WEBVTT - Monologue: AI Isn't Replacing Software Companies, Calm Down About Claude Code

0:00:03.200 --> 0:00:07.560
<v Speaker 1>Media. Hello and welcome to this week's Better Offline Monologue.

0:00:07.600 --> 0:00:17.880
<v Speaker 1>I'm your host ed ze trn SO. In the last

0:00:17.880 --> 0:00:20.480
<v Speaker 1>few weeks, there's been a dramatic sell off in software stocks,

0:00:20.560 --> 0:00:23.400
<v Speaker 1>driven by the anxiety that companies will, instead of paying

0:00:23.440 --> 0:00:27.200
<v Speaker 1>for someone like Salesforce or Microsoft, simply build their own software.

0:00:27.880 --> 0:00:31.200
<v Speaker 1>It is a genuinely stupid assumption based on analysts and

0:00:31.240 --> 0:00:34.159
<v Speaker 1>reporters that simply do not care about the truth. In

0:00:34.200 --> 0:00:37.599
<v Speaker 1>their mind, one can simply type build me Salesforce now

0:00:37.600 --> 0:00:40.080
<v Speaker 1>into claud code and have it barf out an identical

0:00:40.159 --> 0:00:44.360
<v Speaker 1>functional clone that's compliance secure, stable, or because somebody was

0:00:44.400 --> 0:00:46.040
<v Speaker 1>able to bunk it on the head enough times to

0:00:46.040 --> 0:00:49.000
<v Speaker 1>spit out something that sort of looked like a tool

0:00:49.080 --> 0:00:53.239
<v Speaker 1>like Trello maybe, or a personal website. Look, when you

0:00:53.320 --> 0:00:56.560
<v Speaker 1>pay a software company, even a dogshit mediocre one, a

0:00:56.600 --> 0:00:59.680
<v Speaker 1>monthly fee, you're not just paying them to access the software,

0:00:59.720 --> 0:01:01.880
<v Speaker 1>but to take away the burden of maintenance that comes

0:01:01.960 --> 0:01:05.800
<v Speaker 1>with running a software company. Minor things like currency changes

0:01:05.959 --> 0:01:09.480
<v Speaker 1>or time zone shifting can cause major problems, and systems

0:01:09.520 --> 0:01:12.680
<v Speaker 1>that aren't built with intention, you know, like something in LLM,

0:01:12.720 --> 0:01:15.560
<v Speaker 1>would spit out and things get even more complicated when

0:01:15.560 --> 0:01:18.520
<v Speaker 1>you start connecting other systems of record like billing or

0:01:18.520 --> 0:01:21.760
<v Speaker 1>a customer's personal information, especially if they're in Europe. By

0:01:21.800 --> 0:01:24.480
<v Speaker 1>the way, plan to have any of that information actually

0:01:24.520 --> 0:01:27.440
<v Speaker 1>connect with a customers systems or you're gonna need a

0:01:27.520 --> 0:01:29.440
<v Speaker 1>sock two order and you're definitely going to need to

0:01:29.440 --> 0:01:31.800
<v Speaker 1>make sure it's got rock solid security so that nobody

0:01:31.840 --> 0:01:34.240
<v Speaker 1>can swipe all of that data and then you get sued.

0:01:34.640 --> 0:01:37.200
<v Speaker 1>I also assume you're going to effectively take an engineer

0:01:37.240 --> 0:01:40.200
<v Speaker 1>off of one of your team's probably for good, to

0:01:40.240 --> 0:01:44.280
<v Speaker 1>be honest, to maintain your new internal salesforce Monday Microsoft

0:01:44.280 --> 0:01:47.760
<v Speaker 1>three sixty five and Notion clones you Trello cu clone

0:01:47.760 --> 0:01:50.240
<v Speaker 1>as well. Probably you're a sana What I mean, how

0:01:50.320 --> 0:01:53.040
<v Speaker 1>much software are you gonna build? Good? Thing? You've got

0:01:53.120 --> 0:01:55.680
<v Speaker 1>clawed code to help speed that up, right, Just make

0:01:55.680 --> 0:01:58.400
<v Speaker 1>sure you read everything it writes, because every little fuck

0:01:58.480 --> 0:02:01.320
<v Speaker 1>up just became your problem, and you've got nobody to

0:02:01.400 --> 0:02:04.360
<v Speaker 1>scream at, because while your company is saving fifteen dollars

0:02:04.360 --> 0:02:07.080
<v Speaker 1>per seep per month, you've also fired the people whose

0:02:07.120 --> 0:02:09.320
<v Speaker 1>job it is to make sure your nasty little software

0:02:09.320 --> 0:02:13.120
<v Speaker 1>subscriptions actually fucking function. And while you may fear that

0:02:13.160 --> 0:02:15.079
<v Speaker 1>a boss might try and force this down your throat

0:02:15.120 --> 0:02:17.600
<v Speaker 1>to save money. I must be clear how impossible this

0:02:17.720 --> 0:02:21.120
<v Speaker 1>task is. Even the most annoying, frustrating suffer as a

0:02:21.160 --> 0:02:24.880
<v Speaker 1>service contract protects you from the grueling, underlying maintenance and

0:02:24.919 --> 0:02:28.000
<v Speaker 1>infrastructural bullshit that goes into making sure the thing you're

0:02:28.040 --> 0:02:30.920
<v Speaker 1>paying for actually loads and functions wherever you load in

0:02:31.320 --> 0:02:33.200
<v Speaker 1>and even on the browser that you want to load

0:02:33.200 --> 0:02:36.760
<v Speaker 1>it in. In most cases, thank you riverside. The people

0:02:36.800 --> 0:02:39.880
<v Speaker 1>pushing this narrative are either fundamentally disconnected from how the

0:02:39.919 --> 0:02:44.800
<v Speaker 1>world works or actively incentivized to mislead you. I've seen

0:02:44.880 --> 0:02:49.120
<v Speaker 1>this narrative propagated on multiple different business television shows and

0:02:49.160 --> 0:02:53.120
<v Speaker 1>supposedly respectable outlets, and it makes me genuinely worried that

0:02:53.160 --> 0:02:56.960
<v Speaker 1>we don't have a media industry prepared to dissect fundamentally

0:02:56.960 --> 0:03:01.160
<v Speaker 1>deceptive narratives. Just because it's possible for a non CODA

0:03:01.200 --> 0:03:04.200
<v Speaker 1>to cobble together a website that looks near identical to

0:03:04.240 --> 0:03:07.120
<v Speaker 1>a model's training data in the space of an hour

0:03:07.440 --> 0:03:10.920
<v Speaker 1>doesn't mean that we're replacing every software company, nor is

0:03:10.960 --> 0:03:14.239
<v Speaker 1>its ability to do so. Any indicator that will be

0:03:14.280 --> 0:03:17.280
<v Speaker 1>able to do more than that in the future. I

0:03:17.440 --> 0:03:20.320
<v Speaker 1>plead with the media, please stop filling in the gaps.

0:03:20.639 --> 0:03:24.480
<v Speaker 1>Please stop seeing every incremental improvement as proof of whatever

0:03:24.600 --> 0:03:28.000
<v Speaker 1>marketing slot Warrio Ama Day or Clammy Sam Altman is

0:03:28.040 --> 0:03:31.520
<v Speaker 1>trying to cram down your throats. You're being played, You're

0:03:31.560 --> 0:03:35.200
<v Speaker 1>being conned, and by extension, you're conning your listeners, your readers,

0:03:35.240 --> 0:03:37.960
<v Speaker 1>and your viewers. And once the bubble pops, I believe

0:03:38.000 --> 0:03:41.840
<v Speaker 1>they will demand an explanation from you. I certainly will,

0:03:42.600 --> 0:03:46.280
<v Speaker 1>and really, in this era, I think people are under

0:03:46.440 --> 0:03:49.080
<v Speaker 1>selling how big a reckoning there will be when the

0:03:49.120 --> 0:03:53.040
<v Speaker 1>bubble collapses. How are we meant to trust anybody who

0:03:53.360 --> 0:03:57.640
<v Speaker 1>vociferously pushed this or who got obsessed with AI once

0:03:57.680 --> 0:03:59.880
<v Speaker 1>all of this falls apart. When I say falls apart,

0:04:00.200 --> 0:04:03.000
<v Speaker 1>I mean that the current rates that you are paying

0:04:03.280 --> 0:04:06.520
<v Speaker 1>are not even close to being sustainable. I'm currently working

0:04:06.600 --> 0:04:08.400
<v Speaker 1>on a piece called the Hater's Guide to Anthropic on

0:04:08.400 --> 0:04:12.120
<v Speaker 1>my premium newsletter, and I found a mathematical study that

0:04:12.200 --> 0:04:16.240
<v Speaker 1>found that on a twenty dollars a month CLAWED subscription,

0:04:16.680 --> 0:04:20.760
<v Speaker 1>you can spend over one hundred dollars of anthropics API tokens,

0:04:20.960 --> 0:04:23.600
<v Speaker 1>and on the two hundred dollars a month about two

0:04:23.720 --> 0:04:28.320
<v Speaker 1>three hundred to seven hundred dollars. You got to I

0:04:28.320 --> 0:04:31.279
<v Speaker 1>guess it takes money to lose money. But this is

0:04:31.520 --> 0:04:34.599
<v Speaker 1>the reality of the AI bubble. Everything you were using

0:04:34.680 --> 0:04:39.359
<v Speaker 1>is being subsidized. It's being subsidized by these companies in

0:04:39.400 --> 0:04:41.400
<v Speaker 1>the hopes that it kind of bakes into your life,

0:04:41.440 --> 0:04:44.800
<v Speaker 1>except it's not good enough to do that. The frenzied

0:04:44.920 --> 0:04:49.680
<v Speaker 1>media push under clawed Code existed to kind of beguile you,

0:04:50.000 --> 0:04:52.200
<v Speaker 1>to make you think that there's more happening than a

0:04:52.240 --> 0:04:54.000
<v Speaker 1>tool that might be able to kind of build a

0:04:54.000 --> 0:04:56.960
<v Speaker 1>website sort of or build you some half functional software

0:04:57.000 --> 0:04:59.719
<v Speaker 1>that maybe works sometimes, and when it comes to actually

0:04:59.720 --> 0:05:02.120
<v Speaker 1>building that into something you could sell your shit out

0:05:02.120 --> 0:05:04.360
<v Speaker 1>of luck. You actually do need to code. And even

0:05:04.560 --> 0:05:08.160
<v Speaker 1>the things that you vibe code weren't really secure or functional.

0:05:09.279 --> 0:05:13.039
<v Speaker 1>And even that mediocre web slop you're seeing spooched out

0:05:13.040 --> 0:05:17.240
<v Speaker 1>by claud code is heavily subsidized. If people had to

0:05:17.240 --> 0:05:20.120
<v Speaker 1>pay the real rates, those people you see jacking off

0:05:20.160 --> 0:05:22.640
<v Speaker 1>on Twitter about claud code, they'd be paying two hundred

0:05:22.680 --> 0:05:25.680
<v Speaker 1>plus dollars a day. Do you think they'd actually pay that?

0:05:26.040 --> 0:05:28.680
<v Speaker 1>Do you think that's actually happening because I'm sick and

0:05:28.720 --> 0:05:31.200
<v Speaker 1>fucking tired of hearing all the people rambling about claud

0:05:31.240 --> 0:05:34.400
<v Speaker 1>code writing all the goddamn code. I'm sick and tired

0:05:34.440 --> 0:05:36.440
<v Speaker 1>of it because I don't see what the end point is.

0:05:36.800 --> 0:05:40.160
<v Speaker 1>I don't hear an actual result from this. I think

0:05:40.160 --> 0:05:42.919
<v Speaker 1>people should be laid off. But if this was doing

0:05:42.960 --> 0:05:45.200
<v Speaker 1>the thing that they were saying it would that al

0:05:45.240 --> 0:05:47.960
<v Speaker 1>would be happening, and that would be definitively connected to AI,

0:05:48.200 --> 0:05:51.640
<v Speaker 1>or more software would be being shipped. That's actually good.

0:05:52.520 --> 0:05:56.680
<v Speaker 1>Just more software existing doesn't mean the software's functional or useful.

0:05:56.920 --> 0:05:58.960
<v Speaker 1>And if you look at VIBE coded apps, they all

0:05:59.040 --> 0:06:01.560
<v Speaker 1>kind of look the same. And that's the result of

0:06:01.240 --> 0:06:04.719
<v Speaker 1>the more using the same training data. And it's why

0:06:04.960 --> 0:06:07.919
<v Speaker 1>these things are not really good at building nuanced or

0:06:08.440 --> 0:06:11.400
<v Speaker 1>unique software, because that's not what they do. They copy

0:06:11.480 --> 0:06:14.040
<v Speaker 1>software that's already been written, and they do so with

0:06:14.120 --> 0:06:18.279
<v Speaker 1>no intention, no real plan, and nothing you can look

0:06:18.320 --> 0:06:20.560
<v Speaker 1>back on and say, oh, right, that's why they built

0:06:20.600 --> 0:06:23.280
<v Speaker 1>it this way. And perhaps that's not a problem as

0:06:23.320 --> 0:06:25.599
<v Speaker 1>you're a solo person building your own deadly little lap,

0:06:25.800 --> 0:06:28.120
<v Speaker 1>but expand that to one thousand people or ten thousand

0:06:28.200 --> 0:06:31.760
<v Speaker 1>people let's say five ten years passes, and let's take

0:06:31.800 --> 0:06:35.720
<v Speaker 1>one of these what I think are bullshit statements about how, oh,

0:06:35.839 --> 0:06:39.120
<v Speaker 1>ninety percent of our code's written by AI. Great, what

0:06:39.240 --> 0:06:41.080
<v Speaker 1>happens in three years or four years when you go

0:06:41.120 --> 0:06:43.120
<v Speaker 1>and look back and that, you go, shit, why did

0:06:43.120 --> 0:06:46.320
<v Speaker 1>we do that? Are the person left or the person

0:06:46.320 --> 0:06:48.800
<v Speaker 1>didn't leave. They've just been drinking heavily and they don't

0:06:48.800 --> 0:06:50.919
<v Speaker 1>really remember what they were prompting the model with. And

0:06:51.320 --> 0:06:53.440
<v Speaker 1>oh man, how do we actually see why they built

0:06:53.440 --> 0:06:56.599
<v Speaker 1>this or what the reason was. This becomes a huge

0:06:56.600 --> 0:06:59.040
<v Speaker 1>problem when things start breaking, when you have to make

0:06:59.080 --> 0:07:01.599
<v Speaker 1>things more efficient, and then people will say, oh, we

0:07:01.720 --> 0:07:05.240
<v Speaker 1>use AI to fix them. At some point, you're just

0:07:05.360 --> 0:07:08.679
<v Speaker 1>looking at a kind of a rat's nest of crap

0:07:09.560 --> 0:07:16.240
<v Speaker 1>of code written unintentionally spooched out meaningless. How do you

0:07:16.280 --> 0:07:19.560
<v Speaker 1>fix that long term? How do you build an efficient

0:07:20.000 --> 0:07:22.440
<v Speaker 1>software package out of that? How do you build software

0:07:22.480 --> 0:07:24.840
<v Speaker 1>that continues to run well into the future. Is the

0:07:24.880 --> 0:07:28.400
<v Speaker 1>idea that you just build AI on AI? God? No, anyway,

0:07:28.440 --> 0:07:30.840
<v Speaker 1>I think once the real costs are charge, people are

0:07:30.840 --> 0:07:34.119
<v Speaker 1>gonna drop this shit like a bad habit. I can't wait,

0:07:34.280 --> 0:07:36.000
<v Speaker 1>and I'm gonna be smug about it. I'm gonna be

0:07:36.000 --> 0:07:39.280
<v Speaker 1>annoying about it, not gonna lie. But I think based

0:07:39.280 --> 0:07:42.360
<v Speaker 1>on the YouTube comments, you're all gonna join me. Now.

0:07:42.440 --> 0:07:44.600
<v Speaker 1>Next week we've got more Hater season, and I'll be

0:07:44.640 --> 0:07:46.960
<v Speaker 1>back with one of these monologues. Maybe orerhaps I'll give

0:07:46.960 --> 0:07:50.000
<v Speaker 1>you multiple interviews, done some weird ones. Got Victorious Song

0:07:50.080 --> 0:07:53.320
<v Speaker 1>on wellness, Got Adam Becker on Epstein. I mean, got

0:07:53.360 --> 0:07:56.560
<v Speaker 1>all sorts of shit coming. Looking forward to bringing you more,

0:07:56.800 --> 0:07:57.680
<v Speaker 1>and I love you all