WEBVTT
00:00:01.600 --> 00:00:05.839
Welcome back everyone to another episode of Dynamics Corner.
00:00:06.000 --> 00:00:07.440
What is MCP?
00:00:07.519 --> 00:00:09.839
Is MCP AI agent?
00:00:10.000 --> 00:00:10.480
I don't know.
00:00:10.560 --> 00:00:11.199
We'll find out.
00:00:11.279 --> 00:00:12.720
I'm your co-host, Chris.
00:00:12.960 --> 00:00:13.919
And this is Brad.
00:00:14.000 --> 00:00:17.120
This episode was recorded on February 16th, 2026.
00:00:17.359 --> 00:00:18.719
Chris, Chris, Chris.
00:00:19.039 --> 00:00:22.079
MCP, AI, open source, and more.
00:00:22.239 --> 00:00:25.120
Today we had a full conversation with Tobias Fenster.
00:00:51.119 --> 00:01:01.200
Like anybody who doesn't embrace it, yeah, it's not perfect, but if you don't embrace it, you're um you're going to be in for a long ride.
00:01:02.240 --> 00:01:03.359
A rude awakening.
00:01:04.319 --> 00:01:05.280
Good afternoon, sir.
00:01:05.359 --> 00:01:05.840
How are you doing?
00:01:06.000 --> 00:01:07.040
Good day, Tobias.
00:01:07.920 --> 00:01:09.040
Very good, very good.
00:01:09.120 --> 00:01:09.920
How are you?
00:01:10.239 --> 00:01:11.040
Very well, very well.
00:01:13.040 --> 00:01:16.159
Chris and I were just talking about this new thing called AI.
00:01:16.319 --> 00:01:19.040
Uh while we were waiting for the podcast to start.
00:01:23.040 --> 00:01:29.359
We were just sharing stories about this weekend, and um, you know, uh, I'm trying to build a website.
00:01:29.439 --> 00:01:35.359
I'm not a web developer, never, I mean, maybe back in the day with front page or whatever.
00:01:35.519 --> 00:01:39.040
Um, and I just like, hey, I just need to build a website, a front page.
00:01:42.560 --> 00:01:45.120
These days it's fun tour and not good news, right?
00:01:45.439 --> 00:01:45.760
Exactly.
00:01:47.120 --> 00:01:47.760
Exactly.
00:01:48.000 --> 00:01:59.840
So it took me a couple hours to build a full site, a full web stack of like where it's hosted, the the different languages, TypeScript, you know, next dot JS and all that stuff.
00:01:59.920 --> 00:02:02.319
And I'm like, I would have never known any of those.
00:02:02.560 --> 00:02:08.960
But uh yeah, took me maybe two hours to have a working website and I can host it.
00:02:09.199 --> 00:02:10.159
It's awesome.
00:02:10.400 --> 00:02:11.280
That's crazy.
00:02:11.919 --> 00:02:26.240
It's it it is, and it's uh we were also saying that if um if you're an organization or a person that uh hasn't adopted in some way, then you're a little behind, uh little behind, I think.
00:02:26.479 --> 00:02:37.039
And it's not just coding anymore, it's um and that's sort of where I was talking with Chris before some of the stuff is coding is just a small aspect of where it could be utilized.
00:02:37.199 --> 00:02:43.199
If you're looking at documentation, you're looking at project management, you're looking at a number of uh different things.
00:02:43.360 --> 00:02:46.159
So it's uh it's it's the buzz for the day.
00:02:46.319 --> 00:02:48.719
Uh unfortunately, you can't escape it.
00:02:48.879 --> 00:02:52.800
But um uh we do appreciate taking the time to speak with us.
00:02:52.879 --> 00:02:54.639
Uh, I have a lot of things we wanted to speak with you out.
00:02:54.719 --> 00:02:58.879
You do many great things, but uh before we get into that conversation, you might tell us a little bit about yourself.
00:02:59.759 --> 00:03:01.759
Yeah, so my name is Tobias Fenster.
00:03:01.840 --> 00:03:18.240
I'm uh currently one of the co-founders and uh managing director at 4PS in Germany, um, but also working as what is called a chief uh engineer at the Hilti organization, because uh 4PS by now is part of Hilti, and uh that led to that uh little career change.
00:03:18.319 --> 00:03:21.520
So I'm basically spending my time 50-50 in between.
00:03:21.680 --> 00:03:28.479
Um I'm also a Microsoft MVP for Azure for Business Central, um a regional director, and a Docker captain.
00:03:28.560 --> 00:03:35.039
So also active in the community thanks to my blog and podcast and a number of things like uh conference speaking, et cetera, that I do.
00:03:35.680 --> 00:03:41.520
Yes, and uh uh all on that background, there's a many different things that uh we wanted to catch up with you on.
00:03:41.759 --> 00:03:44.479
Uh one uh being your podcast.
00:03:44.639 --> 00:03:45.520
You have a podcast.
00:03:45.680 --> 00:03:47.919
Um and there's a lot to podcasts.
00:03:48.000 --> 00:03:57.439
I see a lot of individuals start podcasts, I see a lot of individuals have podcasts, and I see a lot of individuals with uh uh podcasts that run for a short time.
00:03:57.520 --> 00:03:59.599
So uh tell us a little bit about your podcast.
00:04:00.319 --> 00:04:08.319
So I I have to start basically at a non-tech podcast I also do, um, and that is uh for the local sports.
00:04:09.439 --> 00:04:15.039
That's that's the the local sports uh at the local basketball club is playing at the highest German level.
00:04:15.199 --> 00:04:29.920
And um a while ago, someone who runs a sports podcast on the internet called Sport Radio 360, a German one, um, asked for people who wanted to, you know, talk to a number of people around that and and just be there, uh get some some statements after the games.
00:04:30.000 --> 00:04:35.199
And I've never done anything like that before, but I've always been into basketball for my local team.
00:04:35.360 --> 00:04:37.279
Um so I said, well, let's give it a try.
00:04:37.360 --> 00:04:39.040
And that was, I think, in 2016.
00:04:39.120 --> 00:04:41.680
So that got me started on the whole podcasting thing.
00:04:41.839 --> 00:04:53.519
And that has no connection at all to technology, but I realized actually I I enjoy doing the podcast, but what I enjoy most is talking to people who are professionals in basketball and just being able to ask them questions.
00:04:53.600 --> 00:04:55.680
I would never be able to ask them otherwise.
00:04:55.920 --> 00:05:04.319
And that then triggered the idea for my tech um window on technology podcast, which is um basically the same idea.
00:05:04.480 --> 00:05:14.000
I I have people I would like to talk to, I have topics I want to talk about, and the podcast just gives me a very good excuse to ask them questions I I otherwise never would ask them.
00:05:14.160 --> 00:05:27.680
Like, you know, experts in the field, uh being from Microsoft from the community, whatever, where I probably wouldn't have the time or wouldn't make the time, and they certainly wouldn't to have a conversation for 30 minutes with me about a random topic, so to speak.
00:05:27.839 --> 00:05:30.480
And and that's basically how I run that podcast.
00:05:31.439 --> 00:05:34.639
I smile a little bit because there's there's a lot in there.
00:05:34.800 --> 00:05:41.360
One, I'm sure it's refreshing to be able to talk to people about something other than what you do for on a day-to-day basis.
00:05:41.680 --> 00:05:42.000
That's true.
00:05:42.319 --> 00:05:46.639
To step away from uh the technology and still to be able to talk to people on the podcast.
00:05:46.800 --> 00:05:54.160
And I think the other point that you hit, I think is something Chris and I have spoken about before, and it's kind of why we do this.
00:05:54.240 --> 00:05:58.720
And even when we talk with individuals, I think it mentioned with you as well in the pre-podcast call.
00:05:58.879 --> 00:06:16.079
I have a conversation as if even nobody were listening, but I do appreciate the opportunity to talk to many of the individuals that we speak to, like you said, because you get to have an individual conversation with them, even though it's recorded for many people to listen to, you get to ask questions and have a one-on-one conversation with many people.
00:06:16.160 --> 00:06:20.160
And it's um fascinated and appreciative of that with everybody.
00:06:20.319 --> 00:06:29.279
Um I say that because it's it's um I've learned so much by being able to ask those questions uh and met many individuals as well.
00:06:29.839 --> 00:06:34.079
So no, you do the I I didn't know about the basketball podcast, but I do listen to the tech podcast.
00:06:34.560 --> 00:06:34.959
That's awesome.
00:06:35.120 --> 00:06:43.920
Um, the one I caught up with recently, I think, was with um Jeremy uh Visco on there, um, talking about uh some of his projects and stuff too.
00:06:44.000 --> 00:06:46.240
So and Andy Wingate, I think that was the last one.
00:06:46.480 --> 00:06:48.560
Yeah, or maybe the second to last one.
00:06:48.639 --> 00:06:53.199
I don't I'm not certain of the ones, but that's the one that uh I listened to the that's the storytelling one.
00:06:53.360 --> 00:06:55.600
Yeah, that's the one that Jeremy was saying.
00:06:55.680 --> 00:06:56.319
Yeah, that's good.
00:06:56.720 --> 00:06:58.079
No, that's uh no, it's good.
00:06:58.240 --> 00:06:59.920
Jeremy's doing some great things as well.
00:07:00.000 --> 00:07:01.360
So absolutely, yeah.
00:07:01.600 --> 00:07:04.560
Um, so we do appreciate you taking the time to speak with us.
00:07:04.720 --> 00:07:10.399
Now, yeah, you had mentioned that you do a lot of conference presentations, which is another topic that we talked about talking about.
00:07:10.480 --> 00:07:13.920
So let's let's dive into conference presentations a little bit here.
00:07:14.000 --> 00:07:17.120
Uh a little bit away from the tech of actually like coding and developing.
00:07:17.199 --> 00:07:18.720
Let's talk about conferences.
00:07:19.040 --> 00:07:19.680
Yeah.
00:07:20.240 --> 00:07:30.079
So maybe what got me into that or what I like about it is um basically very early in my career, I self-defined as a very good developer, I would say.
00:07:30.240 --> 00:07:37.600
Um, probably a bit over ambitious because in my first few years I wasn't so such a great developer, but still that was my how I saw myself.
00:07:37.759 --> 00:07:40.480
And then I had a very nice uh boss at that time.
00:07:40.639 --> 00:07:43.439
And what she told me was, look, you're good at development.
00:07:43.600 --> 00:07:47.519
She didn't want to say me, I'm mediocre or I don't know, but um, you're good at development.
00:07:47.600 --> 00:07:52.639
But what actually makes you stand out is that you can still talk to business people in a way that they can understand.
00:07:52.800 --> 00:07:54.800
So she she gave me that feedback.
00:07:54.959 --> 00:08:02.959
Look, it's you're good at tech, but basically the thing that that maybe helps you differentiate along your career is that you can also talk about it.
00:08:03.120 --> 00:08:07.279
And that gave me that idea to, you know, first of all, go into consulting.
00:08:07.360 --> 00:08:15.360
So I went basically from a developer to a development consultant when I taught customers and partners how to use our products or the products that we sold.
00:08:15.519 --> 00:08:24.720
And then it also was a bit of a segue into the whole conference speaking because I felt like, okay, if I can do this for customers, I can maybe do it on a wider for a wider audience.
00:08:24.800 --> 00:08:29.279
And I still really very much like to do that for many two reasons.
00:08:29.439 --> 00:08:30.399
One, it's just fun.
00:08:30.560 --> 00:08:32.080
I I just enjoy doing it.
00:08:32.240 --> 00:08:41.840
I mean, that doesn't stop me from being quite nervous when I step onto whatever kind of an audience that is, or um whatever kind of stage that is I step on.
00:08:42.080 --> 00:08:45.679
I'm always kind of nervous, but it is still very much fun.
00:08:45.840 --> 00:08:56.000
And the other thing is it sometimes helps me to focus on a specific topic or drive a specific topic that I always wanted to learn about, that I wanted to understand about, dive deeper into it.
00:08:56.159 --> 00:09:00.080
And sometimes I I just hand in a session description with an idea.
00:09:00.159 --> 00:09:02.720
I'm I'm pretty sure that it should work, not 100%.
00:09:03.120 --> 00:09:16.960
I I typically let the organizers know in advance that this might might not work out in the end, but it gives me an excuse and helps me to focus on, okay, this is something I really want to understand, I want to prepare, and then um be able to share it with others.
00:09:17.120 --> 00:09:20.639
So those are for me the two aspects of speaking.
00:09:20.799 --> 00:09:32.159
It's fun, I like to do it, I think I'm okay at it, and it helps me to focus on specific topics I have on my agenda and might otherwise not find the time or make the time to really dive into.
00:09:32.799 --> 00:09:41.679
Yeah, you you hit on some points there where if there's nothing like having to present a topic uh to force you to really understand it.
00:09:41.759 --> 00:10:00.879
So it is, it's almost like we talked about with the podcast uh being selected uh by the committee to present a topic at a conference also gives you the opportunity or forces you for the opportunity to learn, you know, have to really submerge yourself or immerse yourself into that topic to understand that topic.
00:10:01.120 --> 00:10:05.360
And you also um uh mentioned a couple of things with the nervous.
00:10:05.600 --> 00:10:16.960
I I literally said that to someone the other day because over here in the United States, uh, we have uh in April and May, you have the Directions North America Conference, and then you have the DynamicsCon conference in May.
00:10:17.120 --> 00:10:22.720
And they both started sending out, I think they both started sending out one of them started sending out, but I don't I can't keep track.
00:10:23.120 --> 00:10:25.279
Directions started sending out the directions, yeah.
00:10:25.519 --> 00:10:27.919
Directions, and I think Problems come be sending this out soon.
00:10:28.000 --> 00:10:31.039
They started sending out their approvals, and that's what I was talking with somebody about.
00:10:31.360 --> 00:10:37.200
Is it doesn't matter how often you do it or how many times you do it, you still get nervous.
00:10:38.000 --> 00:10:41.440
And you get after you get selected, you get a little nervous about the topic.
00:10:41.519 --> 00:10:57.360
Even if it's a topic that you could literally walk into a room and talk about freely, you still get nervous because I know personally I have to do the presentation, I have to make sure the delivery, you have to make sure, and it's it's and it's a little bit different than if you and I were just sitting down with the same topic.
00:10:57.440 --> 00:11:01.120
So you ever feel when you do those things, you feel like the imposter syndrome.
00:11:01.200 --> 00:11:10.879
Like, I certainly have gone through that where like I know the topic, but then like everyone looks at you and it's like, man, did I say what I say something wrong, or was I off?
00:11:11.120 --> 00:11:15.519
You get that it was like maybe I don't belong in this stage, I'm not sure what I'm doing here.
00:11:15.679 --> 00:11:17.039
And then it just gets in your head.
00:11:17.120 --> 00:11:22.080
But yes, it is uh nerve-wracking when you get chosen to go ahead and present.
00:11:23.120 --> 00:11:30.159
And then even afterwards, when you get the feedback, I don't know how it is for you, but there's the rational part of me that knows you can't do it right to everyone.
00:11:30.240 --> 00:11:31.519
Not not everyone will like it.
00:11:31.679 --> 00:11:33.360
That's just a numbers game, right?
00:11:33.519 --> 00:11:48.879
But still, I I can still tell you for I think the second or third conference talk I had, it was very well received, like I don't know, 20 positive feedbacks, but I I still know there was one person who said, um, seemed very well prepared, but also boring.
00:11:49.039 --> 00:11:51.200
I I still remember the exact words, right?
00:11:51.360 --> 00:11:56.399
So that's you know, I I could have focused on the 20 that were positive, but that one stuck in my mind.
00:11:56.480 --> 00:11:58.639
And that's also part of it why I'm still nervous.
00:11:59.039 --> 00:12:00.480
Is it gonna be a boring one again?
00:12:00.639 --> 00:12:01.200
I don't know.
00:12:03.840 --> 00:12:04.879
I don't mean to laugh.
00:12:04.960 --> 00:12:05.840
No, that's a little rough.
00:12:07.200 --> 00:12:09.200
I've gotten those before.
00:12:09.919 --> 00:12:12.240
Yeah, I mean, you get a lot of different comments.
00:12:12.399 --> 00:12:18.080
I mean, I I do value the feedback, I do value the feedback that's constructive as well.
00:12:18.240 --> 00:12:34.639
Um, because again, anybody who attends a presentation should know that I haven't met anybody who does present that doesn't want to honestly hear uh ways that they could change the presentation or if the pres or what they're doing is being done well.
00:12:34.799 --> 00:12:38.240
I mean, not even saying they have to change it, but I always uh say the same thing.
00:12:38.320 --> 00:12:47.120
Uh, you know, give me the feedback because it helps make sure that future presentations uh may be more um uh favorable to some.
00:12:47.200 --> 00:12:48.399
But like you said, you can't please everybody.
00:12:48.480 --> 00:12:50.480
I mean, there's nine billion people on this planet.
00:12:50.639 --> 00:12:58.799
Uh the chances are that there's a few of them that aren't going to be uh in agreement with uh your presentation uh what you have.
00:12:58.960 --> 00:12:59.759
So that's good.
00:13:00.000 --> 00:13:03.440
Um do you have any lined up this year already to present?
00:13:03.519 --> 00:13:06.480
Or they haven't released the ones in the European conferences at this point.
00:13:06.799 --> 00:13:07.360
Yeah, indeed.
00:13:07.440 --> 00:13:09.440
So I I have submitted to a number of them.
00:13:09.600 --> 00:13:11.600
I have gotten accepted to a few.
00:13:11.679 --> 00:13:27.039
So early in May, I'll be at one that's called the ADC Advanced Developers Conference, um, and then one that's the uh BIS app summit in in Cologne, where I also talk about Business Central with the MCP server, um, and at Dynamics Minds in in Slovenia.
00:13:27.279 --> 00:13:30.720
So those are the three that I already know, and I've I've submitted to a few others.
00:13:30.799 --> 00:13:32.399
So let's see what happens.
00:13:32.879 --> 00:13:34.159
Ah, that's interesting.
00:13:34.320 --> 00:13:41.919
So speaking of development and business central, um, I do want to jump into a whole open source conversation with you as well once we get there.
00:13:42.000 --> 00:13:44.320
But you mentioned uh MCP.
00:13:44.639 --> 00:13:49.759
Um you're doing a presentation on the business central MCP server.
00:13:50.000 --> 00:13:50.240
Yes.
00:13:50.720 --> 00:13:54.240
In your words, what is the business central MCP server?
00:13:54.399 --> 00:13:56.000
Well, business central MCP.
00:13:56.080 --> 00:13:57.120
Is it even a server?
00:13:57.360 --> 00:13:58.799
Let's kind of break it down.
00:14:00.000 --> 00:14:00.639
Yeah, indeed.
00:14:00.879 --> 00:14:14.399
And not certainly not a physical server, but but MCP is kind of the standard that allows um those LLMs, those models to talk to backends that that had to be manually coded or very individually done in the past.
00:14:14.720 --> 00:14:19.039
And now there's this protocol called MCP that is kind of a generic adapter.
00:14:19.279 --> 00:14:30.559
They say it's USB for AI in the sense that you suddenly no longer worry about what is it um what's the the Apple I Lightning or whatever, or is it USB A or whatever?
00:14:30.639 --> 00:14:33.200
It's it's just USB-C nowadays for a lot of those.
00:14:33.840 --> 00:14:33.919
Yeah.
00:14:34.159 --> 00:14:38.399
And it and that's um that's kind of the thing the idea for MCP.
00:14:38.639 --> 00:14:43.919
You can plug in anything, it's the same standard, and now suddenly your model can talk to any backend.
00:14:44.240 --> 00:15:02.480
And the BC MCP is um that one that Microsoft created and and made available mostly or most easy uh through the Copilot Studio, but also possible um to use from other clients where you can just ask a question like, what are my top 10 customers in my business central instance?
00:15:02.639 --> 00:15:10.960
So you can start talking to your BC system, to your BC environment, and get meaningful business uh uh answers and have a conversation.
00:15:12.240 --> 00:15:19.759
That so I think the MCP came out version 27, 27.1, 27.2, around that time.
00:15:19.919 --> 00:15:30.720
Um I it's so hard for me to keep up with these versions because I know now the agent preview came out with 27.4, or that's the new name Agent Preview uh from previously what many knew is the agent playground.
00:15:30.960 --> 00:15:34.240
And I think MCP was at 20 that again is a version 27.
00:15:34.320 --> 00:15:35.440
Uh you need as well.
00:15:35.759 --> 00:15:40.000
I also don't know because um they they do stuff, then they do stuff in preview.
00:15:40.080 --> 00:15:44.159
Um sometimes it's private preview that we see as MVPs, right?
00:15:44.320 --> 00:15:48.080
So it's difficult to keep track of what happens when correct.
00:15:48.159 --> 00:15:49.360
But I do know it's in the product.
00:15:49.679 --> 00:15:52.000
I do know it's in the product now, so we can talk about it.
00:15:52.159 --> 00:15:57.519
So that's uh as at the time of this recording, it's in the product um that someone can install.
00:15:57.840 --> 00:16:12.080
So MCP allows other agents, models, or other AI, I guess we could say, you know, we'll just generalize with AI, be able to communicate with a business central implementation.
00:16:12.399 --> 00:16:12.720
Yep.
00:16:12.960 --> 00:16:15.360
Can it just read the data or can it write the data?
00:16:15.840 --> 00:16:17.840
That's up to how you want to configure it.
00:16:17.919 --> 00:16:23.039
So to put that up front, it's not like everything's open and it's free to do whatever it wants.
00:16:23.120 --> 00:16:25.039
Um, but you can configure what you want.
00:16:25.200 --> 00:16:32.480
Uh, for example, you can say this is only a read-only, or you can say uh only want to expose my customers and my vendors, or something like this.
00:16:32.639 --> 00:16:42.559
So there's a big configuration that that um keeps you in control of that, but you can also enable it for write access, and then it can also create stuff, delete stuff, modify stuff.
00:16:43.600 --> 00:16:55.759
So with the MCP configuration, you can determine which data is available, and then you can determine which data is available for reading and which data is available for writing.
00:16:55.840 --> 00:17:07.119
So theoretically, I could take a request of show me all my customers in their balance and the ones that have late payments or overdue, excuse me.
00:17:07.279 --> 00:17:14.880
And then I could also write back and say, update and block all the customers that have an overdue balance uh unpaid.
00:17:15.279 --> 00:17:24.240
That's pretty impressive because now we've gone from having to run, it's almost having to go from, and I was thinking about this over the weekend about something else that I've been working on.
00:17:24.480 --> 00:17:38.880
We've gone from having to run reports to see this information to take an action to just saying update my uh I was playing with the the age and preview, you know, find the customers again with that same scenario with the overdue balance and block them.
00:17:39.119 --> 00:17:49.119
And now you can do that, you know, with external systems, so you can do a different analysis because maybe you can uh do additional analysis, you know, with another MCP, with another product for that customer.
00:17:49.200 --> 00:17:50.000
That's pretty well.
00:17:50.240 --> 00:17:54.960
So and and so sorry to take that one step further, just just to mention it.
00:17:55.119 --> 00:17:58.079
What's also then interesting, it's it's not only BC, right?
00:17:58.160 --> 00:18:01.039
You you can also now integrate with other systems.
00:18:01.200 --> 00:18:15.359
I mean, if you have another um tool in your company that you use for, I don't know, time tracking or whatever, you could also say pull the time tracking data out of system A and plug it into my BC because um there I want to create an invoice.
00:18:15.519 --> 00:18:22.079
So it's it's not only that you talk to one backend, but you can also talk to multiple backends if configured in the right way.
00:18:22.319 --> 00:18:24.880
Does it increase the chances that everything works perfectly?
00:18:25.039 --> 00:18:26.079
I mean, obviously not.
00:18:26.160 --> 00:18:30.240
You you will still have to um work on it and understand what it does.
00:18:30.559 --> 00:18:47.759
But the the possibilities become a lot more interesting now suddenly that you're not only reading data and it's not only one system, but you can, as you mentioned, also write data and you can have it across um the whole scope of whatever system you you're using if they have um some kind of an MCP support.
00:18:49.200 --> 00:18:55.039
So with so business central has an MCP so that other systems can connect to it.
00:18:55.200 --> 00:18:58.960
Can Business Central natively connect to other MCP servers?
00:18:59.680 --> 00:19:01.599
No, no, at least not that I know of.
00:19:01.759 --> 00:19:11.599
Um that one that kind of orchestration would be on the AI side, so something like Copilot Studio or other tools that you might want to use, they would talk to those different MCP servers.
00:19:11.680 --> 00:19:13.119
And that's kind of the idea.
00:19:13.200 --> 00:19:27.119
I would say that you no longer create those point-to-point connections, but instead you have a generic way how you can connect to it, and then you have um those orchestrators or however you want to call them that are able to talk to the different environments and the different tools.
00:19:27.680 --> 00:19:28.319
Okay, perfect.
00:19:28.400 --> 00:19:35.440
So you have just uh I'm I always like to repeat what I hear sometimes just to uh you know put it all together, right?
00:19:35.680 --> 00:19:52.799
So you'll have so BC has access available to it, other systems will have access available to it via MCP, but then you need some orchestrator that will determine which systems, uh, which systems to communicate with, which systems to talk with, to read, write data, and perform different actions.
00:19:52.960 --> 00:19:54.400
That's wow.
00:19:55.519 --> 00:19:58.319
They did a handshake between two orchestrators.
00:19:58.720 --> 00:20:02.720
It's um I just feel like I'm in the Jetsons.
00:20:02.880 --> 00:20:08.960
You remember the Jetsons when I was like the like it's it's almost all the stuff that you saw growing up.
00:20:09.039 --> 00:20:14.400
It's almost like even the cartoons from the 50s and 60s uh are predicting the future.
00:20:14.559 --> 00:20:19.119
Like who would have thought it's that uh we would be here today with some of the stuff that I'm seeing?
00:20:19.200 --> 00:20:27.440
And it's it's um I almost want to go back to just riding my bicycle and having a dime and going to the can't penny candy store um to do that.
00:20:27.759 --> 00:20:30.559
Or arcade, man, like you know, penny arcades.
00:20:31.279 --> 00:20:31.599
Yeah.
00:20:32.000 --> 00:20:43.359
So with the MCP, so you can create your own MCP within Business Central um to determine which aspects of the system you want to expose.
00:20:43.920 --> 00:20:51.599
So it's uh the the MCP in itself is a general component, but you can configure it to make sure you're not exposing something that you don't want to be exposed.
00:20:52.799 --> 00:20:56.720
And it handles authentication and filtering.
00:20:56.960 --> 00:21:15.519
So if I let's just say have customer data and I have salespeople, so uh would it limit, would it have the bounds, boundaries so that only certain pieces of data can be consumed by whomever's accessing it, or do you have sort of one open access point with the orchestrator?
00:21:16.000 --> 00:21:20.640
So what you what you typically would do is that you have you could have multiple configurations.
00:21:20.799 --> 00:21:31.200
One would be the sales configuration, one would be the I don't know, um technical um configuration and the admin configuration and the finance configuration or whatever.
00:21:31.359 --> 00:21:37.359
So you could do that first, but also the MCP is typically connecting using my account.
00:21:37.599 --> 00:21:48.079
So I'm I'm running the client, I'm telling I want to do a connection to Business Central, and that one would then prompt me for my credentials that would be used to log into Business Central.
00:21:48.240 --> 00:21:51.039
So that way, whatever I can do, the agent can do.
00:21:51.200 --> 00:22:03.039
Um, but if my permissions are properly set up and I'm not just the super admin or whatever, which you definitely shouldn't do, as we all know, um then The agent also can only do what I'm doing.
00:22:04.640 --> 00:22:05.839
The good old super.
00:22:06.079 --> 00:22:07.440
I I was working on something.
00:22:07.599 --> 00:22:09.440
I think it was setting up API access.
00:22:09.519 --> 00:22:14.880
And I think it actually prevents when you're doing API setup, it prevents you from giving the account.
00:22:14.960 --> 00:22:16.160
It gets created super.
00:22:17.200 --> 00:22:22.559
Which I hope that they do with MCP as well, so that you don't inadvertently open uh yourself up.
00:22:23.839 --> 00:22:37.599
You had mentioned you had mentioned also, um, I want to get into more maybe the technical construct of MCP, is is how you say you configure and you build it in a minute, but you also one thing you had just mentioned was um you could set it up where you have multiple ones, is what I heard.
00:22:37.680 --> 00:22:41.599
You had mentioned sales uh in another uh area.
00:22:42.720 --> 00:22:48.640
So, what would be a good strategy for setting up the MCP servers if you have multiple ones available?
00:22:48.880 --> 00:22:53.039
Um as far as the uh why you would do that and how you would do that?
00:22:53.599 --> 00:22:55.519
I I think there's two aspects to it.
00:22:55.680 --> 00:23:06.559
One is the the configuration, that those different configurations that you could have, they could limit the access that the agents then have because agents are getting quite confused when they have too much access.
00:23:06.799 --> 00:23:12.000
There's also something that's called dynamic tool um whatever.
00:23:12.319 --> 00:23:16.160
They can dynamic tool discovery or something like this, that that helps with that.
00:23:16.400 --> 00:23:23.279
But generally speaking, the more you can limit the tools, the higher your chances are that the agent is doing the thing that you want it to do.
00:23:23.440 --> 00:23:27.599
So from that perspective, it could make sense to limit those tools that are available.
00:23:27.759 --> 00:23:39.599
So have a limited scope of your configuration, and then specifically talk to that configuration when you know, okay, now I want to do something that's sales oriented, or now I want to do something that's purchase or finance or whatever.
00:23:39.839 --> 00:23:42.400
So you could set up those different configurations.
00:23:42.559 --> 00:23:52.640
It's still kind of the same MCP server, but when you configure that um that connection, then you would say, okay, now I want to use this uh configuration, I want to use that configuration.
00:23:52.799 --> 00:23:58.799
And that would help you to limit the scope, and that increases the chances that the agent is doing what you want it to do.
00:23:59.119 --> 00:24:01.599
At the same time, of course, it also limits the scope.
00:24:01.680 --> 00:24:08.960
So if you think about that super orchestrator agent that can do um a lot of things, then it would need access to a lot of tools.
00:24:09.119 --> 00:24:24.799
But um generally speaking, it would allow you to limit those um that access scope basically and make sure that the agent uh only sees the things that you wanted to see, not only from a permission perspective, but also from an um what is what is even visible perspective.
00:24:25.279 --> 00:24:26.799
It's uh that's a best practice.
00:24:27.039 --> 00:24:36.079
I think I've I've been hearing that quite a bit when you are building these um MCPs and having it, you know, do certain things.
00:24:36.160 --> 00:24:44.000
You have to be very specific to the intent of that, of the purpose of that uh, you know, uh MCP you're building so that it you're right.
00:24:44.079 --> 00:24:45.200
So it doesn't wander off.
00:24:45.279 --> 00:24:50.240
You're it's doing exactly what you need that we need it to do every single time.
00:24:50.480 --> 00:25:06.480
And eventually, as you had you know uh mentioned about having an orchestrator uh to also help and guide that our orchestrator as well to know when to pull which agent or MCP that he needs to communicate for a specific specific purpose.
00:25:06.720 --> 00:25:22.240
So I I think that's a a good starting point, at least understanding that before going into building uh an MCP server, and then and then you've you wonder why it's not working according to what you're trying to do because it's too many, too many options.
00:25:22.400 --> 00:25:23.920
You have to narrow it down.
00:25:24.240 --> 00:25:25.119
Yeah, indeed.
00:25:25.200 --> 00:25:31.039
And and I have to say, um it's probably similar to a lot of people who are nowadays talking about AI.
00:25:31.119 --> 00:25:36.079
It's not like I have 10 years of experience because I mean, literally, this is really very, very new stuff.
00:25:36.240 --> 00:25:51.359
So I'm always cringe a bit when I hear, and that's not meant as criticism towards you, but if you say something like best practice, that a year ago, if someone had said best practice, then I would have said, okay, that implies uh five years and a hundred projects, and now you can talk about a best practice.
00:25:51.440 --> 00:25:51.599
Yeah.
00:25:51.759 --> 00:25:56.480
Now suddenly someone who has opened it and done three projects is the expert and is sharing best practices.
00:25:56.559 --> 00:25:58.799
So and I fully understand there's no other way.
00:25:58.880 --> 00:26:00.960
That was just why I'm I'm reacting a bit.
00:26:01.039 --> 00:26:02.799
That it's it's moving so fast.
00:26:02.960 --> 00:26:11.839
What used to be a best practice, quote unquote last month, suddenly now there's this new super tool that can solve this automatically and it's no longer required.
00:26:12.000 --> 00:26:15.119
And that's time we limited.
00:26:15.680 --> 00:26:24.720
I've always been uh uh a little apprehensive every time I heard the word geez, every time I hear the word best practice, because I say who determines what the best practice is?
00:26:24.880 --> 00:26:34.799
Because oftentimes, in my opinion, uh again, it depends on what you're talking about, but sometimes best practices is also um subjective.
00:26:35.039 --> 00:26:35.680
Subjective, yeah.
00:26:35.759 --> 00:26:38.079
I was just gonna say subjective to the person you're speaking with.
00:26:38.240 --> 00:26:40.079
And you're talking about last month.
00:26:40.400 --> 00:26:43.680
Last month is years in AI life.
00:26:43.759 --> 00:26:49.119
Uh you know, it's it's what was we use the word again, best practice.
00:26:49.279 --> 00:26:50.960
What was common last week?
00:26:51.119 --> 00:26:54.240
You wake up on Monday morning is no longer common.
00:26:54.480 --> 00:26:58.480
It's um it's it's uh it's it's a way to do things.
00:26:58.640 --> 00:27:08.400
Uh so when you say you do just a small comment on the best practice, because that that is also something I learned from because you mentioned the directions conferences in North America.
00:27:08.559 --> 00:27:09.440
I was there once.
00:27:09.519 --> 00:27:16.960
Um, one of actually my my first couple of um appearances at a conference was at the Directions North America.
00:27:17.200 --> 00:27:25.839
And what I learned there is what I considered good practice or interesting or whatever suddenly was no longer interesting because the audience was so different, right?
00:27:25.920 --> 00:27:26.640
It was yes, yeah.
00:27:26.960 --> 00:27:40.240
Directions in here, and then you have directions North America, and I talked to developers at the same time, but they had completely different backgrounds, they had completely different paths, the setup was so different, and I felt like, am I stupid?
00:27:40.400 --> 00:27:41.039
Are you stupid?
00:27:41.359 --> 00:27:46.640
You know, I try to not think about other people as stupid, but I was like, there was a disconnect.
00:27:46.720 --> 00:27:53.279
Let me put it like this there was an obvious disconnect that that we didn't talk about the same things, and I hadn't felt that before.
00:27:53.359 --> 00:28:01.279
And then I learned, okay, the the North American business central ecosystem, how partners work, what they do, that's just a completely different story.
00:28:01.440 --> 00:28:05.119
And neither of us were stupid, we were just coming from a different direction.
00:28:05.200 --> 00:28:09.759
And what was a best practice for me didn't make sense at all to the person I talked to.
00:28:09.839 --> 00:28:12.640
And that's also part of why um I agree with you.
00:28:12.799 --> 00:28:18.000
Best practices sometimes can be um a bit arrogant to think that you know the best practice.
00:28:18.480 --> 00:28:19.839
Yeah, no, it's tough.
00:28:20.240 --> 00:28:22.240
Again, I always hear that about the conferences.
00:28:22.400 --> 00:28:28.880
I keep saying I have to get over to one of the European conferences, and everybody tells me just to see the difference as well.
00:28:28.960 --> 00:28:58.720
It's it's not not to get over and experience the conference in uh whichever location is in, the the culture and the beauty of uh whichever city that they're hosting it in, but just to see also like the difference, as you had mentioned, because I talk with um a lot of Europeans uh will come over, or individuals from Europe or other parts of the country will come over to the Directions North America conference, and they all say the same thing is like it's not that it's it's not that one is better or or worse, it's just different because the messaging is different.
00:28:58.960 --> 00:29:03.039
Um so I I just I really want to get that experience in one day.
00:29:03.119 --> 00:29:04.640
I hope that I uh I can do that.
00:29:04.720 --> 00:29:09.359
I always say I'll do it, but it becomes a logistic uh challenge at some point.
00:29:09.759 --> 00:29:11.279
Yeah, it's it's the rest.
00:29:11.839 --> 00:29:12.240
Yeah.
00:29:12.559 --> 00:29:19.119
No, it's not because in Europe, uh Europe has a lot of countries that are close together, that you have transit together.
00:29:19.279 --> 00:29:26.000
Here we have a big hop to go from you know one continent to the other, so it gets a little challenging.
00:29:26.160 --> 00:29:28.880
Um, so but to jump back to the MCPs as far as setting up.
00:29:28.960 --> 00:29:37.279
So we talked about you uh one strategy is to sort of compartmentalize or have different configurations by maybe area or some sort of subset of data.
00:29:37.359 --> 00:29:48.400
So you had mentioned maybe having a sales which could have access to sales orders and customers, possibly, but maybe you have some customer-related uh configuration just as the customers and the ship to addresses, and then you have a sales order one.
00:29:48.480 --> 00:29:53.920
You could have a purchase order one or purchasing one, same type of breakdown, then maybe an inventory one if you want to do inventory values.
00:29:54.079 --> 00:29:57.359
So uh you have a different segmentation between those.
00:29:57.440 --> 00:30:08.960
So then also we'll help the uh orchestrator know which set of tools to use so it doesn't have to scan the entire set of tools to choose the best one, it just knows that this configuration is for sales.
00:30:09.200 --> 00:30:11.279
Now we're talking about configuration.
00:30:12.160 --> 00:30:14.480
How do you configure it within Business Central?
00:30:14.559 --> 00:30:17.759
Like what is it made up of uh on the underlying points?
00:30:17.839 --> 00:30:21.279
Um, and how do you determine what access the customer is?
00:30:21.359 --> 00:30:24.000
So what's the construct of it within Business Central?
00:30:24.720 --> 00:30:35.039
Yeah, so what you have is basically a configuration, um, a configuration page where you can just say, okay, this is the name of the configuration and this is what it's allowed to do.
00:30:35.200 --> 00:30:40.640
On this and that um functionality, does it have read, write, um, modify?
00:30:40.799 --> 00:30:50.400
So read, create, modify, delete, and then you have the bound actions um that basically trigger some code on an entity, and that is what you can do.
00:30:50.559 --> 00:31:00.319
You can say, okay, um, on that customer list I want to allow read, I want to allow create, and I wanna uh I don't want to allow modified and delete.
00:31:00.480 --> 00:31:11.839
And that would be the configuration that you set up, and that then um changes what is available as the um as the MCP uh surface, so to speak.
00:31:12.640 --> 00:31:16.319
So you you're maybe mentioning entity, so it sounds like it's API driven.
00:31:16.400 --> 00:31:17.599
So yes, absolutely.
00:31:17.759 --> 00:31:18.000
Yes.
00:31:18.240 --> 00:31:18.559
Okay.
00:31:18.720 --> 00:31:25.200
So the MCP is uh configuration is uh a collection of APIs in essence.
00:31:25.440 --> 00:31:37.039
Um so with that that allows end users and partners if they're comfortable creating API pages, because in Business Central, that's what they call this API pages.
00:31:37.279 --> 00:31:43.279
So you could create your own API configuration configurations.
00:31:43.359 --> 00:31:46.240
I don't even know what to call things anymore, uh, to be honest with you.
00:31:46.319 --> 00:31:48.400
The terminology seems to change all the time.
00:31:48.640 --> 00:32:00.240
So you have your own, you can create your own API configurations or API endpoints that you can put as part of an MCP uh on top of the APIs that Microsoft has available for you as well.
00:32:00.640 --> 00:32:00.960
Yeah.
00:32:01.119 --> 00:32:13.519
And and the thing is you have those um you have those pages that that those API pages that are exposed, and then you have something like list customers, and then you also see the ID of the page.
00:32:13.759 --> 00:32:26.720
So if you still want to see what is connected to it, you you basically see in those tools what it's what it's called in um in an MCP, you still see what the actual page is with the ID and and what what you can do with it.
00:32:26.880 --> 00:32:38.559
And that gives you a pretty direct connection between that tool, like list customers page XY and um the actual page in Business Central, so you know exactly what it's called.
00:32:40.319 --> 00:32:41.039
Interesting.
00:32:41.279 --> 00:32:44.799
I am going to have to set some more of these up and play with them a little bit.
00:32:44.880 --> 00:32:54.480
And uh after my weekend working with some orchestration stuff, uh I want to see if I can get it to orchestrate with uh Business Central.
00:32:54.559 --> 00:33:06.880
And you had some uh your points that you hit uh touched upon you know integrations and communication with other systems, whereas before a lot of times organizations would have to create connectors of some sort to connect the two together.
00:33:07.200 --> 00:33:12.559
Now it seems to me, it sounds, I understand it's still I don't even want to say new.
00:33:12.640 --> 00:33:22.880
I guess the whole thing is new because you know this technology is not um uh as mature, I guess you could say, as some of the the long using technologies uh as we talk about changing.
00:33:23.279 --> 00:33:28.240
But it it just makes it a little bit easier for them to be able to put all this together.
00:33:28.400 --> 00:33:35.440
Um and like you had said, hopefully you can limit the scope to to have a little more accuracy.
00:33:35.599 --> 00:33:41.119
Um and I don't want to say they don't say you're supposed to call it hallucinations because that's negative.
00:33:41.279 --> 00:33:42.559
What are you supposed to call it?
00:33:42.880 --> 00:33:50.000
Supposed to be uh no, that's what I was listening uh to something the other day because hallucinations is negative.
00:33:50.079 --> 00:33:52.319
I don't know, I'll have to look that up anyway to go back to that.
00:33:52.400 --> 00:33:54.400
So no, that's it's it's uh the same thing.
00:33:56.079 --> 00:33:57.279
It's fabrication, Brad.
00:33:58.400 --> 00:33:59.920
Yeah, that's what they're saying.
00:34:00.240 --> 00:34:00.559
Thank you.
00:34:01.039 --> 00:34:04.559
Fabricates new stuff that has nothing to do with it.
00:34:05.359 --> 00:34:06.000
Yeah.
00:34:06.240 --> 00:34:09.519
So no, that's that's what that's that's the interesting conversation then.
00:34:09.599 --> 00:34:11.039
You do you have that orchestration layer?
00:34:11.199 --> 00:34:11.920
Okay, first step.
00:34:12.159 --> 00:34:19.519
Second step is what does that mean for functionality that currently actually partners create for customers, your small PTEs?
00:34:19.920 --> 00:34:25.039
Is that something that should still be in PTE or should it now be something that is created by your agent?
00:34:25.280 --> 00:34:28.880
Similar to the conversation we had in the past around the power platform.
00:34:29.360 --> 00:34:33.039
Am I creating a PTE and Business Central or am I creating a Power App?
00:34:33.199 --> 00:34:35.840
Now the question is: am I creating a PTE?
00:34:35.920 --> 00:34:42.480
Am I creating a Power App, or am I just doing letting an agent do the magic and trust that it actually works?
00:34:42.719 --> 00:34:47.840
And then you have the layer on top that that's AI in the product, but you can also do AI in development.
00:34:47.920 --> 00:34:52.320
So what's stopping me from saying, create a PTE for me that does exactly this and that?
00:34:52.400 --> 00:35:17.679
And it just opens up so much, and that's the software development part only if you now think about requirement management, if you think about documentation, if you think about uh quality assurance, it's at the core of it and where I find it it started was basically software development having a bit of a helper on that, um, and then letting it go and create the whole stuff once it got agentic.
00:35:18.000 --> 00:35:35.039
You see similar things in M36 uh M365 Copilot, where yes, it could help me to set up um or to create the summary of my meeting that I had and come up with to-do's, but now suddenly it can do a bit more, it can create stuff, it can um uh finish to-do's actually for me.
00:35:35.280 --> 00:35:40.320
And then if you think about next step, okay, why are we in between that as partner?
00:35:40.559 --> 00:35:43.039
Why why can't an end customer do it themselves?
00:35:43.119 --> 00:35:45.920
And who are the people at the end customer who are doing this?
00:35:46.000 --> 00:35:47.280
Do we need those anymore?
00:35:47.440 --> 00:35:49.760
What's what's the role gonna be in the future?
00:35:49.840 --> 00:35:53.039
And there I'm starting to struggle.
00:35:53.599 --> 00:36:10.719
Yeah, it does it does lower the lower the barrier of uh for entry for things that you could do with business central, but you know, at the same time, if you have someone that does that, um you know, you you you run into issues that they may not understand it.
00:36:10.880 --> 00:36:27.280
Uh yes, it lowers the barrier, but if they don't understand it, that's when people like us come in and be like, okay, yeah, I mean it got you there fairly close, but you don't, you know, maybe you're not quite understanding the um the approach or the the architect.
00:36:27.599 --> 00:36:32.079
So uh dangerous world for a minute, uh I'm sure I'll prove it.
00:36:32.400 --> 00:36:44.719
You hit a point, and I think the it was a challenge that we had with the Powell platform when it was first was released to be able to work with Business Central.
00:36:45.199 --> 00:36:54.159
This is that uh to another level where the the years ago software development was the key.
00:36:54.320 --> 00:37:04.559
So you know, the ability to create the to receive a design from somebody who created it to develop that, even if you self-created.
00:37:04.719 --> 00:37:14.320
I'm saying somebody created a design, and then somebody was given the design, and they came back with like magic, and now you have something that does something.
00:37:14.719 --> 00:37:31.039
Now it's almost where it's changing, where it's you need to know the technologies to build the proper design, and then this is where AI is becoming good at creating your uh your design, right?
00:37:31.119 --> 00:37:36.559
It's the it's it's almost like giving it off to someone and doing the magic, but it's also how you orchestrate.
00:37:36.639 --> 00:37:54.239
Again, you talk about the MCP orchestration or somebody uh an orchestrator managing the different MCP MCP configurations, now understanding the technology stack, understanding the technology and the functionality to be able to now review what gets done by those different pieces.
00:37:54.400 --> 00:38:00.159
See, this is what I keep saying is uh everyone kind of laughs at me or something, but I say software development's dead, right?
00:38:00.239 --> 00:38:02.000
I don't mean it's dead like it's going away.
00:38:02.159 --> 00:38:06.880
I'm saying it's dead as we know it because now it's like the developers have to come up a level.
00:38:07.119 --> 00:38:18.559
And in business, I'm talking within business central as well, because now you have to understand the application, you have to understand the technology and the options to then be able to give it to an AI, to be an AI, I don't even know.
00:38:18.719 --> 00:38:30.159
I call it AI because it means different things, to give it to a model or to give it to an agent or give it to whatever you'd like to call it, to create something that you then review to make sure it meets the stack.
00:38:30.960 --> 00:38:50.400
So the role is sort of being pulled back, and then also from the functional side, it's being pushed down to where now you have that hybrid of you need to be functional from business point of view and technical from architectural point of view to know that all these things exist to help sort of guide the road that we go on.
00:38:50.480 --> 00:38:51.679
That was my long-winded speech.
00:38:51.760 --> 00:38:52.159
I'm sorry.
00:38:52.639 --> 00:39:04.079
No, but I I fully agree, and I think that's one of the benefits of being a BC developer compared to, let's say, a C sharp or Java developer, because BC always was very business focused, right?
00:39:04.239 --> 00:39:22.000
There's no way you can be a good, at least in my opinion, that you're a good BC developer without having a good understanding of the business part of it, understanding how it all works, having a clear understanding of the requirements, because there is not really a lot of plumbing code or technical code that doesn't do business value, right?
00:39:22.239 --> 00:39:25.519
Everything we do in BC is connected to the business requirement.
00:39:25.599 --> 00:39:28.800
It is for a reason a business requirement of the customer.
00:39:28.960 --> 00:39:42.079
Now, if you compare that to someone who is doing C sharp or Java or Rust or whatever, there might be that technology code that does nothing but put the database together with the cache and your business logic or something like this.
00:39:42.239 --> 00:39:58.480
And those people who were very far away or are very far away from the actual business application, I think for them it's an even more disruptive change because that I'm pretty sure an AI will do very, very soon and will do better than any human can do.
00:39:58.719 --> 00:40:12.880
But that the job that a business central developer typically does, understanding a business requirement and translating that, that is something that that is very, very valuable still in the world of AI, because that's actually what I need to tell the AI to do.
00:40:13.119 --> 00:40:21.039
Now, if I'm the person who derives most joy out of writing the lines of code, that that might be a challenge in the future.
00:40:21.199 --> 00:40:37.199
But if I'm the person who enjoys understanding the requirement and making sure that the code happens and we arrive in a good place where we have code that actually does what I want it to do and is maintainable and is very solid from a quality and performance perspective, that is something that's not gonna go away.
00:40:37.519 --> 00:40:38.960
Will I still write the code?
00:40:39.119 --> 00:40:39.760
Probably not.
00:40:40.000 --> 00:40:43.119
Do I need to understand the code and review it at some point?
00:40:43.360 --> 00:40:44.159
Probably yes.
00:40:44.320 --> 00:40:49.920
I mean, you also see people who say, okay, I just let it also create test automation and then I'm good.
00:40:50.559 --> 00:40:55.519
There, I'm not convinced that this really works for complex requirements yet, but let's see where we go.
00:40:55.679 --> 00:41:04.559
Um, but the need to understand the business need or the the business requirement and translate it into code, that's basically the same, I would say.
00:41:05.039 --> 00:41:05.760
Spot on.
00:41:06.079 --> 00:41:06.880
You nailed it.
00:41:07.119 --> 00:41:14.800
I think in it, I think business central is a little more challenging because as you had said mentioned, we're in a box already.
00:41:15.039 --> 00:41:19.760
Like we have an existing, it's not green field create an application that does this.
00:41:20.000 --> 00:41:25.679
It's we have a box that we're in, and we need to enhance that functionality.
00:41:25.840 --> 00:41:27.280
So it's a little more challenging.
00:41:27.360 --> 00:41:34.559
So again, as you had mentioned, it's it's imperative, important for someone to understand the application, the flow of the application.
00:41:34.800 --> 00:41:48.480
And reading code is much different, much different than writing code, because it's much easier to read a book, in my opinion, and understand a book and then make suggestions than it is to write a book from scratch.
00:41:48.639 --> 00:41:52.400
So if you're writing it from scratch, you may struggle and have challenges.
00:41:52.559 --> 00:41:59.920
But if you're reading code that gets output by a model, you'll be able to follow along and say, oh, this is not proper.
00:42:00.239 --> 00:42:01.199
This is uh proper.
00:42:01.360 --> 00:42:03.519
So it's it's there's a clear distinction there.
00:42:03.679 --> 00:42:16.239
So it's not um it's simplifying it in a sense from the technical point of view, uh, making it a little more challenging from the business point of view because you really need to understand those processes and how that works.
00:42:16.400 --> 00:42:21.599
So it's it it's also more challenging, right?
00:42:21.679 --> 00:42:24.000
It's it's very information dense.
00:42:24.239 --> 00:42:33.280
If I'm writing code, there are still moments where I can if I'm in the flow, I can kind of relax because I exactly know what I'm doing.
00:42:33.440 --> 00:42:37.599
Now, if I'm just reviewing code over and over again, I need to understand what is happening.
00:42:37.679 --> 00:42:41.679
At least for me, I I would say that takes a higher mental burden.
00:42:41.840 --> 00:42:55.440
It's it's more exhausting um than creating stuff, reviewing, having an idea, understanding, giving feedback to that machine that always says, yes, perfect, you're right, and then understanding, no, I might actually not be right.
00:42:55.519 --> 00:42:57.440
So um let's take a closer look at it.
00:42:57.599 --> 00:43:06.960
So from that perspective, and and I'm seeing some people um in the wider development ecosystem who are also starting to think about that.
00:43:07.199 --> 00:43:10.000
What does that mean for the developer role then in the future?
00:43:10.079 --> 00:43:11.599
How do we stay focused?
00:43:11.840 --> 00:43:18.159
Are you able to just orchestrate 20 agents running around and creating stuff and reviewing and whatever?
00:43:18.480 --> 00:43:22.800
Is that something that a human is able to do um at the end of the day?
00:43:22.880 --> 00:43:24.639
That's also something we need to learn.
00:43:24.800 --> 00:43:33.679
Sorry, that's not very very much on the meta-level, but uh something that that popped up recently in my bubble somehow, and um it's a fascinating question as well.
00:43:34.239 --> 00:43:37.519
It is well, I'll go along with your your philosophical question.
00:43:37.599 --> 00:43:43.360
It is a fascinating question, and it's it's it's you have to think about it because it'll come.
00:43:43.599 --> 00:43:50.719
It's coming, but what do we do when no one knows it?
00:43:51.119 --> 00:43:56.800
And uh this is again philosophical because now we're talking this technology is so new.
00:43:56.880 --> 00:43:59.760
We're talking months, we're not talking years.
00:44:00.239 --> 00:44:02.960
Sorry, or maybe a year or about, right?
00:44:03.039 --> 00:44:28.159
The LLMs were created many years ago, but to the to the extent of where code is being generated, and the code is being generated properly, and even for AL coding, I've created some stuff with AI or with the models that is spot on and it's much better than I would have done, or many of the developers that know would have done, including documentation, including you know, here's a list of a summary of changes, and this is how you should test it.
00:44:28.320 --> 00:44:29.920
Like all comes with it.
00:44:30.400 --> 00:44:35.280
But I can look at it because I've been developing for so long and know that it's okay.
00:44:35.519 --> 00:44:44.400
Now let's just take now we have everybody that's a developer today elevate them to the role of this orchestrator of business knowledge.
00:44:44.480 --> 00:44:47.119
Let's just say everybody falls into a nice place there.
00:44:47.280 --> 00:44:52.960
We're all going to get old, we're going to retire and no longer be uh working uh in the space.
00:44:53.119 --> 00:44:56.159
Hopefully, I get to retire at some point in my life.
00:44:57.360 --> 00:44:59.360
What about new people coming in?
00:44:59.440 --> 00:45:10.320
Because if they're not brought, I don't want to say brought up, but if their career doesn't progress with them learning the fundamentals to be able to make that distinction, what happens?
00:45:10.480 --> 00:45:17.119
So, what happens in 10 years when all of the old timers, as we'll call them, retire?
00:45:17.519 --> 00:45:20.559
And 10 years is probably a defensive guess, right?
00:45:20.639 --> 00:45:21.760
That change might happen sometimes.
00:45:21.920 --> 00:45:23.280
I'm trying to be nice.
00:45:24.960 --> 00:45:31.679
But I mean the counter-argument, at least for me, to that is do you still know exactly how to write a SQL statement?
00:45:31.840 --> 00:45:34.880
Do you know how to create assembler code?
00:45:35.039 --> 00:45:42.400
That that was we leveled up that one abstraction level, and especially AL or CAL in the past is then even one abstraction level on top of that.
00:45:42.559 --> 00:45:44.719
Maybe AI is just the next abstraction level.
00:45:44.800 --> 00:45:50.880
Maybe we don't worry about that anymore in three years, three months, five years, ten years, I don't know.
00:45:51.039 --> 00:45:56.320
But at some point, maybe we don't worry about that anymore because AI just does it perfectly.
00:45:56.480 --> 00:46:09.599
We all we need to do is be able to describe the requirement in the right way, and then AI is able to do it perfectly, the same way as C sharp code hopefully nowadays um uh transpires correctly into assembler code.
00:46:09.760 --> 00:46:17.599
That I don't know, but that could be an outcome that we we're just one abstraction level higher and everything still works.
00:46:17.920 --> 00:46:22.719
The learning is typically it helps to be one level below what you actually need.
00:46:22.800 --> 00:46:27.760
So even if you're writing AL code, it's still very helpful if you have an idea how SQL works.
00:46:28.000 --> 00:46:37.760
Not for 80% of your work, but for the interesting 20 or maybe even 10% of your work, it helps a lot because then you can dig into the performance issues that you see.
00:46:37.920 --> 00:46:42.880
So if that's the right the right idea, then we should probably still understand it.
00:46:43.039 --> 00:46:44.079
But do we have to?
00:46:44.239 --> 00:46:44.800
I don't know.
00:46:45.199 --> 00:46:49.199
Yeah, that's just like uh I put that similar to like uh pilots, right?
00:46:49.360 --> 00:46:50.320
We still need pilots.
00:46:50.400 --> 00:46:59.119
I mean, clearly the plane can fly on its own most of the time, but you still need someone to be seated and know what to do if that doesn't work.
00:46:59.280 --> 00:47:05.280
So you I I don't think it's necessarily gonna go away, but you still gotta learn how to do it manually if needed.
00:47:05.360 --> 00:47:13.760
Uh, and I think that comes into that uh that you had mentioned earlier is understanding the business logic or the business space in that in that area.
00:47:14.000 --> 00:47:25.280
But it is tough because I had someone ask me if their kid, should their kid be learning how to develop and should they still go to you know uh learn all that stuff.
00:47:25.519 --> 00:47:27.280
I said, yeah, I said I my answer was yes.
00:47:27.599 --> 00:47:40.480
I think they should still, but the again it goes back to that barrier of to entry is much lower because then they can just get all that information as quickly as possible and then build what they're and you know intended to build.
00:47:40.719 --> 00:47:41.920
So that's gonna be a tough one.
00:47:42.000 --> 00:47:43.599
I think it's gonna be a couple years, man.
00:47:43.679 --> 00:47:47.119
Uh I know you said 10 years, but uh you're being super nice.
00:47:48.719 --> 00:47:49.280
I am.
00:47:49.360 --> 00:47:51.280
I my true feeling is different.
00:47:51.360 --> 00:48:06.559
I I liked your response, and it really puts a good perspective on it if you're talking about the understanding one layer lower than where you need to be, because I was thinking who I did learn assembly.
00:48:06.639 --> 00:48:07.760
I'm that old.
00:48:07.920 --> 00:48:09.840
Uh they taught me assembly again.
00:48:09.920 --> 00:48:15.119
That was the one level below, because I first started with uh other languages and then I had to know assembly.
00:48:15.199 --> 00:48:19.280
I didn't learn punch cards, which was the predecessor of that in essence.
00:48:20.400 --> 00:48:21.440
Will it even matter?
00:48:21.599 --> 00:48:22.320
I think you're correct.
00:48:22.400 --> 00:48:24.239
Um, how many people know assembly language today?
00:48:24.320 --> 00:48:37.360
I'm sure there's a small few that need it for something, maybe in there, uh depending on where they are, but it's not wide-scale use of assembly for uh developing software.
00:48:37.440 --> 00:48:40.880
Uh, maybe for some hardware controls or some other things like that, it is.
00:48:40.960 --> 00:49:01.920
So I think that's is an interesting interesting perspective because it's a skill living today we think that we need, but we have to be a little more abstract and in the future say, just like some of the skills that were needed in the 1900s, not many of us know how to make our own clothes or make our own soap or or make our own milk and cheese.
00:49:02.079 --> 00:49:03.840
Um, because we don't need to.
00:49:04.000 --> 00:49:08.000
But there are still a set that does that's a very uh valid response.
00:49:08.159 --> 00:49:09.519
You just shot me down now.
00:49:09.599 --> 00:49:12.719
So now I'm going to that was absolutely not my intention.
00:49:12.880 --> 00:49:16.000
That's no clue if it would be right, but I think that's a perspective you can have.
00:49:16.239 --> 00:49:16.639
That's the point.
00:49:17.039 --> 00:49:20.559
No, no, it's your perspective was when I say I listen, my that's my sarcasm.
00:49:20.639 --> 00:49:28.400
Everyone kind of knows my sarcasm's a little harsh sometimes, but no, I think uh your perspective was good, and um it's it's a great way to look at it.
00:49:28.559 --> 00:49:33.119
Um, so we'll have to see where we are um uh in in the future.
00:49:33.360 --> 00:49:34.320
See, I could talk to you all day.
00:49:34.639 --> 00:49:35.119
Next month.
00:49:35.280 --> 00:49:36.480
It could be next month, man.
00:49:37.360 --> 00:49:37.920
That'd be interesting.
00:49:38.239 --> 00:49:39.039
It could be quick.
00:49:39.280 --> 00:49:39.599
Let's see.
00:49:39.920 --> 00:49:42.559
By the time this comes out, this might be outdated.
00:49:42.719 --> 00:49:46.079
So, just like at the beginning, yeah.
00:49:46.559 --> 00:49:50.559
Just like the beginning of this podcast, when we were talking about you know a weekend project.
00:49:50.719 --> 00:49:55.920
I mean, I I would have had to like go online course to learn how to create websites.
00:49:56.079 --> 00:49:57.039
I didn't have to do that.
00:49:57.199 --> 00:50:01.760
I mean, it two hours, it's that's what it feels like right now.
00:50:01.840 --> 00:50:03.199
So yeah, it is.
00:50:03.280 --> 00:50:04.719
It's uh it's amazing.
00:50:05.039 --> 00:50:07.920
So another we have uh a lot of things we wanted to speak about.
00:50:08.000 --> 00:50:17.119
Um I had to be respectful of your time, but another thing that's uh we're pretty um interesting topic is this whole topic of open source software.
00:50:17.360 --> 00:50:23.440
Um, because I know it's another thing that um you like to speak about and talk about as far as open source software.
00:50:23.760 --> 00:50:27.519
And now we'll even take this a little bit further.
00:50:27.679 --> 00:50:32.400
Let's we just spent some time talking about AI generating code.
00:50:33.679 --> 00:50:39.760
And I did hear an interesting conversation over the weekend about how AI is an impact and could impact open source software.
00:50:39.840 --> 00:50:44.880
So let's talk briefly about some open source software and uh your take on open source software.
00:50:45.840 --> 00:51:09.199
And in generally, in the past, I I really enjoyed being part of some of those things and and contributing in in very small ways, but still contributing to open source, but because it felt like so much business software was built on it, so much money was made from it, but still there were people who were basically doing it for free for the sheer joy of just doing it.
00:51:09.360 --> 00:51:12.159
And that was I I was very much attracted to that.
00:51:12.239 --> 00:51:13.039
Let's put it like this.
00:51:13.119 --> 00:51:24.159
The whole idea very much appealed to me that someone was willing to just spend their time to do something for the fun of it and and making sure that something got better without getting directly paid for it or anything.
00:51:24.400 --> 00:51:26.239
That that's a very nice concept.
00:51:26.400 --> 00:51:43.280
Um, the point is now with AI that you're seeing more and more struggle with that because A, some AI bots are just flooding open source um repositories with meaningless or maybe meaningful, but still not good enough pull requests and feedbacks, etc.
00:51:43.599 --> 00:51:53.599
Maybe you saw that story that I saw over the weekend as well, where someone declined um a pull request by a bot, and then that bot spun up a blog and wrote about it.
00:51:53.679 --> 00:51:54.960
I mean, that's crazy, right?
00:51:55.679 --> 00:51:56.800
Where's the limit to that?
00:51:56.880 --> 00:51:57.760
What's what's the point?
00:51:58.000 --> 00:51:59.280
What are we even doing here?
00:51:59.519 --> 00:52:13.679
And then you have more and more open source libraries that say, look, uh, the whole idea of why I did this was I wanted to get attention, I want wanted developers to see my solution and then maybe upgrade to the premium tier or get some license money out of it.
00:52:13.840 --> 00:52:26.960
And now suddenly no developer is seeing it anymore because a bot is crawling or uh an agent is crawling that and uh just um getting the information, using it, the developer doesn't even know that they are using this and that great open source library.
00:52:27.119 --> 00:52:31.760
So they decided to shut it down, it's no longer free, it's no longer open, you have to pay for it.
00:52:31.920 --> 00:52:36.960
You can still contribute, but to use it, you have to pay for it because that's the only way how it can be monetized.
00:52:37.119 --> 00:53:13.199
And that also might have an impact on um how we create software, maybe less so in the business central world where it's not that much open source, but um in other parts of the ecosystem, that suddenly also becomes very relevant because the way how open source was monetized, or at least a big part of it, um the the appeal that it had that you can, as an open source author, also get some um visibility and maybe land a job through your open source project, or you operate to the premium level and you get some money out of it, all those aspects suddenly go away if it's just an agent that is pulling information out of your repository.
00:53:14.239 --> 00:53:20.800
So it is the open source market is going to have some sort of disruption.
00:53:21.039 --> 00:53:31.840
And it this listening to you speak about that uh uh drives me to another philosophical question of what happens when we're saturated with all this software?
00:53:32.320 --> 00:54:04.960
Like when does if if if models were trained based on and you may give me the same answer you gave me to the last philosophical question, but if the model was trained on open source, let's just say, I mean, I know it was trained on other things as well, but let's just narrow the scope, open source software, and it was able to read all of these packet these libraries to determine how to create code, and then now it's making modifications to the code, and then it's going to continue learning on that code because it created that code.
00:54:06.320 --> 00:54:08.159
When does the cycle end?
00:54:08.400 --> 00:54:10.639
Do we stop creating new stuff?
00:54:11.119 --> 00:54:15.760
Or is the AI going to create new stuff itself?
00:54:16.480 --> 00:54:18.800
Or that's also what you see, right?
00:54:18.880 --> 00:54:25.840
That it goes off and creates new stuff that it thinks it's need it needs, but it might just be a copy of what already existed because that's what it was trained on.
00:54:26.000 --> 00:54:34.559
So there is indeed a cycle that is interesting to see if and when it will be broken, um, and if not, what that does.
00:54:35.280 --> 00:54:38.960
Yeah, the um that open source, I did see that same thing.
00:54:39.039 --> 00:54:46.559
That was a crazy story on how somebody made a pull request, a bot made a pull request, and then commented back about being the client.
00:54:46.639 --> 00:54:52.320
Now, some of the stuff I'm starting to say on the internet, the whole like not everything you read on the internet's true.
00:54:52.559 --> 00:54:59.119
I'm starting to feel more and more of that now, where you know how much of it's being fabricated versus not fabricated.
00:54:59.199 --> 00:55:09.280
I mean, I had the opportunity um to start working with open claw, multi-clawed bot, whatever name you want to call it, and it's not so easy to set up and start working with.
00:55:09.440 --> 00:55:23.760
So all these people that say they're running, you know, 10 billion agents on open claw, running, you know, large organizations doing this, I have to sometimes question how, or even to what extent are they running your business?
00:55:23.840 --> 00:55:30.239
You know, if they're just running and not doing anything, that's one thing, or running and actually orchestrating many tasks, that's another thing.
00:55:30.559 --> 00:55:42.159
So with the open source market, so do you see uh that the open that we'll just see a decline in open source projects or feel that we have a decline in open source projects with that?
00:55:42.639 --> 00:55:46.719
I mean, that there's two two contradicting things to consider here.
00:55:46.960 --> 00:55:55.360
A, it is a lot easier nowadays to create something that's open source because I can literally just tell an agent, look, this is the library I need, please create it for me.
00:55:55.440 --> 00:55:57.920
Um put it on GitHub and make it visible.
00:55:58.000 --> 00:56:04.000
It will create perfect documentation, it will create automated testing, it will do more than I could even think of.
00:56:04.159 --> 00:56:08.079
So, in that sense, it's a lot easier now to create something like this.
00:56:08.320 --> 00:56:13.039
I mean, I I do run a couple of open source tools that took me quite some time.
00:56:13.280 --> 00:56:18.400
If I now had to recreate them, it would probably be the same as the website you mentioned, Chris.
00:56:18.480 --> 00:56:24.400
It it would probably be only a few hours, which took me days or sometimes even weeks to create in the past.
00:56:24.559 --> 00:56:33.360
So from that perspective, I would say it's easier than ever to create something and make it available um to the public as open source or however you want to.
00:56:33.599 --> 00:56:37.360
But for the same reasons we mentioned before, what's the point now?
00:56:37.679 --> 00:56:49.280
Um I can't remember what it was called, but I saw a uh project running now that basically was a collection of tools for one specific person for one specific task.
00:56:49.519 --> 00:56:58.639
Because it's so easy now to create software if you have a rough idea what you actually want, and just put it on the internet, put it on GitHub, and and that project.
00:56:59.119 --> 00:57:11.440
Need to remember what it was called, but um the idea was basically a collection of tools that had been created by people for one specific purpose, no thinking about reusability, sharing, whatever.
00:57:11.519 --> 00:57:13.920
It just did exactly what they wanted to do.
00:57:14.159 --> 00:57:22.480
And you can imagine that this is now really very, very easy, and you have a huge collection of that, you might find things that are useful for you too.
00:57:22.639 --> 00:57:24.400
So that could be a driver.
00:57:24.559 --> 00:57:36.480
At the same time, for the reasons we mentioned before, the whole business model, the whole appeal of doing open source for some kind of a business value is declining rapidly, I would say.
00:57:37.280 --> 00:57:44.639
So software is becoming disposable because now you can create the tool that you need at the time to do a specific task.
00:57:44.960 --> 00:57:45.280
Yeah.
00:57:45.519 --> 00:57:47.280
And that's it's it is.
00:57:47.360 --> 00:57:53.519
It's almost now uh we're in a certain universe because we're all developers, right?
00:57:53.679 --> 00:58:00.960
Or it's our lens is through knowing many developers or even being close to many developers.
00:58:01.119 --> 00:58:07.280
So our first inclination, or my first inclination now, is I need to do something.
00:58:07.440 --> 00:58:08.400
What are my choices?
00:58:08.559 --> 00:58:15.599
I could search for it, and then when I can search for it, I could try to find an application that has the functionality that I'm looking for.
00:58:15.920 --> 00:58:22.719
Then I have to test it, and then I now I'm skeptical because was it AI created and is going to have problems?
00:58:23.039 --> 00:58:36.480
Or do I just go through a plan with Claude, come up with a good plan, say create it, go to sleep, wake up in the morning, and it's there.
00:58:37.440 --> 00:58:39.280
It's this is kind of where we are in the weather.
00:58:39.440 --> 00:58:43.760
When I and it may not even be going to bed and going to sleep, it may be five minutes later as well.
00:58:46.000 --> 00:59:05.840
Yeah, it's it's uh so it's you know, does software become disposable and is the value of the software um decreased because the ease and the ability to create it, whereas uh years and years ago that magic was to be able to create those things for people.
00:59:06.400 --> 00:59:12.400
Um so it's uh it's another interesting philosophical but yet technical point of view on it.
00:59:12.480 --> 00:59:13.519
So yeah.
00:59:13.679 --> 00:59:20.320
And the same for the question between general purpose software and and specifically for one purpose-built software.
00:59:20.480 --> 00:59:31.599
So if you think about, let's say, time tracking, if you want to build a time tracking solution that caters to the needs of a big industry or even a whole country or something like this, it's very complicated, right?
00:59:31.760 --> 00:59:35.760
Because people have very different ideas of what they need out of a time tracking software.
00:59:35.840 --> 00:59:40.239
It needs configurations, it needs to um align with laws.
00:59:40.480 --> 00:59:45.679
Then if you want to go multi-country, you suddenly have the complexity of different laws in different countries, etc.
00:59:46.079 --> 00:59:57.280
But if I think about one specific company, maybe not a huge enterprise, but a, I don't know, 50 or 100 people um company, they could have very simple requirements for time tracking.
00:59:57.440 --> 01:00:03.440
So it's absolutely possible that they could just let Cloud do that or some other agent do that and create it for them.
01:00:03.679 --> 01:00:07.840
What does it mean for the market of the general purpose time tracking providers, right?
01:00:08.400 --> 01:00:15.679
Do we for which purposes do we still need professionally built software that that caters to the needs of a lot of companies?
01:00:15.760 --> 01:00:26.480
And what are the kind of problems, what's the the class of problems that we can solve with your one-off solution that maybe works for a year or two, and then we just create the next one, or maybe a month or two, right?
01:00:26.639 --> 01:00:28.000
And then we create the next one.
01:00:28.159 --> 01:00:32.400
That that is also an interesting disruption that might happen in the software market.
01:00:32.559 --> 01:00:45.440
And if you look at the some of the stock market reactions for SaaS vendors, some of it is already visible that there seems to be a decline in trust into how much of that we still need.
01:00:46.239 --> 01:00:46.800
Yes.
01:00:47.039 --> 01:00:50.159
No, that those are very interesting questions.
01:00:50.480 --> 01:01:01.760
And I'd love to have conversations or hear from many people on their perspective and viewpoints on that, because it is it's to the point where you're creating something for a task.
01:01:01.920 --> 01:01:12.960
I've created tools, and I'm only saying it myself because it's uh not that I'm anyone special, it's just a real world use that I've created tools I used once and now they're gone.
01:01:13.280 --> 01:01:16.239
You know, I was able to do what I want, and I wasn't even worried about it.
01:01:16.400 --> 01:01:27.280
Like I I got to that point of I'm not even concerned because if I need something, I can reuse this and build it or or um add to it.
01:01:27.440 --> 01:01:30.079
So uh have you done anything like that yourself?
01:01:30.480 --> 01:01:31.039
Yeah, absolutely.
01:01:31.199 --> 01:01:35.760
I mean, the the example that comes to mind is the podcast that we that I mentioned initially, the sports podcast.
01:01:36.000 --> 01:01:46.400
The guy who did it decided to go in a slightly more professional direction, and that meant that my part of it kind of went out of it, and I had to set up my own podcast, but I didn't want to lose the history.
01:01:46.559 --> 01:01:57.599
So I created a small tool that just grabbed the RSS feed from the old one, um, downloaded everything, put it in the right order, created a new one for me that I could import in my new target.
01:01:58.079 --> 01:02:10.800
That would have been basically impossible in the past because I'm pretty sure I would it would have taken me two, three days to create it, or maybe just one day, and I most likely wouldn't have wanted to spend that much time.
01:02:10.960 --> 01:02:17.679
Um now I could just plug it into the agent, GitHub Copilot in that in that instance, um, let it run.
01:02:17.760 --> 01:02:20.400
It took me two, three attempts, and then I had what I needed.
01:02:20.480 --> 01:02:23.679
It wasn't perfect, but it was good enough, and then I just threw it away again.
01:02:24.000 --> 01:02:31.440
I mean, that's it, it literally was a problem that would have been not possible for me to solve with a meaningful amount of of work.
01:02:31.599 --> 01:02:35.519
It would have just been too much time, too much effort, and now it suddenly wasn't a problem.
01:02:36.239 --> 01:02:36.639
Wow.
01:02:36.960 --> 01:02:40.719
Welcome to 2026, where you can have custom.
01:02:41.119 --> 01:02:47.280
I mean, if you think about that, that's custom, purposeful, developed software for your needs.
01:02:47.360 --> 01:02:50.960
You had a specific scenario with specific requirements.
01:02:51.119 --> 01:02:53.039
They weren't general or generic.
01:02:53.280 --> 01:02:56.559
Jeez, I say I get tongue-tied when I start getting excited about this.
01:02:56.719 --> 01:02:58.239
And you're able to just do it.
01:02:58.639 --> 01:03:08.719
It it goes with it's almost like we talked about at the beginning of this episode about talking with people and and hearing what they have to say, learning from them and having one-on-one conversations.
01:03:08.880 --> 01:03:16.480
Now you're having that one-on-one conversation with software to have your specific needs, and like you said, good enough, right?
01:03:16.800 --> 01:03:20.639
Good enough so that you can get the job done.
01:03:20.880 --> 01:03:29.360
A lot of times that good enough comes into play even in professional, yeah, what you call professional grade software, professional grade work, because it gets the task done for you.
01:03:29.519 --> 01:03:38.000
And again, it probably took minutes for that to, you know, maybe you said you went two or three tries, so maybe would you spend an hour or two hours on that?
01:03:38.239 --> 01:03:43.440
Whereas if you would have had to done that to write something yourself, for example, it could have taken weeks.
01:03:43.599 --> 01:03:43.920
Yeah.
01:03:44.079 --> 01:03:47.679
So that's what we're doing is we're condensing weeks into hours.
01:03:47.840 --> 01:03:54.880
Uh, hopefully we all have a little more time to go to the beach, hike mountains, or do other um other good things.
01:03:54.960 --> 01:03:58.559
I'd say start picking up trash, but that's what I think everybody should do.
01:03:58.719 --> 01:03:59.840
Uh absolutely.
01:04:00.400 --> 01:04:02.960
Let's wrap the AI topic on that on that note.
01:04:03.119 --> 01:04:11.920
That's that's uh I don't know, I'm not sure if that's overly optimistic, but if we end up like this, that it gives us more time to do what we actually want to do, that would be amazing.
01:04:12.239 --> 01:04:13.679
I think that would be amazing.
01:04:13.920 --> 01:04:16.639
Well, Tobias, thank you very much for taking the time to speak with us.
01:04:16.800 --> 01:04:21.199
Uh enjoyed um all of your valuable insights and your perspective on many things.
01:04:21.360 --> 01:04:29.840
Uh, if anyone would like to get in contact with you to talk about some of the topics we spoke about or find more information about it, either one of your podcasts, what's the best way to get in contact with you?
01:04:30.239 --> 01:04:34.320
So there is Tobiasfenster.io, my first name, last name.io.
01:04:34.639 --> 01:04:40.719
That's basically my um my homepage where you can find my podcast, where you can find my blog, and also my social links.
01:04:40.880 --> 01:04:43.679
By now, I'm active on LinkedIn and on Blue Sky.
01:04:43.920 --> 01:04:46.400
More than happy to have any any conversation there.
01:04:46.480 --> 01:04:50.480
And um, yeah, let me know if you see anything that you like or don't like.
01:04:50.559 --> 01:04:57.360
Um, as we said in the beginning about conference talks, feedback is uh very valuable, and I would say the same is true for everything I do online.
01:04:57.760 --> 01:04:59.119
And very much for having me.
01:04:59.199 --> 01:05:05.840
It was a really nice conversation, and um, I would love to follow up um and to hear your thoughts on a number of other topics as well.
01:05:06.159 --> 01:05:07.760
Oh, we will we'll have to do something.
01:05:07.920 --> 01:05:08.800
We'll schedule a follow-up.
01:05:08.880 --> 01:05:18.079
I'm into scheduling follow-ups now because the the way the times change, but uh no, I do um greatly appreciate you spending your time with us today because again, time is truly the currency of life.
01:05:18.159 --> 01:05:19.280
Once you spend it, you can't get it back.
01:05:19.360 --> 01:05:27.119
So any moment you spend with somebody, uh, you know, any time you spend, any moment you spend doing something or doing something with somebody, you don't get that time back.
01:05:27.280 --> 01:05:30.480
So uh I do uh value and appreciate your time.
01:05:30.559 --> 01:05:32.079
And I look forward to talking with you again soon.
01:05:32.159 --> 01:05:33.119
Chao ciao.
01:05:33.519 --> 01:05:34.400
Take care.
01:05:34.880 --> 01:05:40.000
Thank you, Chris, for your time for another episode of In the Dynamics Corner Chair.
01:05:40.079 --> 01:05:42.079
And thank you to our guests for participating.
01:05:42.400 --> 01:05:43.920
Thank you, Brad, for your time.
01:05:44.079 --> 01:05:47.679
It is a wonderful episode of Dynamics Corner Chair.
01:05:47.840 --> 01:05:51.199
I would also like to thank our guests for joining us.
01:05:51.360 --> 01:05:54.079
Thank you for all of our listeners tuning in as well.
01:05:54.320 --> 01:05:58.159
You can find Brad at developerlife.com.
01:05:58.320 --> 01:05:59.760
That is D V L P.
01:06:01.039 --> 01:06:08.480
I f e dot com and you can interact with them via Twitter D V L P R L I F E.
01:06:09.119 --> 01:06:16.880
You can also find me at mattalino.io M A T A L I N O.io.
01:06:18.159 --> 01:06:22.000
And my Twitter handle is Mattalino16.
01:06:22.719 --> 01:06:25.679
And see you can see those links down below in the show notes.
01:06:25.760 --> 01:06:27.039
Again, thank you everyone.
01:06:27.199 --> 01:06:28.960
Thank you and take care.
