Katyella · Episode 3 · September 17, 2026 · 49:32

What Is Programming Anymore?

The week the AI discourse turned dark: a possible coming attack on open source, Primeagen on protecting your energy, and whether letting agents make all the decisions causes skill regression. Plus legacy migration as the consulting angle AI de-risks, and why Nina built apps for years without ever sharing one.

Episode three. Jet lag, AI backlash week, and a quote Matt swears he never said.

What we talked about:

  • 0:18 Catching up: Matt back from the States, Nina spinning plates.
  • 0:41 The week the mainstream AI discourse went negative, and why it reads like a coming attack on open source.
  • 1:16 Primeagen’s reaction to “Protecting Your Energy”: how much AI coding takes away from the process.
  • 8:35 Migrating the legacy dinosaurs: why AI makes the rewrite pitch less risky, and how banks only just got mobile apps to work.
  • 20:01 Skill regression. If something else makes all the decisions long enough, what is programming anymore?
  • 31:49 Nina’s confession: years of finished apps, none ever shared. This is the year she fishes.
  • 32:33 “I’ve never been one of those greybeards,” a quote Matt does not remember saying.
  • 49:22 Closing remark: planning is everything, or you just meander.
Transcript

Automatic transcript, lightly corrected. Also available as WebVTT.

Matt 0:18

Hello, Nina.

Nina 0:19

Hello, Matt. How are you doing?

Matt 0:22

Good, good. Can't complain. Just flew back from the States. So, a little jet-lagged.

Nina 0:27

Oh, yeah.

Matt 0:28

What about you?

Nina 0:31

I am... I'm doing... I'm spinning plates when it comes to projects, as usual. It was kind of a busy week. There was a lot of buzz in mainstream media about AI that was kind of on the negative side. This week that was taking up my headspace seems like just a potential future attack on open source software in general, in my opinion. But I don't know. I, I do you think that we should stop AI development? I don't think so.

Matt 1:05

I don't think we have a choice.

Nina 1:07

We don't have a choice. Exactly. So when I was looking at one of our previous episodes and I came, I also came across this content creator on, on YouTube named ThePrimeagen, his name's Michael B. Paulson, and he created a kind of a vlog video reaction to another dev entitled Protecting Your Energy in terms of AI use and just how much we use it for coding, what is taken away from the process with coding. And there were some good quotes in there, and I thought we could talk about it.

He says that it's easier to get into code, and because the barrier for entry is lower there, the bar for a good product is much, much higher because of the oversaturation. You can streamline tools, he says, but it's harder to streamline your ideas into gold. And i wanted to get your opinion on that that take because i'm kind of in that camp where i agree with him on that i think it's it's it's harder to get noticed with a product

Matt 2:26

Yeah i mean it's always been hard to get noticed with the product like what's the only product that's like been good enough to be good on its own like, Linear what else has been like oh this product's so good we're going to use this or it's not just like salesmanship and marketing that's getting like that's the go-to-market, build it they will come like hasn't worked in 15 years right so that was not an ai problem.

Nina 2:56

Well there there is an aspect of good software is is hard to come by because there are people who are just building things so quickly so he brought up that the slow build and he was he was he was writing out like his process when he was building applications on like a whiteboard and just like how he would build it piecemeal by piecemeal and construct the architecture as he went along. Maybe that's just him. That's like the type of dev he was, but he described the process as, the process itself is the refiner and how AI kind of takes away like that step to refining a product and figuring out what works piece by piece from like a user perspective.

And I kind of agree with him there because I feel like everything that I did before AI development tools was a little more intentional up front. And then I modified my process as I went, whereas a lot of the decisions are made up front for us right now by Claude.

Matt 4:05

Are you just talking purely technical or are you talking UX? Yeah.

Nina 4:09

No, like purely technical, too. I mean, I don't know how you... Do you make all the decisions for server and everything up front? Server, database, program?

Matt 4:22

No. I mean, usually I take a theater-of approach and just kind of piecemeal things together over the long run. So, I mean, that's just... I mean, I'm giving Claude, like, okay, let's do this small now. Let's just do this part here and then okay now let's add a real database let's not just use a local one okay okay let's let's get this deployed yeah so.

Nina 4:50

Yeah yeah i think i think in his case he's spinning multiple agents and he's he's doing multi-agentic work i think but i try to when i'm when I'm programming to create an application in the same steps that I would if I didn't have AI piece by piece. I personally do try to do that. Like start with some like fake, fake data to begin with some static data and then work on the backend later. And I start from like front to back. But yeah, what'd you, what'd you build this week? Speaking of that.

Matt 5:29

Um, I have no clue. Thank you. What the hell did I do this week?

Nina 5:37

It feels like that, right?

Matt 5:38

Yeah, that's been a rough one. I did a lot of taxes. Business taxes are coming up. A lot of my Claude stuff has been building up taxes, pulling in all the forms. I didn't even look at anything this year. I just gave it all to Claude, and then it put it together. I mean, it's giving it to an actual CPA. If he's doing Claude, I don't know. But I basically got a drive that I give Claude access to. And then it also has access to my emails.

So whenever he emails me, I just tell Claude to read the emails and answer for me. And I give it like a quick once over and then try to like... Some things I need to do because TurboTax is annoying. Not TurboTax. What's the... Keeping one. QuickBooks is annoying yeah and i have a separate thing that does my accounting or not count bookkeeping, and so they they will require me to use QuickBooks and so i just have to like manually export some reports to pdf and then Claude's able to read the pdfs i don't think, QuickBooks will ever integrate MCP that works so like oh is it also back to your stuff like you don't need a good product.

Nina 7:02

See what see what i mean though that's a pain point now for for people that use products that have

Matt 7:07

Ai oh yeah well no one's gonna change that that's like saying that there's better Salesforce too but no one's switched in better Salesforce in 25 years.

Nina 7:17

But maybe maybe that's how you know a product is good if they get their if they get their internal agents correct in there i don't know No, it has to change. I refuse to believe it's not going to change.

Matt 7:30

I mean, there's early adopters that pick the easy stuff, but there's just like, the human stuff behind it that people don't want to. No one goes in and wants to be the guy that changed the sales system and it goes poorly. So it's like a first mover advantage. Once you're in, you're in. Like, there'll be cool new AI stuff.

Nina 7:50

Right, yeah. If something is made from the ground up with AI as a center of the process, if there's something like QuickBooks for doing your ledgers and stuff, but it's coded AI first in mind,

Matt 8:10

Probably.

Nina 8:11

But yeah, that's such a pain.

Matt 8:13

There's Pipedrive and there's Linear, and those are better tools that no one really picked up. There's everyone still in Jira Basecamp like in Basecamp small, it's like they're niche products so people build ai niche products the big ones will still just be gigantic, fizzled out vcs that either worked or didn't work.

Nina 8:39

Well that's where that's where devs like us could come in to migrate or actually just migrate all the legacy dinosaurs or build something new internally for every company's use case, like for individual companies. I don't know.

Matt 8:57

Yeah.

Nina 8:59

Oh, go ahead.

Matt 9:00

Do you make, like, it's just, like, they're so rarely rock the boat like that, that you get a new app. Like, I feel like banks just got mobile apps to finally work like in the past five years.

Nina 9:15

But that's like an angle for consulting or something, you know? Pitching that you know somebody can do it fast with ai so it's not like a risk you know it's not it's not as big as a risk or a huge money suck because it could it could get done quicker if you hire somebody like for their particular system you know like i don't know this this week i i built so many tools that i kind of just had a moment like you where i was like what did i what did i do this week and i had to write it all down i worked on the eta brain fixing a fixing a bug with LLM responses i that's the that's the the podcast search app that we talked about in a previous episode i created a plan for creating kind of a a word cloud that's built it's built from like the supply side and it's weighted by how many how many phrases an episode repeats multiple times so that's that's like on the next for that i think that'd be interesting for user experience and that would be something that so i i planned that i release an app for a Discord group that i'm a part of that has 20 users and i migrated from JavaScript to TypeScript in a day I think it was like 76 modules, added a testing pipeline, and then I made an app that was for testing me in my learning style and my memory style, how I remember things for algorithms.

And it has an AI code reviewer. And it tells you, based off if it's an efficient algorithm, they'll say, like, you suck or you're doing great.

Matt 11:09

I said you put an AI in LeetCode.

Nina 11:13

I could have used LeetCode, right? Like, that's the thing. Like, there are products that exist that we can use. Just like, you know, there's Jira, there's Bitbucket. But you said there's two other competitors coming up. What did you say their name was?

Matt 11:29

For Salesforce is Pipedrive and then there's Jira there's Linear for Jira test management.

Nina 11:35

Type stuff yeah so like in a similar vein i could have used i could have used LeetCode but i wanted something that was catered to me that that you know i could control like a queue of questions and then repeat the queue of questions because that's how i get better at like better at like memorizing like new like operators that i that have slipped from my mind since using Claude or regex that slipped from my mind like adding a flash card mode rapid fire time like that stuff it it's completely catered to how i learn best and that's why i made it also you have to pay for LeetCode right so

Matt 12:17

Yeah i forget it's been a while.

Nina 12:19

Yeah you have to pay for LeetCode so So that's programming around a paywall was also was also the incentive.

Matt 12:27

But yeah, but those are tools. Those aren't businesses.

Nina 12:30

So right, right. But just speaking from like, like us, just it being software that people are going to use for whatever for tools or for businesses, you know. Yeah, did you want to add anything on the business side of that argument then?

Matt 12:50

I mean, there's many angles to it, right? There's just the pure programmer angle, which I feel is just very purest programmer whining that they're not in control of the code anymore and what they do is not being as special as it used to be. That's what I don't really care about. And then there's the business side where everyone got really scared that all their sassas were going to get eaten up, by Claude and Claude was just going to spin off a bunch of stuff to do this.

Or somebody else would just come in and code your app. Business is so rarely the product. It's like a third product and a third sales and a third marketing. And it's all those three need to hit perfectly to have a successful business right it just to, and you can throw enough money and only have like one or two of those pieces fit but it's really rare that you're just like so good at product or so good at sales or so good at marketing that doesn't really matter for the other ones.

Nina 14:00

Right i guess i guess my my kind of perspective of this is kind of a very it's it's a little myopic and just focused on the product being good being the main thing that indicates the success like there are the other factors as well like absolutely yeah Yeah, I don't know. Well, when you have businesses or people who are going to strategically market a software, I guess I'm trying to think of a follow-up to that. What do you think changes for them because of AI? Because they're using that in their day-to-day in a way where they actually probably are evaluating the worth of products based off using AI because people are becoming so dependent on it.

How do you think that affects their judgment when it comes to presenting a product?

Matt 14:55

I mean, one, AI is just bad at their jobs still. It's getting better. Men can do some of the lay work, but, you're not just seeing companies spin up marketing SEO that's just taking off. You're not seeing companies spin up AI podcasts for their companies that are just taking off either. Like even podcasts that they got running they put a little ai on it to like, ease production but that's like it's not, it's not the marketing movement it's not the sales movement for these ai things, you got like companies taking in like phone calls through ai just to like record it down, that's like just a fraction of the sales funnel right like you're like still hodgepodge in these tools together and it's up to human to piece it together into a way that functions for the customers.

Nina 15:59

Right so in in your experience with with your business How do you how do you focus on all three of those in a way that's balanced marketing sales to actual product since we're business topic of it? You know, I yeah,

Matt 16:21

Things go a little faster now. Like I can update my website super fast. It's still some back and forth to like get things laid out the way I want and I have to rewrite everything. Thing but to kind of get like i would spend more time before ai like just structuring a website, and then doing actual copywriting but the copywriting still me Claude's not great at it yeah i could do like some cooler things like i could think of an sv an animation and like, i can have Claude make the SVG that i could have never done i would have had to hire somebody to do, so that's the marketing part and sales I mean I was never going to use Salesforce in way too small, but I was using things like Trello to manage that and now I just have like a markdown, in a repo that Claude kind of manages all this stuff and like gives me a daily report on where things are at, yeah and then product is just execution of code from my business and so that's just Straight up what we're all used to, coding.

Nina 17:35

Yeah. It's like everything has become a pipeline. Everything feels like a workflow pipeline or something now. I'm still not to that point. I did plan out a project this week where in order to get better at multi-agentic coding for everything from automating, answering emails, sending out resumes or or invoices like i i did get the idea to just get better at programming by making kind of like a harness or a gui for multi-agentic work to to build apps and then also for for anything that i need to do on my computer but it's a gui in which is catered to me and my needs and my what what i do on a daily basis haven't haven't started programming

Matt 18:36

Well like what is getting better at programming these days like what's the motions you go through.

Nina 18:42

I you know you still have to build some stuff from scratch

Matt 18:47

Yeah so you're building stuff like that's.

Nina 18:49

Like but but i mean i mean like when you actually there's a difference between like building it and Claude Claude building it with ease and then building it from the ground up yourself programming every line of code like i i don't know i i used to be that person so like before before agentic programming before ai i had four i had 42 GitHub repos of various full-stack projects. And I would spend hours. I didn't really... There was a period I didn't really go out much. I was just programming every single night.

And through that, I learned how to just build products better and Like, as I was solving, I was doing algorithms last night, right? Because I'm trying to stay fresh because they still make you solve algorithms at interviews. I was realizing how much I missed, like, actually coding. And I got a little sad last night while I was coding. Like, yeah, Claude can make all these decisions for you. But, I mean, there is a skill regression there. There is a skill regression when it comes to potentially problem solving if something is long term making all the decisions.

So that's what I mean by, you know, what is programming anymore? Well, I don't know. Something that's futile and It doesn't matter to most people who are going to hire you to do something. They just want the end result, right? But there is value, I think, in building things from scratch and running into the issues yourself, running into them often. Yeah. Yeah.

Matt 20:45

I don't know. But you're doing two different things. So you're building products, but you're also studying algorithms.

Nina 20:52

So right right right right but when when you just just typing out just typing an algorithm and running into something as silly as like a syntax error it just it brought me back to just just actually just actually having my hands on the keyboard and the code editor open that's all i'm saying is that it it did make me think like i mean there's a hundred ways to solve one algorithm to. So there's like, it's just as there's like 100 ways to write a function a different way or to write a group of functions to do one thing a different way.

So it's like that aspect that I feel like I miss and is missing. Like, I don't know. I know I'm a minority here.

Matt 21:39

Oh, you're at least half of this podcast, so.

Nina 21:42

Wait, what?

Matt 21:43

You're at least half the people in this podcast, so that kind of. So we had this stuff with cloud, right? Before it all started, you need to know Nginx. You need to know how to run the PHP daemon on a box. You need to know how to set up the box and connect it to the LAN in the closet. You need to know how to get the IP through. But all that stuff now in the cloud is just like, you just put it in an EC2 instance and you let an Amazon load balancer configure that all for you.

Nina 22:17

You're right. They're terrifying. Just the way it used to be where you had garbage collection in the past, and then we had new frameworks and new things that came about. I think compared to you, I'm kind of a young dev. How many years of programming experience do you have? 20?

Matt 22:34

20. Yeah.

Nina 22:35

20? Yeah. I've got seven and a half. So... So I don't have a perspective of, I guess, as a whole, from personal experience, how tech has evolved and how this is just another step in improving the process.

Matt 22:55

It's just another abstraction. And there's always a bunch of devs who get upset, that their thing got distracted. And then they don't realize they're getting paid to build products and not getting paid to write code.

Nina 23:10

Okay so okay then i'm i'm probably that one that one person who's just like an old man about it like i can't believe so you're saying it every single at every single evolution in tech that you've experienced there's always been those those guys those people who are just like what do you mean i'm not doing my own garbage collection what do you mean i'm not freeing up memory what do you I mean, I'm not, you know.

Matt 23:38

Usually people complained about the back and forth to the cloud. Millisecond-wise was unacceptable.

Nina 23:44

They did that?

Matt 23:45

Yeah, because, like, usually, especially if it was an internal tool, you would run a premise for the company, and everyone had to work in the office back then. And they're just like, the relay back and forth, it's like, could add, at worst, a quarter of a second. And like listen they just find things to hang on to and they'll be like well what if something happens to an Amazon box and it's just like well what if something happens to your box like right, there's still problems there's still solutions and that's what you're there for is to figure it out.

Nina 24:25

Yeah yeah well outside of getting things getting things built results, right? Because now, you're right, we are, it's all about the product. How do you, what do you think is, is, what do you do to, hmm, how would you say you, I'm thinking about the podcast, the previous podcast that I mentioned, but how do you streamline your ideas into gold in a saturated market? And correct me if I've already brought, brought this up, this podcast, but.

Matt 25:01

Yeah. So like, what I think what they call it is, is go to market strategy. And there's like a million ways to do that, and you've got to find the way that you're good at for me it tends to be in person so i try to do conferences, i'd like do speaking engagements also why i'm doing this podcast like impersonal stuff i'm better at than like phone calls i'm really bad at cold email, and seo just doesn't seem to work for a consulting style product, at least for me so it could be another just skill issue.

Nina 25:41

I mean there's ways to you can use agents now for SEO but everybody's using agents now for SEO

Matt 25:48

I mean I'm trying it we'll see and like there's, ads I tried that too with Claude and I was doing it before Claude and I've also done it with agencies and just like what sticks is a little a lot of effort and a little bit of luck, turns out the luck's very important.

Nina 26:09

Yeah, I mean, so that's the, okay, so now that we're breaking down like using AI to help with like marketing or cold emails or SEO, everybody's doing it, right? So if everybody's doing it and everybody has their app optimized best for the search engine, oh crap right like

Matt 26:34

No because i just described like seven of them that i i personally failed at, like there's no way there's no reason they shouldn't work other than me.

Nina 26:44

I could i could try to do your search engine optimization and stuff

Matt 26:48

I've done it but yeah like so like maybe yeah so that's just like somebody else finds their thing that they can get Claude to do okay and make it work right so, it's it's not like uh everyone's got the same skill now, because that's not happening that's definitely not happening from what i see i still see a big difference in skill levels and devs i see big difference in skill levels in marketing and sales.

Nina 27:18

Okay so when you did the search engine optimization amongst like the other stuff were you using an ai tool to like help you get it right at the time like to help or was it

Matt 27:28

Before yeah i was doing it before as well and i did it more after with Claude.

Nina 27:35

And Claude did not did not help help with it

Matt 27:39

It did yeah i was going up like a hundred visitors more a month and now i'm, going up like 250 a month after some things Claude did.

Nina 27:51

Okay. Well, on my website, I do get traffic, but it's not search engine optimized. I think it's just from another platform, but I'm wondering if I could help get some traffic through.

Matt 28:11

Yeah, you could, but to me, it won't matter unless the traffic leads to client calls.

Nina 28:19

Right. Since being on LinkedIn, it's gone incredibly up. So I don't think it's coming from the entertainment part of my life. I think it's coming from the software part. So maybe, maybe, but yeah, I can help out with all that stuff. But like, yeah, but my, my point is, is if everybody's doing it, it's just harder to stand out. So it's like, how do you stand out? You have to be somebody who makes maybe software for a particular realm or I don't know if you've heard the term attention economy.

Yeah. Wait, what?

Matt 29:00

Yeah, but this has been true since the 90s.

Nina 29:03

You're right. You're right. But nowadays, attention economy stuff is like gimmicks and just weird, weird content that you see when you're scrolling on LinkedIn. Now LinkedIn is strange. I don't know if you've been on LinkedIn lately and seen.

Matt 29:22

Yeah, it's funny. I started doing LinkedIn like two years ago, and now it's just useless.

Nina 29:29

Well, I could do it. I could help with that because I'm going to start sharing stuff. But it's like memes and stuff.

Matt 29:36

Yeah that's i give enough i'm just like it's just Facebook for businesses now i have no interest in this.

Nina 29:41

But it seems it seems like that helps

Matt 29:45

I mean yeah i don't i don't know what it's helping with is like i don't see how that's leading to leads for like any business i'd see like people still do messaging on LinkedIn and that seems to work, like instead of doing cold emails you do called messages on LinkedIn with like some basic like engagement with them before you send out the message right and like hey so you posted this and blah blah blah.

Nina 30:15

Well leaning into video format for like a little tech demo of some of the products but have like a gimmick at the start to grab their attention and then once they're hooked just show the product and maybe I don't know

Matt 30:30

It's like guess they haven't it's weird.

Nina 30:33

I've i've been brainstorming how to even even do this for my for myself like when it comes to recruiters like i'm working i'm working with five recruiters right now but i haven't posted on LinkedIn it's just been you know it wasn't like an organic like fishing and pulling them in

Matt 30:54

Yeah i mean but that's also that's not another go-to-market strategy i maybe for product being yourself, you're doing an affiliate deal with these recruiters and they get a cut of your offer that comes in.

Nina 31:05

Right, right. I don't know all the business terms.

Matt 31:12

Yeah. They all apply.

Nina 31:14

They all apply.

Matt 31:15

Because business people kind of know what's going on in businesses.

Nina 31:19

They better.

Matt 31:23

Usually.

Nina 31:24

Usually. But yeah, that's, that's like another, that's another part. Maybe, you know, that's, that's a part of my, my personal dev life that I think needs refinement is like the business side. I make, I make so many apps before AI, I was making so many apps and putting so much time in, but I never shared a single one. Yeah they were just for me and it's like i don't know why that is but i think i think it's going to change this year this year is is the year i'm going to start putting out apps and trying to attract i'm trying i'm gonna fish i'm gonna fish and see what i can catch online because i don't know i just did it for the love of just making things but that doesn't always just because you got something good it doesn't get you anywhere you know if you're not putting it out there so build software you can still build software you don't have to write code I'm looking at one of your quotes from a previous podcast where you said I've never been one of those greybeards what did you mean?

Matt 32:40

I don't think I actually said that but.

Nina 32:42

Wait you didn't?

Matt 32:43

No maybe Claude's hallucinating or maybe I did I also never remember those types of things.

Nina 32:50

I'll have to look at the transcript to see if see if I swear I remember you saying that

Matt 32:59

I swear it sounds like me though also yeah yeah, but like I was talking about the people who just are all about the code like to me it's not I'm not doing this as an art form doing this as an economic way of life Yeah, yeah.

Nina 33:18

Yeah, now. Well, okay, you're right.

Matt 33:20

Like, if you want to take a picture of something, it's way easier to use the camera than to paint it. Like, people still like painting.

Nina 33:29

That's how I felt about painting. I got out of painting because I was just like, I'm just a human printer. Maybe I should get into code. Oh, that's a good idea. That's what happened. And I, I, no, seriously, you go, you, you go to art school and everybody's very artsy fartsy and very like kind of how I'm being towards code. It's like, honestly, and you can, you got to a certain point where you could do something really well. And, you know, like, like just drawing as realistically as possible. Like that was my goal.

And then, yeah, I, I was like, okay, I've done this. I'm bored. I don't feel excited by this, and that's how I got into code, because I wanted to constantly be thinking and refreshing and problem solving. But in 2026, I need to remember that the problem solving has changed. It's not about the code. It's about the end result, and it's about how it can apply. It could be applicable somewhere. So not to get artsy-fartsy about code, but I guess I am being one of those in this day and age.

So because we can build things so fast, like in this podcast, This guy, so Michael was reacting to another developer who was saying, I was ripping through tokens till 2 a.m. And he looks at the camera and he goes, I got a wife and four kids. What am I doing? And I guess, like, how do you not fall down that hole of, because you can make things so fast, and you could probably build things, I mean, you can build Rome in a day. How do you not, how do you take care of yourself, you know, like from burning through tokens, you know, getting, getting into this state where you just keep adding features, making sure everything's MVP for whoever you're working for.

Like what's, how do you set boundaries with AI if you need to at all?

Matt 35:47

Yeah. I mean, it's been the same before. Like for me, it's always been, there has to be a monetary outcome to it. So like just coding in a corner isn't going to work for me like yeah i do like would do pre-sales or like do marketing and do like kind of out research like that, like at least after doing some coding and so just that need that interaction would that requires time you can only be in a conference for so long, you can only do so much seo updates before you have to sit and let it marinate in Google, and then you can do some code stuff like in the beginning when like Opus 5 first came out and fable 5 like want to lose nights of sleep but I'd code a little bit harder getting hyped up about it but, not like regularly.

Nina 36:49

Oh, okay. Okay. Well, it seems that this YouTube video I found is like a very particular type of dev that I think.

Matt 37:00

Yeah. ThePrimeagen also has this video where he does like 72 hours. They rented out a lighthouse and him and 10 dudes try to program a game and they fail miserably.

Nina 37:13

They failed because they have 10 people, not because it's...

Matt 37:17

I don't know if it was actually, it was a lot. I'd say it felt like 10 people. I didn't watch the whole thing. But they had like two shifts going, like going around the clock. Wow. The startup bros, it's like what they live for.

Nina 37:32

I don't know. That sounds kind of fun. Like, that sounds really chaotic, but I don't think somebody could do that with 10 devs. I think maybe four, four, three. But 10 devs, you just, why?

Matt 37:46

I mean, I guess they'd have five on five shifts.

Nina 37:52

I see on shifts. Okay, that makes more sense. Even so. Even so. Yeah, have you watched him previously? Any of his?

Matt 38:01

Yeah.

Nina 38:04

Never seen the 72 hours one, but that's interesting. Yeah, so it's more about the business side for you and programming to get that end result. But that's a good thing. That is a very good thing, I think, that I'm trying to also figure out how to do and how to do business strategy. Well, I can ask Claude. I can ask Claude about GTM strategy. I can ask it about marketing and stuff like that. So, yeah. But did you say you built anything this week, or was it just kind of like a...

Matt 38:37

I wouldn't say build. Yeah, no, just use different things. I'm kind of doing my invoices now.

Nina 38:44

Mm-hmm. That's right. You said you did the invoices. I i've used it to read a a severance document i've used it to read a what else have i used it to read um

Matt 39:02

What did it tell you from the severance document was it just like this is bogus.

Nina 39:08

No it no it said everything looks about average and you and it was like clarify on maybe one or two things I read them over and I decided I wasn't going to clarify on them.

Matt 39:20

I think it's legally required to tell you to clarify two things. That's what it always says on PR reviews now.

Nina 39:26

Wait, really? Is that like a...

Matt 39:28

It's like that's a pushing back module that seems to be injected into it recently.

Nina 39:35

Recently? I think I was on Opus when it did do that. Said take take a look at these two lines ask to see if they're required to do a b and c because this contradicts this you might not get paid or something and i i looked it over i was like that's not what i'm reading here so i ended up reading it anyways so

Matt 39:58

Yeah it seems to crush straws for two to three pushbacks if it doesn't have any it's something i'm fighting right now in the pr review one.

Nina 40:08

Yeah yeah okay well that's good to know now and i've used it to to web scrape all my Amazon purchases for taxes so i've used it too

Matt 40:19

Because it was really good it's just crunching data.

Nina 40:24

It's good at putting stuff in a CSV and doing the math from the CSV without ever having to type a CSV like function or, you know, that would take so much time for like bookkeeping or budgeting. Having to do like the little, what are they called? I don't know what they're called. They're math functions in Excel, but I don't know the actual word for them. Yeah, those, yeah, terrible. I have to look them up every time, too. But yeah, Claude's great for that. It also... Do you have a harness of some kind? Like a workflow thing?

Matt 41:07

Nah. Oh. I mean, I do when it's like an actual application, but just for myself. I usually have scripts, and then I tell it to use scripts. I don't have workflows. Workflows, I more think of like a code pipeline. that has steps that it calls out to Claude for something to bring back, and then it moves to the next step in a workflow.

Nina 41:31

Yeah, yeah. So what I was going to build was kind of like a customizable generative flow chart where you could put agent instances in different parts of the charts and then link them together in a certain order to execute. It's pretty much like every, it's like every workflow app.

Matt 41:51

Yeah. And in our Hermes.

Nina 41:53

Yeah, but it was going to be money.

Matt 41:59

Just make a money, don't you?

Nina 42:02

I think it might be a subconscious thing at this point. Maybe. I don't know. There have been so many times where I've been like, just use this. And I've been like, no, I'll code my own. It was the same when I made a game engine where, you know, someone was like, why don't you just use Unity in Unreal? And I was like, I feel like I'd do it better if I just knew exactly how it was programmed. And then it was just lightweight and for my use case. And so I made a raycaster and an engine for 3D development, 2D, and 3D.

It was WebGL for true 3D, raycasting for efficiency for having raycasting or 3D in the browser, and then 2D. And so I made that engine. there was an interpreter there was an interpreter for code to then have it be read by like a game life cycle that was on the back end and then it was insane like I spent like five years on this thing and then what did I do with it?

Matt 43:11

Yeah, nothing.

Nina 43:13

Nothing. But you know what Matt, I had thought.

Matt 43:17

I mean, so yeah, so you weren't trying to make money. Like you gotta be clear with your goals and when you're doing this stuff what you're doing.

Nina 43:24

Yeah i yeah i don't know it's um yeah it's now that if i could do it over again i do have a clear goal goal i made assets for a horror game and i'll probably bring that to light when i stopped juggling the other plates but yeah having clear goals Have you ever taken that Myers-Briggs test thing? The Myers-Briggs personality type thing?

Matt 43:48

Is that the four letters one?

Nina 43:50

Yeah, it's the four letters one.

Matt 43:51

Yeah, a few startups have made me take that.

Nina 43:55

Really? They make you take that?

Matt 43:56

Oh, yeah. It's like their favorite team-building exercise, I feel like.

Nina 44:01

Oh. Well, there are these two types in Myers-Briggs that you find in developers. That I think this is just an interesting thing to bring up. INTJ and INTP. An INTJ is all about the end results and getting to the goal as efficiently as possible, and they are farsighted in their goals. An INTP is like a mad scientist who just loves the journey or doesn't even know where they're going. And I'm an INTP. And I wish I was more of an INTJ but you actually find that those two personality types make up like a majority of developers which is very interesting I can't remember the study I read but but yeah the INTP has like the creativity and they're systems thinkers and they just need the guidance of an INTJ if you find two of them on a dev team apparently they're meant to be a good pair

Matt 45:09

Uh-huh yeah i think i was the j1.

Nina 45:12

Are you the j1 yeah that that makes sense because you you are you're looking forward at like okay well what's the point of this yeah i've got i have books on it i've studied how to be more like an intj so i like no seriously so i could make money no really really it's like a thing it is a it is an actual thing where where there is one personality type that is like the Einstein, and then the other one that's like the Musk.

Matt 45:43

Ugh, don't do that.

Nina 45:45

Don't do that, I know, I'm sorry. Not the musk, the gates. Just more the name and end goal person.

Matt 45:53

Just any other sociopath in tech.

Nina 45:57

Okay. Yeah, let's not talk about that guy.

Matt 46:06

Yeah, but I get what you're saying. It's just, you gotta... I mean, would you even be happy if you moved to the other side? You had to look at those kind of goals, too.

Nina 46:17

Yeah, because I'd be focused, and I'd have something completed. You know, something completed. Because there's only so much you could experiment and learn and explore before you have to have... There has to be a means to the end there, you know? So, yeah, I would.

Matt 46:39

Yeah. Don't change your personality because I think it's a good idea.

Nina 46:47

It's really funny, too, when you look at income. Last tidbit about those people. No, really, really. No. So INTPs, if you look at which personality types make the most money, no matter what career they're in, INTPs and INTPs, they're all the way at the bottom. Because they have a hard, they, unless they're in tech. And then on the other side, all the way at the top for the highest income bracket, you will see ENTJ and INTJ. So they're similar personality types, like very similar. And they're both fitted for tech, but one's at the bottom of the bracket, one's at the top. And I really do think it has something to do with like that aspect.

Matt 47:31

When I saw it, I just thought one was sociopaths and the other one were con artists.

Nina 47:37

Con artists?

Matt 47:38

Yeah, those are the empath ones.

Nina 47:41

Oh. No, the INTP is like the where's my classes nerd, more of a machine type person. I think there's a joke that INTPs are the warmest machine and INTJs are the coldest human. There's a joke about that.

Matt 47:59

No the ENTJs are the con artists and the INTJs are sociopaths.

Nina 48:05

Yes yes yeah the INTJs are the magnetos they're the villains but they make money so it's true yeah well

Matt 48:20

So they're becoming supervillains.

Nina 48:22

Yeah we'll become supervillains yeah exactly

Matt 48:26

They know what's going on.

Nina 48:28

Yeah they they know they know look look at that the octopus guy from spider-man i forgot his name uh Doc Ock he makes money right i

Matt 48:37

Think he almost solved energy crisis as well he was helping people.

Nina 48:43

Lex Luthor that guy's that guy's loaded

Matt 48:47

Yeah we can do it.

Nina 48:49

Yeah Peter Parker not that much he's not yeah

Matt 48:54

Making pictures in an apartment.

Nina 48:57

And he can't make up his mind between science and photography and see Satan in the world that's goofy

Matt 49:08

Alright I think on this note I'm gonna have to end I got a hard cut now.

Nina 49:13

Yeah um

Matt 49:15

Closing remarks are we good.

Nina 49:18

No we can have a closing remark

Matt 49:19

I got nothing. So I hope you got something.

Nina 49:22

Planning is everything and having goals is everything. Or else you just meander through life. That's it.

Matt 49:28

Yeah, I'll take it. That's a good one.

Nina 49:30

Yeah.

Matt 49:30

And that's the bet.

Nina 49:32

All right.