AI Mania Is Eviscerating Global Decision-Making w/ Nik Suresh

0:02Speaker 1:Also media greeting, Hello and salutations.

0:06Speaker 2:I am at Zetron and this is better offline.

0:20Speaker 1:We are back. You need to subscribe to the newsletter, you need to download the podcast you're already listening to take your friend's phones downloaded on there. You will have failed me if you don't do this, not really, I love you all dearly. But look today I am joined by the incredible writer and consultant Nick Siesh, who two weeks ago put out one of the best pieces on AI I have ever read, and not just because it confirms my biases and it's called AI mania, is a viscerating global decision making Nick, thank you for joining me.

0:47Speaker 3:Hey, thank you so much for having me on again.

0:49Speaker 1:This piece was terrifying. I just want to be really clear that, like the things that I read in here a lot of them, I was like, I assumed this was happening. I assumed that it was the bad. But to summarize the piece in short, it appears that just AI is this bordering on religious thing within business now that people get into, regardless of whether it even does anything.

1:16Speaker 3:Yeah, yeah, and you know, I think the thing that got it so widely circulated is even that people who use AI in their work and say they're seeing benefits. A lot of the more sensible people are looking at their colleagues and going, what are you guys doing? Right? Like, I think I'm a little bit faster at my job, And here we have my CEO saying, let's just scrap every other project. We'll remove all the employees. We're gonna have one hundred x productivity. And even sensible people that enjoy it are getting a little bit worried by the level themania's gotten to.

1:48Speaker 1:Yeah, so give me, give me an example here. The first section is called AI investments are generally total failures. And I realize the title really does say it all there, but can you break that down for me? Yeah.

2:00Speaker 3:So it's interesting because you know, most projects at companies do fail, especially big companies. They're just bad at running tech projects. I've written about this well before the AI thing was big. What is fascinating is that they have an even higher failure rate for the AI projects. And again, like just to we don't need to get into an argument with people about whether it works or not, right, we can just start with the basis of normal it projects fail a bunch, right, AI projects are novel technology.

2:30Speaker 3:So these are failing even more. And the protective factor at a companies normally that they don't want to spend all their money on it. They actually want to spend the smallest amount of it possible. Right, it's mostly a thing you're it's a cost for the business. Right With AI, this is totally inverted, and companies are going, I'd like to take as much budget as I can feasibly get, and I'm going to put it all into this thing that I know I'm bad at, and I want to do a harder version of that, but I can't admit to anyone I've just wasted millions and millions of dollars.

2:58Speaker 3:So it's sort of like all the organizations in the planet are going all in on the thing they know they're bad at, playing it on hard mode, and then sort of just lying about it after the fact.

3:08Speaker 1:And when you say lying, what do you mean that? They just being like, Yeah, this is making me one bazillion more productive percents. I'm so good at AI. I love it. Like what is that? Like? What are they doing? What are they actually do? Get this crap?

3:25Speaker 3:Well, okay, so it happens in a couple of ways. Right when I say lying, that sounds like very intentional. And there are people that lie, because there are people that lie everywhere, but usually it's two things. Like internally it can be a lot of delusion. There are a lot of executives out there who have funded AI projects and they don't realize that the whole organization has sort of shifted into presenting them with the rosiest view of what is happening that is possible. Like if you're a CEO who has heard good things about AI, and let's say you've got two guys underneath you gunning for promotion, Right, those two guys are now at an arms race to talk about their AI productivity and how much AI they're rolling it across the business, Like that is the goal.

4:08Speaker 3:It's not to make money, it's to be the most AI, whatever the fuck that means. Right, So it's pretty nonsensic, But that's that's self creation. It's it's very so good much And again it's it's an extreme version of what we already see, right, because people already lie to CEOs for promotions. This is just that on steroids.

4:29Speaker 1:So you mentioned in here you say you've seen zero percent success in a year and a half. Really is it really zero? You've seen nothing? I just want to be my owner is going to eat this up, I realize, but like I just want to make sure, yes we should you know, add one asterisk there, because some people misunderstood that me seeing zero percent doesn't mean like it's zero percent across the whole world, because I haven't seen every project in the world. Sadly, some people don't have enough reading comprehension for that, so you have to lay it out or you get like an AI boo who's like, it's not literally zero. Some must be getting some results somewhere.

5:05Speaker 1:I have seen nothing, and I've been doing nothing but sales and on sites at companies like all across Australia in the US for a year and a half. Now I don't get brought in to do them.

5:15Speaker 3:They're like things I see on the side and I go, well, that's that is not working.

5:20Speaker 4:Well.

5:20Speaker 1:The thing I think about when I hear that people like, actually, you haven't saw every software thing. It's just this is how AI boosters work. They want you, as the defender of reality, to be like prove unilaterally this is the AI is terrible. If it's so bad, prove it. But when you ask them to prove anything, they look at you and go, well, you know, just the numbers, just the biggest, the number number I made so big this year, It's the biggest I've ever seen. And they refuse to show you anything real. I think I've been sent the fucking Erdos problem thing fifteen times now, and the next time someone sends it to me, I'm gonna draw a gut.

5:59Speaker 1:That is a joke. That's a joke. I'm not gonna not gonna do anything with a gun, but I am just gonna possibly imagine the saw from Jigsaw, sorry, Jigsaw from saur Even they're gonna get me in the comments with that one, just kind of tying them to radiate and say, sure, sure, tell me what you think the Erdos problem means. Tell me what it means, Tell me what you think eros means, because that's the thing. These people don't have anything tangible to share, and your peace kind of I read it in an airport and I was just sitting there going like fucking hell. And I did get a few looks because I was so surprised by it. But it feels like your peace crystallizes everything.

6:35Speaker 1:I've been thinking for two years that just this is this weird death of stylein style situation where people are just saying, God, damn, I love AI because if they don't say it, they'll make their boss feel bad, their customers feel bad, and their colleagues feel bad despite none of them getting anything out of AI.

6:55Speaker 3:Yeah, you know, the simple way to think about it is that any given business, you know, you have one problem generally that is bottlenecking how you're making most of your money, and whatever you think AI can do, it's really only the correct thing to deploy that solves that one problem. It doesn't matter if you write more code, if writing code isn't the way to make more money. And I mean it's basically just like as simple as that, right, people are going all in on a thing that doesn't really make sense sort of for context they're inhabiting.

7:23Speaker 1:So you mentioned within the piece that you had to basically stop offering AI services or a specific one forgive me if I forgive me if I misread that, but you just won't mention AI services within software projects because people go nutrageous over them.

7:38Speaker 3:Yeah. What we found was, you know, we're a relatively small business and we optimize a lot for working with people.

7:44Speaker 1:We entry a software consultancy.

7:46Speaker 3:Just to be clear, Yes, we're a software consultancy and data specifically. So my postgrad degree was a Masters in data science. Like I was into AI nominally before the whole LM thing took off. And yeah, what we found was whenever we mentioned AI, two things would happen. And the one I mentioned in the piece was people would become rabidly interested in buying from us. But they were also really bad customers, Like they made bad decisions, they had cash flow issues, they didn't think clearly, Like, these were not people we wanted to work with as a matter of lifestyle. But also, you know, it's like a lot of startups and stuff, the sort of customer where used to take all the money upfront because you're worried they're just going to scam you, Like, these are not people you want to be in bed with. And the people we really liked, we would mention it just as an aside, and they would immediately become so disinterested because they just go, well, you know a lot of people are trying to scam you with this right now, I'm just not interested. So we just don't talk about it at all in any context. We don't like doing it.

8:49Speaker 3:The work's really boring anyway. But yeah, like it's just here's another thing I can tell you. Our clients that don't talk about AI at all are like far in a way, the most profitable because they're focused on whatever their biggest problem is at any given moment. It's very very simple, Like this is not rocket science.

9:07Speaker 1:Yeah, it's just like it's like the companies that are doing well are focusing on solving business issues. The companies that are doing badly are the ones getting very expensive digital fidget spinners.

9:18Speaker 3:Yeah, like it's it's as simple as that. I don't know many companies where llms unblock, like whatever their core issue is, there's stuff I can think, like maybe customer support, but I don't know much about customer support anything where I have some subject matter expertise. I'm like, an LLM is not the most important thing you need right now.

9:37Speaker 1:I can't think of it with two, three, four boobless years into this bullshit, and I am a spreadsheet hog and it doesn't do spreadsheets very well. I will say, people can be people will be very mad at me for this. I've found one use of llm's one. And I think I've told you about this the Bloomberg terminal, because the way that you the way that you request stuff on the Bloomberg terminal is the three Python. I don't write Python, so you'd be like Bloomberg, l of them tell me out to get this, and it gives you the code and it runs it boom simple. Probably the least you like, literally just sexy search. Well done, that one point one trillion dollars of capex. I'm sure that was all worth there. But what's funny is the example you've given. The piece of the AI chatbot thing for Snowflake feels like the things that LM should be good at, which is a database request, right, Like, yeah, why doesn't it work? Surely that is the thing that this should be good at. What is the issue there and why don't LLLM solve them? Yeah?

10:39Speaker 3:The complexity there is you know, I don't know much about Bloomberg terminals, although it's very funny everything you've told me about.

10:45Speaker 1:If you tell it's an autism portal, it's so good.

10:50Speaker 3:Yeah, So you know, I imagine they're fairly constrained in what they can do. Yeah, For example, you know, you mentioned writing Python, which means someone has written a Python library that interfaces with the Bloomberg terminal and constrains what sorts of things you can ask it. So if it's very well designed and you put all the text in there, the same thing that lets me a human interrogate it with, let an LM do it. What's very hard at a business is a lot of that stuff is just not written down, or it exists in someone's head or the real world is very, very complicated, and the very people who are Let's use this example, right, why does cortex not work? So you go in and you're the CFO and you type in I would like to know much revenue we made last year. It doesn't know all the little nuances of your business, and for something really important you need it to be correct. But if you're making use of an LLM and you're the CFO, you can't debug whatever it put out. And that's okay when the number is totally wrong, because you'll go, hey, this isn't correct.

11:47Speaker 3:If it's off by like five percent or ten percent, you probably don't know and exactly would like with your Bloomberg terminal. Right, if you looked up something really important. You might reach out to a friend and say, is this Python court because I'm about to spend a lot of money on this.

12:02Speaker 1:Yeah. And to be clear, when you make a request on the terminal, it does the Python. It's just the Python code to request Bloomberg's actual data, like it hasn't. I imagine that I don't write Python, but I imagine there is a library in this, that this is just a layer on top of it so that people like me can actually use it without just punching the typing in like three letter things and making things pop up and then spending hours fucking with it, which of course I never did before. I use the LMS side homes of course.

12:28Speaker 3:Yeah, yeah, that's pretty great.

12:30Speaker 1:But with Cortex, the search side, it seems that the problem is also that is it making stuff up? Because I'm not being an ignoramus here. Surely looking over a big data based on information and pulling stuff out is what an l of them is good at. Or is the suggestion that the databased layer is not something ALMS can solve.

12:49Speaker 3:It's just very complicated.

12:51Speaker 1:You know.

12:51Speaker 3:So we were at a client site yesterday and you know they have this ledger internally, which tracks like all the transactions coming in and out, and there's like forty different account types stored in all these different tables, and in different situations you want to use different types of transactions to calculate certain things around like service utilization, and and LLM looks at that and it just goes, well, I haven't worked here for eight years, so I don't know how to run this calculation.

13:15Speaker 1:Like it's so the data is so it's that the data isn't labeled properly and all these others. So it's how the data is presented and looked over. Yeah, yeah, you.

13:25Speaker 3:Can even label it properly. But there's a reason we have human specialists. It's the same reason I don't get an LLLM to write all my contracts. I could, but it's that I go to my lawyer Eed, and I go, what should I actually do here?

13:38Speaker 1:You know, like the Yeah, it's experience versus data.

13:42Speaker 3:Right, you know, Like when we started the business, he wrote a contract for me that you know, it's like a standard one and it has things in there like I wouldn't have thought of, like, for example, the agreement we have with clients on how long we take between email responses and.

13:57Speaker 1:Ye, when I'm still doing PR stuff, that's exactly That's exactly what I had in there, like basic things that you wouldn't think of unless you'd experienced the thing. But it's funny because even now, even people who are anti LLM, I will hear them say, but it's good for searching large databases, and it sounds like it isn't like AI is just not good at that.

14:19Speaker 3:I specialize in this specific area, which is just like big information retrieval. It's pretty good at you know, the a database specifically in tech means something. It means systems like Postcrest or whichever, which means nothing to normal people. I would say they're pretty good at reading gigantic amounts of texts and then pointing a human at the correct place. So a good use is if we're like in a new client site and they have a million lines of some fucking thing, I can go, I need to look for the places where this logic might happen, and they'll point me at them and it's still wrong sometimes, but at least I can look. That's not useless. But it's like how much what I pay for that?

15:01Speaker 1:I don't know?

15:02Speaker 3:Maybe like twenty dollars a month.

15:06Speaker 1:But that's not that. But that's the thing. It's like, even with the terminal, it's the same thing. It will get things wrong. It never gets the numbers wrong because when it runs the numbers, it's pulling them from a specific place. And I still tell it, if I ever touch the LM stuff, run it again, run it again, just like and then I'll google it and I'll look up the stuff because I don't trust anything that touches them. But still, it's just it is a phishing expedition. You can't rely on it to have any logic, which I think is what most people want from search when the human beings are pretty terrible about asking for exactly what they want, and I feel like it just falls down.

15:41Speaker 1:And the reason I keep going on about this stuff is because searching big databases and making decisions feels like ninety percent of the LLLM integrations I've heard about, like, it's just how do I do this? And what does this mean? Is how people are kind of seem to at least be describing how it's useful in their business, but it can't do that, which is bad.

16:01Speaker 3:Yeah, it at the end of the day, as long as we're constrained by having to go in and check whether the results are correct, even if we grant that they're correct most of the time, which is not always true. It's very domain specific. That just means you can't get one hundred times faster. And that's what's justifying the bajillion million infinity dollars of investment. Right if we went, hey, this makes you fifty percent faster, Like, all right, is that worth a trillion dollars? I I don't think so.

16:30Speaker 3:I don't think it's worth trillion dollars, Like the market thesis depends on the psychosis claim.

16:36Speaker 1:Also, I mean it's meant to be artificial intelligence, just to be clear, that's the intelligence part is in the name. But everything was saying makes it sound really fucking stupid, just like it's just it's not smart to do. It doesn't know anything and it does dumb guys stuff. Oh you didn't ask me in exactly the way I needed to be asked. I don't understand, mate. Sorry, I'm not going to be helpful at all. And also, what we are describing is very simple stuff. It's not a trillion dollar, it's not even a fifty billion tam industry. It's just even if it got everything was saying right, it still wouldn't be that exciting.

17:15Speaker 3:Yeah, you know, I think it was. It was very interesting when it came out. It was like pretty novel. When GPT three came out, I went with the hell like this was not available when I was at university. This is fucking wild. I think this really sort of overstate. It's welcome, and it's just so interesting because we've taken a technology that I mean, if you're willing to set aside many many ethical claims around stealing from artists and environmental things, like if we just handwave all of that, which I'm not happy to do, but in this context for an argument, I will.

17:46Speaker 1:Yeah, yeah, if you.

17:47Speaker 3:Handwave all of that, It's like, all right, we've made something pretty neat, and then we've still somehow completely managed to fuck everything up. Like all we had to do was be reasonable for ten seconds about how we use this in advent than it, and we have completely failed. Everyone has gone insane.

18:14Speaker 1:Yeah, it's and I'm trying. It's like if everyone did platform as a service and take it to a church, It's like, Wow, you could just put a single line of code. Now your entire ecosystem is working perfectly on this. Yeah, it sounds like the promise that was made with pass and then just didn't happen. I worked on a few of those clients. It was fucking terrible, explained to journalists. Platform was a service meant. But again, no, everyone didn't suddenly say that no one was running their own computers anymore, no one was running their own databases anymore. And it's just it feels like it's completely escaped reality and you are very much in the trenches with this. And one thing I did want to ask about as well, was this my favorite section of the piece, heretics will be shot.

18:54Speaker 1:And the way you describe it is that there is this genuine religiosity behind it, this this aggressive kind of tech evangelism that says, if you were not in with the AI stuff, you were out of the crowd.

19:09Speaker 3:Yeah, and it's actually kind of scary to watch. Like, normally when I write something, I'm mostly angry about it, and you also wasn't a lot of anger in that piece because I was more wowly realizing. Yeah, I was like, hey, there are not adults running our hospitals, you know. It turns out, I think the worst thing was realizing that they were always sort of immature and childlike and not responsible or brave, like they've always been that way. This has just allowed them to coordinate around a thing that allows them to behave this way at the same time. Yeah, and anyone that does not get on board, like God, I've seen some pretty wild stuff. We're sort of at the point where if you tell as CEO at maybe most companies, especially big ones, right if they're worth over a billion dollars, if you tell certain people, certain executives that you do not believe that AI is the future, and I mean those words right, talking like have a sound business plan or you believe a specific initiative will work.

20:04Speaker 3:If you will not sign off on statements like AI is changing everything, it will transform this business, they will just fire you. I've never seen anything like it. Even you know, we had the cloud a while ago. People still behave the same way. It's like a little bit culty. They go, we've got to move everything to the cloud. They got to move everything to the cloud. I don't think I've seen someone get fired for expressing doubt about that. But if you say something like I don't think AI is the way for me to do my job. CEOs.

20:32Speaker 3:I've seen multiple CEOs say things like, well, then you don't belong at this company, you know, pack your bags, get out of here.

20:38Speaker 1:What the fuck? What's the justification?

20:42Speaker 3:Well, you're against AI, which must mean you're bad at your job, because we won't need people who don't use AI anymore. You know, It's just I don't know what to say, man like, because the question of why implies there's like a coherent reason they've gone insane.

20:57Speaker 1:That's just because you mentioned in here as well that doubting AI also scares their customers and that it's kind of a chain of idiocy if you could talk about that a little bit.

21:09Speaker 3:Yeah, So what was really interesting was I was chatting with a Fortune five hundred executive, right. I thought it was really interesting because some of the you know, two brain cell dudes on hacker use were like, Ah, this this cad be real, and I'm like, I mean, it's it's a Fortune five hundred executive, Like, would you what do you want me to say? It's just it's the truth. And this guy was pretty smart, but his company had said some pretty crazy things. So well, you know, do you like, between you and me, we're sitting here having coffee, is it true or do you believe that? Or you know what is making you say this? Is it just because you need to hit certain sales numbers? And it's like, yes, some people are very gullible and they buy in to this stuff. We don't really believe it internally. Some people believe it internally, but whatever, it's crazy people everywhere. It was that at this company because they're huge, they have tons of enterprise contracts and some of those those CEOs have deals with them, and that CEO says something like, we have a one hundred times speed up in productivity,

22:09Speaker 3:and the original guy, the guy was talking to, goes on the news and says, well, one hundred times is not possible. He has embarrassed his customer. And these people can be very, very petty, so they'll cancel the contract. And the first guy gets fired for insulting a customer. And where this weird game theoretic prisoner's dilemma, you know, standoff where you can't even contradict another executive because you can't tell if they're cynically playing the game, because then you can say it and they don't care or if they really believe it, because then you have embarrassed them in public. So all of them have just sort of gone.

22:43Speaker 3:Even pretty responsible people have gone. I'm not going to disagree with them. I'm just not going to talk about this because it's too dangerous to talk about.

22:51Speaker 1:That is fucking insane. I realize we to say this a lot today, but it's fucking insane. That is one of the craziest things that if you go on television and say you don't like something, you will lose a contract over it, which kind of explains what happened with the PR firm, I guess. I mean, I don't know how they would have worked out. I don't like AI, but maybe there was some clues somewhere. But this is an industry gone mad. This is this is the craziest thing I've heard in any and this is like an actual information hazard situation. I'm I am actually a little bit chilled by it.

23:25Speaker 3:Yeah, I think it's fascinating because it's always been bad to insult your customers, right, obviously, but usually all your customers aren't saying the same thing at the same time in such a destructive way. When they're all saying one thing, that topic is just off limits because if you agree with them, I mean, you haven't really accomplished anything. They don't care if you agree with you. You know, they're just like you know, for example, if you were the CEO and zero or something, right, you're gonna have huge contracts with a bunch of big companies. You can't actually say what you think on this topic because you might upset someone. And I think what's been really interesting thing as well as seeing the pressure from boards. I've had so many executives right in since I wrote that and go, yeah, my board has ordered me to use AI. I don't believe in it, but they said if I don't use AI, yeap boards are saying we But I've also had board members tell me that they feel like if they don't tell everyone to use AI, they'll be removed from their positions.

24:20Speaker 3:So like, it's just this, I like, I don't know what to tell you. It's insane, it doesn't make any sense. It's like this huge mass delusion event that I feel like we might be writing about for the next twenty years. Like it's just fascinating.

24:34Speaker 1:Well, I think what the contra to all of this is It also feels like it could reverse very aggressively and suddenly, because if everyone's just kind of sitting around being like, yeah, I actually fucking hate or like there's a contingent of people that's impossible to measure, of people who has kind of fucking sucks. I kind of fucking hate this, but I'm going along with it, and enough people start just saying they don't like it, it just feels like one day that the mood could shift. Because this is why I've been predicting the whole time, that just at some point everyone will stop doing AI or someone they like and respect will say fuck this, I'm not doing it anymore, and then everyone will walk away.

25:12Speaker 1:They're just looking for permission.

25:14Speaker 3:Correct. We're basically talking about coordination problems, right, which is a thing where a bunch of people want to say a certain thing. If one person puts their hand up, that person will be killed. So no one wants to put their hand up first. What you need is just some event for everyone to say the same thing at the same time. And what's interesting is we're here because a different problem was solved, which is a lot of executives and board members and people are very easily excitable.

25:40Speaker 3:They always had this sort of latent level of mania. AI coming out in the form of lns solve that problem for them, because if you put your hand up and you said, I'm going all in on the metaverse, like, that's kind of a joke. It was silly back then. So we first solve the problem by giving them something to coordinate and be frantic around. And now we're just waiting for the first like hyperscaler to collapse for the reverse event.

26:06Speaker 1:Or just sach In the Della or probably not so Dabashi with the way that Google's earnings just went. But someone is going to eventually just go fuck this. I truly I think it could be meta. But also who knows what that wacky son of a bitch Mark Zuckerberg would all get up to next. I hear he's working on a tamagotchee. You know, whack, you wacky son of a gun. It is really cool, though it kind of reminds me of your other piece I Love, which is a brainwashing executive today. But the most extreme version of LinkedIn psychosis, where it's just everything you do is an artifact of how much LinkedIn you consume, except everyone is LinkedIn.

26:47Speaker 1:Now everyone, can I talk about that brainwash piece for a second? Yes, please, please please do it?

26:52Speaker 3:Okay, so you know the whole thing, and the piece is one of my executives. When I still like it didn't work for myself, came in and just asked everyone at the company about a product called Monte Carlo, which is like super niche like fucking nerd thing, and we were like, why are you talking to all of us about this in our one on ones?

27:12Speaker 1:You need talk to us.

27:13Speaker 3:Once a year. This would be like my granddad coming in and going like, tell me about like quantum cryptography, Like there's no way that this is a topic you should be interested in, right, Like it's just so far out of your wheelhouse.

27:25Speaker 1:Yeah, where'd you hear about this? Like when my kid comes in and says, in you would I'm like, yeah, where'd you hear that?

27:31Speaker 2:Now?

27:32Speaker 3:I did hear about this, and it was it was LinkedIn. So they did tell me where they heard about it from, which were good dire. But I don't know if you saw this the CEO of Monte Carlo, not knowing I'd written about the company with you know, I wouldn't say criticize them, but I had some question marks and I was like, who are you advertising to and what the hell's wrong with your website? So she took that piece and boosted it and said, it's you know, being shared with executives and boardrooms around the world. It might be the most circulated thing in boardroom right now. I'm measuring that.

28:02Speaker 3:And then if you look up the Monte Carlo website, you know from someone who has just gone you know, I like this piece. I think it raises some good points Monte Carlo. I don't know if it's always at this, but it's now Monte Carlo dot AI and they've completely shifted their messaging because they are no longer a data observability platform. They're now an agent observability platform, whatever the fuck that means.

28:26Speaker 1:Like Jesus fucking What's great as well, though, is the reaction to your peace and not be looking it up. It is. I occasionally see it with mine, but I'm more direct with how I feel about AI people, so they're less willing to share it. But the amount of people I've seen sharing it who are very clearly the subject of the piece is insane to me. It's just like very funny. It's just like people like, yeah, wow, can't believe people are missing. People are all using AI for no reason. I have course, have nothing to do with this. All my AI solutions are perfect. I love AI so much, but these other room baciles who are using it wrong each Yes, it's genuinely driving me insane.

29:11Speaker 1:If I'm going to be honest, I have a phrase for this, which I think is novel. I might I might have invented this.

29:17Speaker 3:It's the N minus two idiot solution, which is if you look around a room and you go, that guy is the smartest person in here, everyone will agree and in a room full of sufficiently you know, ridiculous or compromise people, they're all okay with it because they assume they're number two, right, So you just need to leave them in enough wiggle room. But what you can't do is go only one person in this room is smart. You can go two of you are smart, and everyone's okay with that. They'll all agree with the first person.

29:45Speaker 3:They all assume they're the second person. And Yeah, a lot of the coverage has been you know this, this insane phenomenon, and you know, credit to possibly the Monte Carlo Ceo. Maybe she didn't say this in so many words, but maybe what she was actually saying was I thin trapped by the situation. My whole company has pivoted to AI agents, even though it doesn't make sense. Maybe that was her gentle way of screaming into the void. I don't actually know, but it is very funny. I kind of want to do a blog at some point. That's just you can stop using AI. It's just like, literally, you can stop this at any time.

30:20Speaker 1:You don't have to do this. If you feel trapped, you can just leave. You kind of did that at the end of the piece. You did a very empathetic thing at the end, saying here's how to deal with the organizational AI stuff. If you're told this, well just quit, And in other situations you could also do something else. It's just it's genuinely sickening. It is actually sickening in a way that it pisses me off. I've said this before in not so many words, but this is meant to be an industry of meritocracy. This is meant to be the industry where people, oh, we do things for efficiency reasons, and we do things because we do things because they get good business results. And now I'm realizing you've kind of told this in the past as well.

31:04Speaker 1:Everyone's lying about that they were just doing stuff that made them feel good.

31:10Speaker 3:Yeah, I kind of wish we had that. You know, every so often you'll hear someone say something like, you know, the issue with capitalism is it optimizes everything for money, right, we to throw away every other value for money in efficiency and productivity, And god, I wish that was true, because we somehow have like the worst of both worlds where we're throwing away every other value in somehow still not achieving peak efficiency, right, Like we're doing all these horrible, hideous things, and then you look, well, how much money are you making? It's like not even that much money we're wasting in all the data centers. I mean, the data centers alone are just fucking insane to me. I've bean what's really good as well is on the subject of meritocracy and numbers, trying to just pull together very basic things like is any of this profitable and asking people and them get angry at you, just very angry.

32:03Speaker 1:Just agitated, but like why did you keep asking these questions? Why'd you keep doing this? And I grow up in a rough naighthood. Have met my fair share scam artists, and one hundred percent of the time, if someone gets angry at you for asking for things like that, that's because they are full of shit.

32:18Speaker 3:Yeah, yeah, you know, I was in psychology before I was in it, and you see that a lot in things like multi level marketing scams and stuff, you know, like they're if you go, hey, isn't this a pyramid scheme? They've got the whole script which is like, what do you mean it's a pyramid scheme? What's what's your understanding of a pyramid scheme? Answer to you fucker, And it's yeah, it's sort of the same thing here. Didn't we Maybe we talked about this previously, but have you seen like that clip of you know, a very very gentle interview with Sam Altman, and at the end, this interviewer, who clearly loves him, very gently goes like, so, what's the plan for profitability? It's just like, yeah, have enough of these questions.

32:59Speaker 1:No, He's just like, we're waking way more than thirteen billion dollars this year's yeah, you know, and if you need someone to buy your stock. And what's funny is sach An Adela was on that call, like just your you're yeah, just such an a della, just sitting there with the I don't know, the magic magic roundabout, just nothing going on up there, waiting for the next input. The world's just like he did the little monkey with the symbol that does backflips, just sitting and setching a della's head is going wait for me to.

33:32Speaker 3:Say, I'm imagining like you know, the little mac os speech ball, just a little spinning widget up there.

33:38Speaker 1:The thing is, Samon will never talk to me. They've already told me he won't. But that's because I know within five minutes he will get pissed off of me, because he would just be really easy to just be like, hey, why don't you make any money? Why'd you lose so much money? Sam, Well, we make money, and then just you'd get more and more immature until he event actually tries to get at you through the screen like a dog that sees it reflection. Easy peasy. He tried and kill me with the other day.

34:07Speaker 3:I mean, there's no reason to write like you know.

34:11Speaker 1:As a pr person, I would I would rather kill myself than put Sam Moltman on a call with me, like I've just that is the that would actually be the appropriate choice, Like whatever you do, do not put exitch you're on the call with Sam Moltman.

34:25Speaker 3:Yeah, it's also it's like only a losing proposition for it. Yeah, because let's say he comes on and he has the greatest showing of his whole life, you know, like he just he feels every question perfectly. That is the same outcome as just talking to another journalist that's gonna jerk him out for an hour. Right, There's no reason for him to take that risk when he can get so much positive coverage from people that aren't gonna ask him any questions.

34:49Speaker 1:And I will only ask him the most annoying I'll just ask him are you happy? And he'll his head will start like a erupting in smoke. Anyway. Anyway, so we've kind of got off top pick. I realized I wanted to actually talk to you and break down something that I've heard with various people, which is this whole issue of demos with AI. The AI is really good at making a good enough demo to sell to someone, and you kind of covered this before A I was as far down the whole as possible with very much an executive. But I get the sense that it's just very good at doing enough to sell an imbecile on something.

35:27Speaker 3:Yeah, and you know, I'm willing to be a little bit more charitable here because it doesn't have to be an imbecile necessarily. It just has to be someone who is not proficient with tech, because when you work with something sufficiently complicated, you understand there's a lot of difficulty around the fine details and deploying a piece of software and the design and achieving business value whatever you want to do. But you know, as a end user, like I don't really know how my oven works. I just have some nobs that I twiddle and then the oven works. Right.

36:00Speaker 3:So if you turned up and you said, you know, I've gotten atobots that assemble ovens perfectly, and it does it and it bakes a pie. If I don't eat the pie or you know whatever, I might go, Hey, that was really really cool and it is genuinely impressive enough that possibly no one would blame me for investing in it initially. And you've got shown something that looked plausible, right, and not just plausible, like literally impossible. Prior to twenty twenty, like you couldn't do this sort of trick, and then the next day you start to find out like, hey, well you know what, like this this pie is really bad and it turns out the oven only has one setting and sometimes it catches fire. And you know, I still might not be an imbecile because what I saw upfront was so impressive, Right, I would be an imbecile if I kept buying it four years later, right, Like there was a period where the overinvestment was understandable.

36:59Speaker 1:Well, this actually gets back to brain Motion Executive, because there's the point in there. The people that buy software often not the people that use it. So you just have to impress the CIO or the CTO or the CEO enough, and that person does not have the practical on the ground experience to make that decision, but or usually does yes.

37:21Speaker 3:And that is basically what makes not just this COLN, but like a lot of enterprise bad software possible. You just go, hey, you know your customer support monkeys, they need a new platform. Do you want to put them on Zendesk? And then someone buys Endesk and is the CEO like, you're never going to log into Zendesk, you should. I actually think, you know, most executives, I mean, aren't that busy. They're pretty busy.

37:43Speaker 1:Yeah, they've got nothing on I'm serious.

37:45Speaker 3:Yeah, Like, but they absolutely could just go down and sit next to someone for like an hour and check out what their work looks like. That might do a lot more for sort of a long term company strategy than whatever they're doing, like going to big meetings. Look, there's probably some level of wealth where that's not true, if you're worth personally worth enough billions, but yeah, tons of people worth hundreds of millions of dollars still take the time to do that. And again, they run very profitable companies and they don't waste their time on this stuff beyond what's necessary.

38:29Speaker 1:It really is just the grimest thing ever. And I'm guessing that the reason, like the reason that llms are so good at creating these demos is you can vibe code together something that has like buttons that work, and yes, the buttons work is good enough.

38:44Speaker 3:Yeah, the buttons do work. And that's the thing. The buttons do work, which is why it's sort of understandable. But then you had other problems, right, like you know, the buttons work until you load it on mobile and then this button disappears and that button doesn't work, or two people click button at the same time and it puts down like you know, it doesn't block inventory. So now you've got minus one in the warehouse. Like all these things, and you know, it's not that an LLM can't solve them, especially with human guidance, it's that it hasn't delivered on the promise of human takes its handoff and software is solved it.

39:20Speaker 1:It's also very likely not running on your infrastructure, right Yeah.

39:24Speaker 3:I mean a lot of the time it's liters running on a laptop and you haven't solved complex deployment issues and that kind of thing. But again I understand, right, like if I'm talking to a random CFO, they don't know what infrastructure means in the context of it, so it's all basically understandable. But it's the same reason sane people don't get super excited about developments that they're not specialists in.

39:48Speaker 1:Right.

39:49Speaker 3:I don't walk around every time there's like any breakthrough in medicine going oh my god, this is going to change everything. I go, I don't know how to evaluate this. I have no idea of this is game changing. I'm just going to shut up and let doctors talk about it.

40:02Speaker 1:But this feels like, actually, I think I've kind of worked out something, which is the problem isn't so well. The problem is the fact that you've got executives buying something that they don't really understand. But that's been the problem before. The problem is they're more incentivized to do it instead of that. One example of an executive getting kind of brainwashed into buying Monte Carlo, it's I will buy anything with AI on it, and in fact, I believe that if I don't do this, I will be under existential threat.

40:34Speaker 3:Yeah, one hundred percent, And you know, for when it's worth that executive didn't buy Monte Carlo and the end, they just wasted a bunch of time. And I suspect if Monte Carlo had AI on their front page, they would have bought it right like it realized instantly. I think they would have been like yeah, especially because you know, with this whole thing where no one is actually using the tools, it lets you tell your CEO that you've made a you know, investment in AI and it's transforming something.

40:58Speaker 3:Don't ask any questions about what it's transforming. It's actually hard to make businesses more profitable, especially established ones.

41:05Speaker 1:Right.

41:06Speaker 3:Startups can do whatever they want because they have nothing to lose. But if you've been in business for five years or ten years or twenty years, there's a lot of things you're doing and it's not as easy as just going claude double my profits. If that worked, we'd already be seeing it. That doesn't work.

41:21Speaker 1:It kind of sounds like most software doesn't solve problems. Well, I don't mean that. I don't mean that no software solves any problems. I just mean people regularly buy software that does not actually improve their companies.

41:35Speaker 3:Yes, that is pretty common. And this is not an lll M point, right, This is.

41:40Speaker 1:Yeah, yeah, no, And I've really I must be clear, I mean this on a much larger scale than l elms.

41:46Speaker 3:Yes, I would say a lot of software does not deliver on its promises. They tend to have much stronger sales teams than they have Firstly, you know literally programmers like the programmers tend not to be very good in a lot of companies. I had to explain to a physio yesterday that we're not a regulated field like this. This blessed, very happy man was just like, oh yeah, they wouldn't let you program a computer without like a license.

42:13Speaker 4:Physio for so long. I know, I felt like I was breaking his heart. That's really fucking good. Oh no, I kind of think you should have to now. I kind of think we need to compute a license.

42:28Speaker 3:Well maybe right, like anything else serious, like you know, if you want to be a firefighter, you train, you take it seriously.

42:35Speaker 1:Yeah.

42:35Speaker 3:I had to explain to him like, hey, man, like, they'll let you program if you've read zero books on the topic, right, you can be a principal engineer. And he went, that can't be true. I've met people who were like three hundred thousand dollars that have never read one book. Like it's yeah, so good. It's to be clear that regulation of to do software engineering and a company you need a license. Maybe not, but I think we can extend it to you can only use social media on a sit down computer.

43:07Speaker 1:That's a good start. And you can only talk about programming if you've read these three books, like you have to qualify to be able to post about it. You can't post about it.

43:16Speaker 3:Yes, I think it would be helpful if we had some sort of accreditation that lent some trustworthiness to the right people. Because I don't even read comments I get on things anymore. I do emails. If someone emails me, I'll read it. But like, the quality of comment is so bad. And it's not just people that you know, disagree with me. A lot of them have literal reading comprehens and issues. I rarely get like an actual critique on what I've written. I get critiques on people think I've written. Yeah, even sometimes people will agree with me, and I'll see they've also missed the point, and I'm like, well, I'm just going to stop reading these. But to answer your question though about software, you know, not being very good, there's the quality issue of a lot of pro grammar aren't very good. You can have badly programmed stuff that is still good enough to deliver value to someone, but a lot of software is fundamentally ill conceived. So an executive will hear some pitch that's like, well, if you get this software, you know, you'll have ten percent more throughput on your customer support whatever. A lot of the

44:19Speaker 3:time that is just not true. So they do the whole migration into a new system and there's just no outcome at the end of it. They just don't see an improvement.

44:28Speaker 1:That's I mean, really just feels like the entire software entered industry is based on nothing at this but like there's just I keep thinking of calamity at this point because we've got tons of people coding with other lambs, not reading the code anymore. You've got software being bought by people that don't use it, don't really understand there, and now just a mass acceleration of that problem that already existed for ten or fifteen years. I think probably the problem with discussing is probably at least fifteen years because it's like a very SaaS based problem because software has a service allowed you to subscribe to software rather than buying an expensive license up front, which sounds good right up until you're in year three and there's been a three price increases and you realize you're paying Oracle.

45:12Speaker 1:But it feels that we're just accelerating dramatically towards something breaking somewhere at scale, Like I really am. I mean, we were talking about GitHub last night.

45:26Speaker 3:Right, Yeah, Oh my god. The GitHub outages have been so funny. I've been really enjoying them. They I think I mentioned that. What was it called? Oh, I god, I'm going to disappoint my friends so much. Forgot the name of the piece. But it's about GitHub and like their massive outages. It's by from.

45:46Speaker 1:Link hub, used by basically every person in software engineering, and it has the uptime of an unstable startup with two thousand dollars.

45:57Speaker 3:Yeah, I would say, start up with two thousand dollars, get it? She higher uptime than that? Like, I think that's actually probably true there. Look, there's some accentuating circumstances like get hub deals with a volume of traffic that a startup doesn't. But they're not broke right, Like, what's what's going on here? And more importantly, they weren't having these issues a couple of years ago. Like they're doing something wrong. I've heard it's related to you know, Azure migrations from inside Microsoft. It probably doesn't help that they're slinging LLLM slop everywhere too. In that post by Ephron, he did talk about how almost every change log they've had has just used the word agent and AI over and over. He never sees the words oh yeah, yeah, there's.

46:45Speaker 1:No like stability or the word improvement. It's just agentic. I think that there are now eleven different companies that claim you can do agents with credit cards now as well. It's just like that. It's like the tech industry is all just k fa.

46:59Speaker 3:Sorry, just what do you mean AI with agents with credit cards? What does that mean?

47:04Speaker 1:Well, that's a good fucking question, nicky. They are suggesting that you can connect an API of some sort to an AI agent and then the AI agent will go and buy something with your credit card.

47:19Speaker 3:I mean, look, as a matter of technical truth, that is true. Why would you do that. That's such a bad idea.

47:29Speaker 1:Well, because you were you you love well. I mean, no one uses it. No one's actually using these things. Door Dash is doing a command line interface now, why good? This whole industry is well, what I think it is is just especially Bay Area people have just been driven completely insane by this stuff, and they are The way I describe it is it's a community of thirty to fifty year olds desperately trying to impress twenty two year olds, and they're all desperate to make the like they're all like on on God, this is permanent underclass, and they kind of look around, they're like, yeah agentic right, fuck command line? Yeah on God, no cap And it's just insane because even like Mark Andrews and being like Based, it's like, shut the fuck up. You are one hundred years.

48:26Speaker 3:Old, say Based.

48:28Speaker 1:I believe I believe he has said Based more than once.

48:32Speaker 3:That's actually worse than all of the AI investments.

48:35Speaker 1:Oh fuck your mosaic browser you fuck with It's just like, it's just like these people don't know anything, and I think we're just because the argument about the Capex, for example, keeps coming back to, well, these are the richest companies in the world, run by some of the greatest tech talent in the world. What if that isn't true. What if that's not true at all? And in fact, what if it's not run by the smartest people. And indeed, as we've been discussing today, there is an active culture to reduce intelligence across the board. And that's the thing.

49:04Speaker 1:If I am right, the entire tech industry is going to experience a depression because everyone went insane, Like everyone got taken into Heaven's Gate, except instead of killing themselves and other people, they're just doing command line interfaces for fucking door Dash.

49:21Speaker 3:I don't understand why you want a CLI for DoorDash. That's baffling. But I'm not even going to talk about that because it's upsetting me because like I don't like, what did they think?

49:33Speaker 1:No, hyphen cli lets you order door Dash directly from your agents, search stores, find the best deals, check out and more.

49:44Speaker 3:See Like, okay, no you can't. You can't trick me to talking about this. I have like eighty different thoughts and I'm not.

49:52Speaker 1:What I love the The first response is Boris Cherney, the head of Claude Code, saying, brb downloading. I don't know much about computers, but do you do you download as CLI? Like I mean, yeah, you type in like you know, Like oh no, I just I feel like these are the people with the money, and they gave me even half this money. I could do basically anything else and make them way more than zero or negative as is happening to them. But it's like, these people are never going to This is a horrifying thing to say.

50:30Speaker 1:I realized these people are never going to learn without suffering. Like it's just it really is that the tech industry. I'm not saying I want anyone to suffer. I really wish there was a soft landing out of all of this, there was a way out of this without genuine depression hitting the tech industry. But I think that's the only way anything changes, because they've never like the dot com bubble was much I mean the telecommunication sabers pict but the dot com bubble. The tech industry is much smaller back then, and I think that something like that is the only thing that could change things because they need an egregious screaming in the face. They need to be told, not just that they're wrong, but they need to be shown it in a graphic way. Yeah, let's they're gonna start learning stuff unless they're gonna start learning things.

51:18Speaker 3:So you know, I had you on my podcast a few weeks ago, and the guests right after you were the authors of a book called crisis Engineering, So they did some of the healthcare dot gov bailout. One of them, Marina, was the CTO of the Veterans Affairs in the US, And the thesis behind crisis engineering is just that organizations don't undergo change unless they're an acute state of crisis, which is not just a problem, I mean a crisis. So bad that there is a complete dissolution of sense making like panic in the streets. This is sort of true everywhere, right, Like people tend to make a big lifestyle changes after a huge divorce or you know, some massive event like that, And yeah, I think you're right.

52:00Speaker 3:I think it's why my advice to most people now is not, like, you know, raise a huge think about this. It's like if someone is actively hurting something in front of you that you care about because of AI, yes, stand up and say something. If you work at a fire department and you just need some money to fix a fire hydrant or whatever it is that is important to you. Right now, we're so far past the point of arguing that these people, I'm like, just ask for your money and ask for an extra ten thousand dollars for chat poth that doesn't work, fix the problem, Just pretend, pretend this doesn't exist. Until it all blows up, it'll be fine, right, Like I'm just waiting to pick up a bunch of really cheap and video shares. That's all I'm doing.

52:45Speaker 1:Well, that's the thing. And then as we wrap up here, that is kind of why I'm telling people, which is, after this blows up and it will. Goodness, it will. There is going to be a lot of money and competency. I think you've already kind of seen this as well. It's just like people are going to realize because of the crisis to follow that the imbeciles have taken over the business. Idiots are everywhere, and I'm already I mean little pr work I've done in the last few months.

53:15Speaker 1:Even then it's people being like, yeah, I keep talking to people and they justy, I generate fucking everything. They don't seem to know anything. And I think that that is going to be everywhere the things that people read. I don't think it's going to be a utopian society or anything, but I think that there is going to be engineering in any kind of business management role a certain degree of Right did you get taken by this? Because I must be clear, once this explodes, I am not going to stop pointing out people who fell behind this.

53:46Speaker 1:Oh I'm going to be I have detailed notes to quote Casey Newton, not speaking about anyone specific about who I might be pointing out here that we will it will actually be I think aucivic duty to remind everyone who fell for this, and I think that result of that is competent people are going to make more money.

54:09Speaker 3:Yeah, and if I could, I have a one bit here where I think I'm a bit less optimistic, which is once this blows up, suddenly everyone will pretend that they were never never for it, right, Oh yeah, And a lot of these people are then going to immediately pivot because they have no principles, right, just chasing money. A lot of these AI experts are going to be to go on LinkedIn and change their byline to AI project clean up crew, right, Like I think the time to be competent is right now. Like I pay my team entirely through not doing AA. I'm the only fucking guy out here in Melbourne. I know that isn't just doing AI projects twenty four seven, right, But the market is there and it's maybe the easiest it's ever been to find sensible clients because now they're all signaling when they're insane. Once the bubble blows up, I'll have to find out by like signing the contract and finding out if they suck on the day. Now I don't have to.

55:06Speaker 3:I just need to see can they make it through a forty minute conversation with me without asking me to pivot their whole business to AI, and if they don't do that, they're almost certainly awesome to work with.

55:16Speaker 2:Well.

55:16Speaker 1:Thankfully, my rautism allows me to document every single person I've ever seen mention AI, and it's really easy to tell. And by the way, every listener of this, show your civic duty when this breaks is take screenshots and respond to anything anti AI someone says when they try and pretend that they never fell for this. This is going to be like the super Bowl of this. You images, Oh, I've been keeping I have hundreds of screenshots of people saying stuff about AI and it like the dumbest shit in the world. I cannot wait. I'm going to be quote posting so much my fucking trek pad breaks.

56:00Speaker 3:I will keep a folder and then I will send it to you right after you invent a death note.

56:05Speaker 1:Right. Oh no, No, I don't want these people to die. I want them to be alive to read how much I'm going to post about But I need them to like, they need to be alive so that they can suffer by by facing the truth again and again and again. And I think it, I think it is going to be funny. It is really it remains funny to me though, the these people still that there are people who are still questioning whether there's a bubble like it's just it's remarkable to me.

56:37Speaker 3:In fact, I don't know anyone sensible who's questioning it. Even people I know who are like I really like AI, They're like, this is clearly a bubble, like it's it's fucking over.

56:46Speaker 1:It's it's jovah, Well you know what, much like much like this podcast, then Nick web quere can people find you.

56:54Speaker 3:At two places? So the first one is my personal blog ludic, the blog that's l U d I, or my company's website, which is hermit dash tech dot com and we have a blog on there.

57:05Speaker 1:Too, lovely, and you can find me of course on Google dot com slash Google and then type in Zitron criminal and ignore all of the other stuff. Just look at the podcast. I of course, we back later in the week with them on log and who knows it's earnings week, so who the what the fuck's gonna happen? Who goddamn knows? Anyway, Love you well, thank you for listening. This has been better offline. Thank you for listening to Better Offline. The editor and composer of the Better Offline theme song is Matasowski. You can check out more of his music and audio projects at Matasowski dot com, M A T T O S O W s ki dot com.

57:52Speaker 1:You can email me at easy at better offline dot com or visit better Offline dot com to find more podcast links and of course, my newsletter. I also really recommend you go to chat dot Where's youread dot at to visit the discord, and go to our slash Better Offline to check out I'll Reddit. Thank you so much for listening. Better Offline is a production of cool Zone Media. For more from cool Zone Media, visit our website cool Zonemedia dot com, or check us out on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.