think I can keep up learning, but that's on my side. On the oil and gas side, I don't even I can't even imagine like the things that you as a drilling engineer are looking at. So I wanted to kind of open up and be like, what are you

Trying if you, I'm sure there's gotta be a lot to it, but like, what are some use cases that you might see on your end of things? Cause it's insane.

Wade03:54

Right.

Wade04:00

Yeah, what you're saying is like, it's funny because I want to feel like, that brings back memories. But those memories for were literally like three months ago. And I think that's the steepness of the slope that we're on right now. I mean, even something like just last week when they went from Opus four point five to four point six. I mean, it was just a huge change in the things that they did and how repeatable it is. But when I first started out like.

Peter Brecht04:21

Mm-hmm.

Wade04:25

I was using Cloud Code, just like what you're talking about, and I'm kind of getting rid of my mundane tasks. One of them was just making PDF searchable, right? Like OCRing PDFs and doing some stuff like that. Because there's scripts out there that'll do it, and I was using Cloud Code to kind of help run these scripts and set up a deal to where anybody could drop a PDF into a network folder. When that network folder gets that PDF, it scans it OCRs, it saves it as now a searchable document so that everybody else can kind of use it and play with it.

But since then, I what you're talking about, I've probably used Claude to write. I mean, I wouldn't, don't think I'm probably exaggerating whenever I say a couple hundred thousand lines of code and the programs that we're building. I mean, right now it's analyzing data in near real time, whether that's frac data or, even like mud reports and things like that. It is taking, not only what's going on right now, it's comparing it with.

What has been happening on this well, it's comparing it with offset wells that we have data on. And then not only that, like the real value in it, and I think the analysis is that it's looking forward to what's going to happen next, or it's kind of saying like, hey, here's what these next stages look like. Here's a similar stage in a similar stratigraphic level that we had earlier in the well, and here's what happened there. You you might be ready for either challenges or we expect it to be smooth.

And it's just doing this, it's gone from when I first started using it was kind of transforming data, making it to where I could kind of use it. I could bring it into Excel or I could do whatever I wanted with it, you know? And now I've built it to where it's doing a little bit more analysis. Now, the scary thing of that is, right, is that it's not always right. I would tell you, I don't always agree with what the consensus is on, either how we need to run our rigs or...

I'm sure, you I don't do much with Frac, but on, probably how you'd flow back a well or how you'd frack a well. And so there's still, there's a big article going around on X right now about, like the software engineers and how, they see it going. And I think there's a lot of that that's true. It is going to cost some jobs and all that. But I think at the end of the day, you have a really hard time replacing, I don't want to say.

Peter Brecht06:24

Mm-hmm.

Wade06:45

Decision-making but kind of that risk analysis and where you kind of fall as an individual or as a company in terms of risk analysis and everything. So it's super exciting. I'm 100 % with you. I use Cloud Code every day. I it started out, I was kind of consulting doing some of that and the expectation was 15 to 20 hours a week maybe doing it. I got so into it. I want to say that I've freed up so much time, but honestly, I'm just using

Any extra time I have to code stuff, to make programs that make me more efficient or make me better because I'm probably not exaggerating when I say it's been 40 to 50 hours a week right now. And this is not like work time. This is usually like nights and weekends, just having an idea and building things. It's a huge dopamine hit for me. I mean, it really is. I love that you can have an idea. And like you're saying, within...

Sometimes, depends on how complex it is, right? I mean, I'm working on something right now that is super, super complex and it's, I've been probably working on it off and on for weeks now. But a lot of the other stuff, like you can get an idea and have an 80 % product within 30 minutes to an hour. And at that point, it's just what I like to do, that my personality is, I like to use it. And then I'm like, well I've got this idea. And I go back and I add to it, or.

I'm using it, there's a problem with it. You go back, so like that iterative process of deploying it, using it, changing it, like it's never ending, which I think is also why it's taking so much time, or why I'm spending so much time with it. But the output of it's amazing. There was one day I had an idea, I just want to scrape all the data from the OCC website. Like I want to be able to see every well within a five mile radius of wherever I'm about to drill. And I want a table automatically output of

Peter Brecht08:10

Mm-hmm.

Wade08:38

What their casing design was and within 30 minutes, like 80 % there. Was, it's mind blowing. It really is. It's crazy the kind of growth. Now the funny thing is, let me get your opinion on this, guys. So I got a question for you. So I'm reading X and they're always, and what they're talking about out there is with AI right now, there's like, I mean, we're kind of, it feels like we're going through this exponential growth, right? Like every iteration is just getting better and better and better and better. Now,

Peter Brecht08:46

That's it.

Wade09:09

Exponential growth can't continue indefinitely. It kind of feels like at the beginning of COVID when they were doing all their projections for sickness and death, COVID was bad, right? But like the early projections I saw, literally on the news were saying 75 to 90 million people, this was in March, they were like 75 to 90 million people will, well, they literally said something like die, I think, by like May, June timeframe. And it was because of that exponential slope going up, it was very early.

I feel like that's kind of where we are with AI. Like we're still on that really exponential slope. The question for me is, is not, candidate in, kind of continuing definitely. I don't think that's possible. Like my question is, where's it going to level off and what's going to be the bottleneck? Is it, is it hardware? Is it energy? Is it just kind of, you've got the low hanging fruit and now we're getting more and more, complex, I guess. And so my question is like,

Where do you feel about it? Do you feel like, cause you kind of feel like, just talking to you, like you're in this euphoria stage.

Peter Brecht10:14

It's euphoric in the sense that I'm able to now solve problems that I wasn't able to solve before. Like I could effectively, I could create something like this for myself with an AI tool or a program before, but I wasn't able to allow others to use it. So what I mean by that is I could master how I was doing something.

Wade10:23

Right?

Wade10:40

Right.

Peter Brecht10:44

But the system couldn't be widely dispersed to my team. So there's that piece that I'm thinking about. That's why I'm euphoric is like, I can now let other people partake in what I could. Like that's the part that gets me going. And your mind starts to go through this sort of pattern where you're like, well, if I'm doing this job and other people are doing this job and there's maybe,

Wade10:59

Right.

Peter Brecht11:11

A thousand other people that would use this as a product, you start to think, well, if I sold it for $10, would people download it? Right. And you kind of go to that, but then you start thinking, this is why this is an important thought. It sort of leads to nowhere because eventually the other people will be able to do the same thing or they can already for the same cost and it's their own thing. So really for me, it comes down to a question of adoption. Right. So when I'm

Wade11:20

Right.

Wade11:34

Right.

Peter Brecht11:40

Scrolling on X, I see all these people or potentially bots most likely that are saying like, I'm, building all this crazy software and, I'm able to achieve and I'm shipping things and there's going to be the first billion dollar company run by one person in their AI platform. It's like, I kind of look at that, first of all, I think they're full of it. I don't think that that's actually going to occur. The reason behind it is

Wade11:44

Hahaha

Right.

Peter Brecht12:07

More people are catching on to the fact that they can go do the same thing because the entry point is the tool. So again, going back to your bottleneck piece, there's, we're I think on not the cutting edge necessarily, but we're pretty close to it where it's like, we're some of the first adopters you and I, other people that might be listening to this are some of the first adopters to build their own things and actually use the tool. I think there's people that build stuff and they go, that was cool. I figured out how to

Wade12:10

Right.

Peter Brecht12:34

Vibe code something, right? And it's like, sweet. Then they just kind of let it go. I'm actually using the stuff that I'm building. So it's going to keep getting better. And then my team is using it too, which I think is even cooler because I can have 32 employees and we've continued growing with those 32 employees because we've saved time and they're more efficient, right? That, that efficiency factor is huge. But if I talk to someone like my wife, she goes, well,

Wade12:36

Right.

Peter Brecht13:02

Like I told her what I built this morning and she was like, that's cool, babe. Like that's really cool. Doesn't under, doesn't see where this, what the inferences are. So in my opinion, and the reason why I'm personally bullish on energy is I think that's the bottleneck. Like I'm really, I really am. I think that.

Wade13:07

Right. Yeah, I've shown my wife to you.

Peter Brecht13:26

We're on an upswing with oil and gas. Think we're going to, I think there's going to be more activity for that because there needs to be the demands are increasing. We had an interview last week with Julie Thompson, who was saying that that's like the number one thing is getting the energy. The chips are going to be an issue, obviously, right? That that infrastructure needs to occur. But I think the adoption speed is going to plateau first. And then you're going to have this

Energy draw that doesn't really start to ramp up until that that because like I think that goes up and then it plateaus and then there's a widespread piece and I can see the same kind of concept with I mean you look at crypto or any of this stuff right or cell phones like let's go to cell phone uses your internet adoption, right? It was kind of a it was kind of this super fast increase it plateaued and all of a sudden now like the internet is like, it's because of smartphones was like the next

Wade13:58

Widespread adoption. Yeah.

Wade14:11

Mm-hmm.

Peter Brecht14:20

Increase so like what's that next inflection point? I don't know, but i'm just i'm looking at this and it's sort of it Let put it this way. I'm also a little scared of it because I had a we had a leak i'm just gonna go through that. We had a leak in our house, right? Woke up six o'clock in the morning or go go downstairs wife took a shower i'm standing in the kitchen pouring my cup of coffee and there's water coming down from the ceiling and i'm like

Wade14:22

Right.

Peter Brecht14:50

Oh man, so whatever, we shut the water off, start doing some testing. It looks like it's just a simple thing. Plumber comes over. Okay, plumber's not gonna lose his job from AI, right? This is the conversation I had with my wife yesterday. He was not gonna lose his job, but I'm looking at my six-year-old son and I'm like, what is he gonna be able to build? Do, think, all these things are different. So like learning a trade is great. But then if you look at the robotics that they're building right now too,

Wade14:51

Right.

Wade15:02

Right.

Wade15:11

Right.

Peter Brecht15:17

The heart surgeons, right? There was, I don't know who went through this example, but if you had to get heart surgery done, okay, you have an option. You can choose the 32 year old heart surgeon that's done 50 surgeries, or you can have this, and he's got a steady hand, or you can talk to the, you can get the 70 year old surgeon that's done 3000 heart surgeries, but maybe, maybe he's not as steady, okay? Who are you choosing, right?

Wade15:17

Mm-hmm.

Wade15:30

Mm-hmm.

Wade15:44

Right?

Peter Brecht15:47

Which one would you go with? I'm curious.

Wade15:49

That's a great question. I don't know. I couldn't answer that in no time. What you'd want, right? What you'd want is the young guy holding the scalpel and the older guy running the show. That's what you'd really want.

Peter Brecht16:03

If you had to pick one, I'm going with the older guy. I'm going to pick the guy that's had more reps. But here's the thing. Then you introduce robotics. And robotics is not quite there yet, but we're already where Claude is writing its own code. Think...

Wade16:13

Right.

Right.

Wade16:21

Well, they already use it for the surgery, like laser surgeries and stuff, right? Like the bigger movements translate to the small movements. Like you're stepping towards that precipice there.

Peter Brecht16:25

Right.

Right.

100 % but at some point the robotics will not only have the steadiness of the younger surgeon It will also have the knowledge of every other heart surgery that's ever been done So it starts to shift a lot of the work But what is that all required still requires power if you unplug the unit? It's not working, right? So

Wade16:40

Right.

Wade16:51

Right. Right.

Peter Brecht16:54

I don't know, there's so much to this to unpack. I'm like, have a notebook that I kind of write and I just put down my AI thoughts. Cause like I come up with things and I'm like, okay, this is what I'm thinking about it today. Because it's interesting to see like the development since I started doing it in like October last year, kind of where like my thoughts were and like those are outdated. It's like, that doesn't like the way I was thinking about it's not the same as what I thought about, yesterday.

Wade17:03

Right.

Wade17:08

Right.

Wade17:18

Right.

Peter Brecht17:21

So kind of like an AI journal, it's, would encourage anyone that's in this space to like, kind of keep a note of like their thought process or like what it's doing, because it's interesting to see that evolution. But I don't know if I answered your question. I think the plateau, I think it's going to exponentially increase. Then we're going to have, it's going to plateau energy is going to have to ramp up and then we're going to continue on the next phase of this, which I don't know, man. I'm just hoping everyone's able to keep it together because it's, it is, it's incredible. I don't know what else to say.

Wade17:34

Yeah.

Wade17:51

Yeah, it's definitely going to disrupt a lot. If it doesn't cost current jobs, it's kind of going to cost future jobs, right? In the oil and gas sector, I'd say a lot of smaller companies, whenever you're looking to build something and exit, you want to cut that pie into as few pieces as possible, right? So that you've made as much return as you possibly can.

And so I think these metrics of having huge companies and doing all this stuff, I don't know that that's what the future looks like. I think AI making all of your A players that much more efficient just makes a more efficient and better company. And so it's going to be interesting to see what this does to the job market, not just in energy, but computer information systems, all these different things. It's going to be...

It's probably not going to be as bad as what some people are saying, it's also probably, it's definitely going to cost some careers for sure.

Peter Brecht18:55

Our theme right now, as I have a guy Bart that is, he's my developer on kind of like, just hand him things and I'm like, okay, can you go figure this out? And it's really cool that discussions have been happening with him, but our theme, the conversation we're having across everything right now is what is the job to be done? Right? So we've got these sort of elevated, guess dreams of what we could do.

Wade19:05

Yeah, right.

Peter Brecht19:23

But it's like, what is the job to be done at the moment? Right? Because you have to take everything in step. You have to have a good filing system. There's still people too, like let's not forget human nature, right? Which is there's still people that are hard workers and will do whatever they need to do to get the job done. And then you still have a lot of lazy people that aren't going to work hard no matter what you do. It doesn't matter. Some people look at AI like, I'm just going to be able to sit back and it's going to bring me my food and

Wade19:26

Right.

Wade19:39

Mm-hmm.

Wade19:45

Right. Right.

Peter Brecht19:53

It's all that stuff like it's gonna know that I'm hungry and it's gonna door dash something for me and it's like I look at it like it's a tool that I get to go utilize to keep building more and I think if you don't have that mindset now, I think you're in for a world of hurt because I mean it's so anyways, I'm stoked. Yeah.

Wade19:56

Right.

Wade20:08

Oh, 100%. You're going to... I think you're right. I mean, I think it's going to make your A players or your efficient people even more efficient. And like you just, you gave an example this morning. I mean, like it expanded your scope, right? Like you were able to deploy something that benefited everyone in your company. And whether that's, just makes them a little bit more efficient or takes away a headache, either one of them is great for the functioning of that company. And so that's awesome. And that's where...

I've kind of gotten to the point where I'm like, all right, I've been building all this stuff and I've been using it internally and I've kind of shared, some of it's just reporting that with some AI analysis and stuff, but I'm really trying to find that like, how can I help other people adopt this, so that on a daily basis, they're using it more and more and it's helping them the same way it's helped me, you know.

Peter Brecht21:01

Yep. Yeah. So yeah, it's a there's a that Pablo Picasso quote, which is, learn the rules like a pro so you can break them like an artist, right? So got to stay on top of this stuff so that we can actually build some cool stuff. So anyways, I'm excited. I can't wait to see some of the things you're building when I come out and visit. Hopefully I get to see you soon and just kind of check that out. But like there's something I'm just like, man, I don't want to share all this yet because I'm too like like

Wade21:12

Yeah.

That's right.

Wade21:21

Yeah.

Peter Brecht21:27

I don't want someone to steal my idea, but anyways, it's, yeah.

Wade21:29

Well, then you better be careful what you're putting out there on GitHub then. I put something out there. I built a chat GPT deal that integrates directly into Outlook. And so every email I have, could click either draft or apply, and it would kind of make a first response for me, or analyze reply. And it would look at mine, and I could talk to it inside of my email. So that way, I'm not copying and pasting outside of chat and all that stuff all the time.

Peter Brecht21:41

Okay.

Wade21:59

And I put it out there on GitHub and I made sure to protect my keys and all that stuff. But I accidentally made it a public deal and like I noticed a few people downloaded it. I was like, well, I don't like that they took it. But at the same time, I made it public and it's kind of cool that somebody, looked at it and saw it and was like, oh yeah, let's, now they could have just been scraping stuff, but at the same time, like I've had to really expand my knowledge about like, what's secure, how to secure keys, how to, all the different things. So.

Peter Brecht22:12

Yeah.

Peter Brecht22:17

Yeah.

Peter Brecht22:26

Yeah, I'm not there yet. I haven't attached any APIs to anything yet. I just feel like I'm a little, I will get there and I will give you, yeah. Yeah. Well, and when I get there, I'm going to call you cause that's definitely a consideration. Like, but, speaking of, have you, have you seen open claw or have you, have you seen what people are building with these Mac minis? Like I know this is a, it is, it,

Wade22:32

I got you. Okay. Just Python scripts and stuff? Okay. All right.

Wade22:53

Yeah, yeah. Like over Christmas it was like, it was the thing. Like I was like, is this my feed just feeding me this or is this everybody, is this really everybody doing this? And it just, it absolutely blew up. That's whenever I was, it was kind of funny cause I felt like about a month, month or two before that, or maybe it was even a few months before that. Like I was using Claude code in a way that I was like, this is game changing, at least for me it was.

Peter Brecht22:56

Yeah. Yeah.

Yeah.

Peter Brecht23:18

Mm-hmm.

Wade23:20

And I felt like that was becoming, that wide adoption was happening kind of over Christmas with ClaudeBot or OpenClaw, all the different iterations you've seen of. The other scary thing is just, I didn't understand it enough to really, like I wanted to do it, but I didn't understand it enough to where I was like, how secure is this? And then also, how much is this going to cost? But I would love to be able to talk to my Claude code and everything through Telegram or.

WhatsApp and be able to continue that process.

Peter Brecht23:51

I am waiting for the first security breach and this episode will come out the first week of March. At least I think that's when it's coming out. And I'm waiting for that first security breach to come through and for everyone to be like, okay, we got to like pivot or hard pivot into this sort of security measure for it. Once they figure that out, I'll be a little more comfortable. But I literally considered buying it a phone, a Mac mini and just going all in on it. But I just, I can't, I couldn't commit.

Wade23:57

Okay. Okay.

Wade24:19

I saw some posts of some guys who were kind of going through and kind of scraping a lot of data and information, I guess, and finding all the people who were running like OpenClaw or ClaudeBot or whatever you want to call it and pretty much had keys out there. There was a lot of stuff out there. The bad thing is whenever you do something like that, now I know they're putting it on the Mac mini to try and kind of isolate it from a lot of their stuff, but...

At the same time, if it's got access to accounts and all that, there's going to be ways. And the crazy thing is now though, it's like as much as you want to use like AI and all this stuff to kind of help with our job function, that stuff, hackers and everybody else, they're coming up with some pretty ingenious ways. Like when you embed essentially invisible instructions into an image or a PDF that only your AI can see, you you open that PDF and you see nothing but.

The PDF that your AI is reading is now taking a prompt injection. That's some kind of scary stuff for being able to either do ransomware or whatever. It's getting much more sophisticated.

Peter Brecht25:33

The cyber stuff is frightening. Just sitting on the, and just from the insurance conversation, it's like how, like what are the bad actors able to do now? Gemini had an article today, Google put an article out today that they, somebody was, there was a couple of groups setting up Gemini accounts and then running a hundred thousand prompts through it at once to try to understand the responses so that they could create the pattern and then duplicate Gemini's

Wade25:40

Yeah.

Peter Brecht26:03

Backend code. So there's other ways that these ha like, there's obviously value there. Correct. And, there, there's going to be a lot of that too. Yeah, proceed with caution, right. As with anything, right. But if you not, if you're not trying it and you're not in it, I don't know, I would be willing, if anyone reached out to me, I'd be willing to be like, Hey, this is how you start and give you some guidance on it. Cause it's really is, it is fascinating. And, I think the more people, good people using it.

Wade26:03

Yeah.

Wade26:07

They're like reverse engineering. Yeah.

Peter Brecht26:32

The more resilient it becomes. So all good.

Wade26:32

How are you using Cloud Code? Are you using command line or how are you using it?

Peter Brecht26:39

I'm not that smart. I just talk to it like it's a human, like a designer. So just the reply, go in there. No, I have it installed on my computer. So it all runs there. I run it locally. And my machine is like a...

Wade26:44

Okay. No, are you talking about on the web, like through your web browser?

Wade26:51

Okay, all right. So you're running desktop.

Does it have access to all your files through that one?

Peter Brecht27:01

I haven't set that up yet. I'm not so here's the thing. I have a Windows computer and the only reason that I downloaded so I was using it on the Internet for a bit the reason that I Installed this was because I was supposed to be to get co-work now as a Windows user but I have an arm 64 device and my computer choices were very poor at the time, but I needed a lot more RAM and it was the cheapest way to get it so Anyways, yeah, no, I'm running it locally now, which is amazing

Wade27:03

Okay.

Right.

Wade27:26

Okay, gotcha.

Peter Brecht27:30

There's a lot of things that have evolved. I've and I'm not technically this is the other part is like I'm good at web design SEO like all that so like I've lived in HTML and building that out, but then you start getting and I went through the Replate 100 days of Python so completed that probably a year and a half ago and Obviously now it doesn't really matter, but I know enough about it But I'm still kind of learning different things like yeah, like like I'm trying to

Wade27:40

Okay.

Wade27:47

Okay.

Okay.

Wade27:54

You were getting primed, yeah.

Peter Brecht27:58

Learn the code side more so that maybe I'm a little less in the dark. By no means am I think the level you're at, you're probably a few steps ahead. And I'm just like, I'm still playing with like some cons conceptual things and trying to figure out like, is this a good idea? I don't know. But anyways, we'll, get there.

Wade28:07

Nah, I don't-

Wade28:15

I think if I'm being honest, I think one of my advantages is, and maybe I would, I'm speaking to myself or for myself. I think one of my advantages is, that I don't have a super strong like coding or see whatever MIS, whatever background, because what that's forced me to do is to rely on it more, which guess what? Like it's the subject matter expert, right? All the time I throw it in plan mode and I'm like,

Peter Brecht28:32

Yeah.

Peter Brecht28:38

It's pretty good.

Wade28:43

You just vision for it. Here's what I'm wanting to do. Here's what I'm thinking. Here's the stuff I've built before. What do you think would be the best way to do it? Now, it's gotten so good. They've made the upgrades so good. It's asking you, well, are you comfortable with this? Or what do you think about this? I'm just going back and forth with it before the deployment phase, before the building phase, just kind of helping to lay that groundwork. I think to an extent,

That that may be an advantage because like whenever I, so whenever I had a geosteering group over at Baker Hughes when I first started, right? Like you bring in these new guys and it was kind of funny because sometimes that asked this question that because you have this foundational level of knowledge, like you always just skipped past it. Like you, I had never really thought of that question before and it seems really basic, but it makes me reevaluate so many things that I kind of hold, as true or hold dear, you know?

And so I feel like I'm kind of like that newbie, of course, being able to use cloud code, because I just lean on it so much for guidance, that I think it may be better than somebody who's like, I want to do this, I want to do it this way, and it's just relying on it to just build code, rather than actually architect an entire system. And so I'd like to think sometimes maybe maybe my stupidity is a superpower in this case, you know?

Peter Brecht30:09

Yeah, yeah, yeah. No, I like that. I love that thought too, because you get to, well, I always think through this too and is never waste a good first time, right? So the first time you do a job, you come across pieces of friction that you have never experienced before, right? So it doesn't matter whether you're learning how to clean something different or working on a car or a piece of code or a new job. There's something that's going to come across your plate and you're like,

I'm like, this is just, it's really hard to get through this for me mentally. And what I've done with anything. Like within our company, my, my, my hope is that I get to sit, sit in everyone's seat. Right? So it doesn't matter if you're a new hire, I want to go through the new hiring process, like a new hire, and I want to observe. And then I want to talk to the new hire and be like, Hey, give me feedback on what was good, what was bad. How can we improve that? The reason you do that is very much like what you're saying. You, you're not, you may not.

I'm not necessarily the subject matter expert on the hiring process, but when you uncover those friction points, you can sort of develop a solution for it. So that's like a lot of these things is coming up because I'm okay. I'm doing the account manager job where I have to do paperwork for a week. These are the things I'm coming up with and it's like, okay, I got to solve. Want to, I hate doing this part of the job. I want to automate it. You fix it.

Wade31:15

Right.

Wade31:28

Right.

Peter Brecht31:30

But you're using, you're able to come in as a new person and using Claude is kind of the same thing. Like you kind of have this wild idea of like, here's the problem and then it solves it for you and then you have less friction. So I don't know. There's a lot of stuff in this and it's like, it's very, there's a lot of thoughts that I have been having about it, but anyways, it's exciting nonetheless. So I love seeing what you're doing with it.

Wade31:50

Right. Well, it is, it's exciting. The downside, of course, to all of what I just said was that, whenever, performance of it degrades, so do I. Like I'll tell you, I think there was a very real degradation in Opus 4.5 in the weeks leading up to 4.6 release. Whenever I was using it, it felt like there was a lot more errors. I was having to go back through and, fix a lot of things that it wasn't errors it wasn't making before, and so I really felt like that. It really kind of slapped me in the face my my lack of what would just my lack of ability, whereas like, if you took a computer away from me, like I could still drill a well right because like, I know the math, I know how to calculate volumes, I know I mean, all those, would I be as fast? No, because I got to sit down and write it out and all that.

Peter Brecht32:23

Yeah.

Peter Brecht32:34

Yeah.

Wade32:45

But at the same time, like I could still do it. Whereas with this, man, like I was still moving forward, but it was taking so much longer because my tool just wasn't, it just wasn't as sharp, you know? And then 4.6 hits and I'm like, oh my gosh, like it started iterating on itself. It was doing tons. I wouldn't have to tell it to do error checking. It was going back and checking the changes that made plus the changes it made in comparison to the previous changes to make sure that everything was working together. It's been.

It's been amazing.

Peter Brecht33:17

A lot of the stuff I've built hasn't been dependent on that necessarily, because a lot of what I have is like static code. So it's like, here's a plugin for a computer and a tool to get the job done. So when you use it, it's like, unless the code breaks, it's gonna keep working, right? So my next step, and this is the fun part for me, is going to download all of the plugins that I've bought from other companies.

Wade33:24

OK.

Right. I got you.

Wade33:35

Right.

Wade33:44

Huh.

Peter Brecht33:44

And understanding how they work and then I can build tools that work around those. You continue using. It's it. We've already so just this was last week. I sent a cancellation for a program that we were paying, I think it was three three hundred seventy five dollars a month for which three seventy five doesn't sound like a lot of money, but three seventy five a month for the function it was providing was was worth it. But.

Wade33:46

That's why they say SAS is dead, right?

Peter Brecht34:13

We built a tool that was a hundred times better for our actual use case than what they could do. And that's okay. So you know what? It's going to, it'll continue to evolve. And if I'm wrong and we cancel something and my stuff breaks, it's on me, but we'll get there when we get there. So can't think that way.

Wade34:28

Yep. Well, and I think that's because that's hilarious because that's one of the things I've really been debating lately is like, I've been able to replace pieces of software and stuff. And then it comes down to the end and you're like, but am I like, am I able to fix this if it breaks? Or do I have the time to support this? Like this could be now a part of partially, a time sync every week is like, if a user is having a problem or if this corner case kind of crashes it or whatever. And so

Peter Brecht34:48

Yeah.

Wade34:56

You've continually got that. But the other thing I really wondered when I first started doing that was like, what? So I think back in the day, two years ago, right, like your patent really was your code, right? Like people couldn't like essentially just rip off your code and do the exact same thing. But now, like you're not ripping off the code, you're using AI to write code, you're you could be ripping off the idea, but your code is now unique. It's nothing it's probably nothing like

Peter Brecht35:18

Yeah.

Wade35:25

What they originally wrote in that software. And so my question was like, is there a gray area here? And where does that kind of where does that line fall? Because like you're talking about those people who reverse and are trying to reverse engineer Google's LLM, like where is the where's the line and being able to do something like that?

Peter Brecht35:35

Hmm.

Peter Brecht35:46

Well, if you read a lot of like SaaS, because I've had to do this for years, just like, cause I, we've all have, I think you have to read a SaaS contract. And a lot of times in that contract, it will say that you cannot, reverse engineer our code or develop something that is similar in nature. But I've always looked at that and I've said, okay, first of all, I'm not smart enough to get to your code. I'm an end user who's going to use your platform.

Wade35:54

Uh-huh.

Wade36:00

Right. Right.

Peter Brecht36:13

In this particular case, it was a forums tool. So basically, how does a client, how do we in, how do we have a client on board with us so that we can move to the next step? And the only thing that was different about it, which was odd was already done with probably 15 other software tools. I just decided that I'm tired of paying 3.75 a month for it when I can just host it for $25 and call it a day. So I didn't rip anything off. Additionally,

Wade36:42

Right.

Peter Brecht36:42

I was like, there's five things that I've asked you guys to do and I've put this into your requests. Nothing's happened over the past three years since we started using it. And I just decided I'm just going to build it all now. So it, and it, and it built it. So there's stuff like that. That just kind of makes me, I don't know, man. I'm excited. I'm bullish. It's cool. So yeah, no, no, no. Yeah. Yeah.

Wade36:49

Yeah. Yep.

Wade36:58

Well, I'm not saying it's wrong. Do not read into that. I'm just saying what used to protect that company, right? Or that idea, whatever, was the execution, was the code. And now that's gone. It feels like the Wild West.

Peter Brecht37:09

The well, moat, it was their moat was the ability code. Yeah, it does. And, yeah, we'll see where this goes. I'm excited. We'll have another conversation about this. Let's do it again in 30 days. So the other fun part about all this, I'm looking at our production schedule and I've been thinking about the podcast, like where all this is going for us. We're, only a thousand followers. So it's like, it's, there's a few people hearing us. It doesn't matter, but I'm enjoying the interviews.

We're coming up on our one year of recording on it's gonna be yeah, we're coming up on our one year It's not as far out because we're recording we record in advance. So within next month. We'll be recording that episode But yeah beginning of last last week of April is when we started so we're coming up on our year I don't know what to do something fun for it. But anyways, I think we'll cut it off here We've been rambling for 38 minutes, and I don't know how many people are gonna make it past the first eight, but Thank you

Wade37:42

Really?

Wade37:49

Now that's yeah.

Wade37:55

Okay.

Wade38:07

This might be a different demographic than normal.

Peter Brecht38:10

Yeah, it might be different, but for everyone in oil and gas, thank you guys for being here. We're thankful to have the few of you who do list on a regular basis. We appreciate your support. Wade, I'll see you again next week.

Wade38:24

Good talk to you, Peter.