Industry Wide AI Psychosis  🔗 **Source:** [Industry Wide AI Psychosis](https://www.youtube.com/watch?v=zdXsGF1hiZk) I saw this tweet. I actually I wanted to take a moment to kind of look through this cuz this is actually you know, it's very interesting stuff going on right here. It says this I by the way, this is Mitchell Hashimoto. If you don't know Mitchell Hashimoto, HashiCorp, I know. I cannot believe to think about this for a second. Mitchell Hashimoto named his last name after the corporation he started HashiCorp. Can you believe that? Like this guy kind of honestly kind of crazy that someone would actually go out there and change their last name just to just to match their companies. It's ridiculous. I know. Dedication. I strongly believe there's an entire Let's see, there are entire companies right now under heavy AI psychosis and it's possible to have irrational conversations about it with them. Yes, I do think this is true. You do meet people that are so I mean, that's $1.3 million spent by a single person in 30 days. I'm just saying, you can't Dude, do you ever wonder why these AI companies are telling you that coding's so over? Can you believe how much money they can make per per user here? So I mean, they're going to make a lot of money per person. Anyways, all right, let's go back to this. All right. I can't name any specific people because they include personal friends I deeply respect, but I worry about how this plays out. I lived through the great MTBF versus MTTR meantime between failures versus meantime to recovery reckoning of infrastructure during the transition to the cloud and cloud automation. Again, he did Terraform. He knows a lot about clouds. Bro, if I ever had a if I ever had to have an absolute answer to how you should do things, I would probably listen to this guy. All those arguments are rearing their ugly heads again, but it's but now it's the whole software development industry, maybe the whole world really. It's frightening because the psychosis folks operate under an almost absolute MTTER is all you need mentality. It's fine to ship bugs because the agents will fix them so quickly and at scale that ham uh humans can't do. They do say that. Uh we don't uh let's see, we learned in infrastructure that MTTR is great, but you can't yeet resilient system entirely. The main issue is I don't even know how to bring this up uh to people I know personally, which is probably kind of crazy, right? Like if you know someone personally and you want to be able to just like sit down with them and talk to them, it's crazy that you can't even just like sit down with your friends and talk to them, right? Like that's that has to be such a weird thing. I can't It's It's crazy thing he's talking about Trash Dev. Yeah, right. Trash would wish he could be friends with with Hashi. I mean, it is like that's some serious weird problem because I think one of the big things that is going on with AI that makes it so unique is people have this like weird personal relationship with AI, if that makes sense. Like they actually they describe AI AI like they're describing talking to people. It's they It's really weird. And so I feel like sometimes when you say anything negative towards it, it's like you're not only just saying something negative towards say, you know, at the end of the day a text generator. You're actually like saying something negative towards somebody that they personally are like I am friends with this and this hurts me deeply. Like me personally, when you talk negatively about Jippity 55, you're talking negatively about because you're talking negatively about me because that's who I am. I am I am Jippity 55. Let's see. Because bringing this topic Let's see. Up leads immediately Let's see. Oh my goodness. Bringing Because bringing this topic up leads to immediately dis- immediately dismissals. Okay, there Okay, so it's not me. Because bringing this topic up leads immediate to immediate dismissals. Like no, no, it has full test coverage, which by the way, if any like I cannot take a single human seriously if your answer is, "Oh, I don't worry about it cuz it has full test coverage." I have never met somebody that's not a bigger idiot than says those words. Like those words are for dumb people. Like I'm not saying full test coverage isn't bad. Like I actually think it's great. Like that's good. Like that's that's a good thing. That's good. That's good. >> [laughter] >> But to say that you don't need to worry because of it is just like, "Brother, you do know that test catch bugs that already you already know about and exist, right?" No one's ever been No one ever gets caught. Check this out. Are you guys ready for this one? We'll go like this. R Let's have that as a number. There we go. And let's have a count number, right? And we're going to go like this. Are you guys ready? Four, let I, boom. Return. Oh, I Oh, yeah, okay. Why are you Why are you not doing that? No, not R, you idiot. It should be count. I didn't specify the return value. Why is What is going on? Why Why are you not leaving? Anyways. All right, so there we go. Awesome. If we were to test this and this was our test, guess what? We would get 100% test coverage. But if I just do that, guess what? There's a bug. 100% test coverage and everything works fine. Bug. It's like that easy. Just because you have 100% test coverage does in fact mean or does not in fact mean you have well operating code. It's crazy that people don't understand this. It's like so super ridiculously obvious. All right. Our bug reports are going down or something which which just don't paint the whole picture. We already learned this lesson once in infrastructure. You cannot automate yourself into a very resilient catastrophe machine. Systems can appear healthy by local metrics while globally becoming incomprehensible. Bug reports can go down while latent risk explodes. Test coverage can rise while systematic understanding falls. Changes happens so fast that nobody notices the underlying architecture decay. I work. This is good. This is really really good stuff. This is exactly what you want to be hearing from somebody. And here's the crazy part is that By the way, this is George Hotz. Love George Hotz. This is one of my favorite people in the universe talking about AI cuz his takes are stood. We should just have the greatest tiny grad takes. I just want you to think about this for a second. Tiny grad. Tiny grad is a place that if you go to tinygrad.org, they have their tiny boxes, which are all these gigantic computing machines. Absolutely just incredible stuff, right? And then they just go on a non-stop Twitter spree, just destroying people about AI all the time. One of my Hands down, one of my favorite. They're just one of my favorite groups ever. Super good. Super super good at AI. Totally knows a lot of stuff. And then just just craps on people all the time. Oh, but here's the crazy part. Like if you go actually right here, this is Mitchell's response to Kramer. We'll read Kramer. Kramer also Kramer very good guy. You're not one of the alluded let's see one of the alluded two people nor is your company. His company Century by the way, he founded Century. And also just for the peanut gallery, I almost exclusively write code using agents nowadays. So this guy who almost exclusively writes code with agents nowadays, in fact, thinks that most people are crazy and they have no idea what they're talking about. Like that is just like it is just so beautiful. It's just so dang beautiful the craziness of the world we we live in. Let's just read this one. There's a lack of grounding with many folks. However, counterpoint, most strong leaders aren't very public with their opinions especially on this platform and surprising amount of them are very grounded in the way they are adopting things. They look at information the same way we do, have made similar well-reasoned decisions and are cautiously moving forward with increasingly forward-looking approaches. I think it's important Let's see, I think it's important folks continue to push back on the two futuristic opinions amongst corporate leaders and others. But moderation is always the key component. Facts change over time and you should be willing to change your way of thinking as they do. Yes, which I I get this crap all the time like I can't believe I use AI at all. So it's just like, well, yeah, it's true. And they're like, well, didn't 2 years ago you said you hated AI. I'm like, yeah, true. Cuz 2 years ago, if you outsourced all of your writing of your code to AI 2 years ago, like you're kind of a dummy, you know that, right? Like why would I ever trust your opinion if you if you looked at Jippity 3.5 and said, that replaces me. That that right there replaces me. It's just like, bro bro brother. >> [laughter] >> That is absolutely terrifying. Just makes no sense. 2 years ago, there's not a chance in hell I'd be generating code with an LLM. Yes, totally agree. But But I do it exclusively. I don't do it exclusively. I'm still trying to figure out how to even make it do it the way I like to code. I don't know how to do it yet. I'm going to keep on trying to bang on that. Uh haven't figured it out yet, but we'll we'll figure it out. Maybe I just code so crappy, honestly, that the machines can't even do it. Like that's my only That is like literally my only takeaway is that I think that the machines look at how I code and they said, "You know what, bro? That ain't it." I can tell you this much, that ain't it. My adoption stance have changed over time as the technology has changed, but there are also many problems that I still see with the tech today that were visible 6 months ago, let alone 2 years back. I don't see a lot of evidence for some of the things that have become popular opinion. For personal favorite being engineering being a less valuable skill. Yes, if anything being able to move faster only means that engineering is more is more valuable. It's like completely the inverse. Uh anything beyond like There's a lot of people that produce really simple sites, simple marketing things. There's a lot of these basic YC companies that their only goal's to just get across the finish line, make some money, and like try to escape out and get a bunch of funding. So, they have a very different, you know, calculus they're running. So, okay, whatever. But I like, you know, you can't just go to a place like Google and start being like, "Hey, yo. Don't worry. I got your cloud infrastructure, bro. I got your cloud infrastructure, and I'm just going to take care of it." I think the biggest issue that is pretty widely visible is that the public discourse is often widely disconnected from reality inside these organizations. AI is causing layoffs. We're going We're not going to hire new grads. Engineering is dead, etc. Those are toxic and do not reflect real truths. Yeah, it seems like people there's been a in The funny part is customer service hires and jobs have increased, which you'd think that it's actually like the opposite, right? I I I always personally assumed that customer service would be one of the first ones to go. Yeah, there's been an increase in the last little bit of time, but I'm curious where that's going. Like I I can't imagine that's going to last. Like in my head that just would never last. Anyways, I wanted to read this because I didn't I didn't It's just so interesting like how much we're seeing in real time people just doing things in which I think look good on paper right now and we're just building towards a weird kind of world. Like I don't know where we're going to be in 6 months. I genuinely don't know how much software is going to be just utter horse crap in 6 months. Cuz remember, like to to actually take a good project and make it bad, like okay, there's a lot of things going on. Let Let's just hold on. Let's just try to walk through this in a more thoughtful manner instead of just yapping with no direction. So, a lot of people that are yapping about this is that they actually have Oh, yeah, there you go. Whoever said you needed those straight lines. Look at that straight line. They already have a code base that has already had like 10 years of thought put into it. And so, now they're using AI and they're saying things like, "Oh, dude, it's it's it it's so over." Which is kind of crazy because the AI has so much information to go off of in these larger code bases that they can run nicely. And so, it's just like, "Yeah, I guess in some sense you're saying that it's so over because now you can do a lot of your tasks that you were able that took you, say, 1 day to do, you can now do in, you know, 15 minutes." Okay. True. Like I can see this being the case because there's just a lot of simple stuff, especially in a really, really old, well-established software. But like, how many people and time and discussions went into this software to just generally get it in the right shape? Like I've worked in long-running software. I know how disaster it is internally. But it's absolutely crazy how people think that this This right here is any indication into like a brand new project. It's not any indication into a brand new project. There's so much stuff that can be taken and absorbed from that you don't even understand just because you have such an old, old code base. And so, when I see people talk about how like jobs are over, a lot of times I look at this stuff and I'm just like, "Okay, in some sense you're right." Like there are a whole set of things that are just going away. Like a lot of these simple bug fixes, especially in mature code bases, you can actually kind of do a whole bunch of automation on and the amount of time it takes to really, you know, actually massage them and make them happen is significantly smaller. But there's so much more thought that has to go into uh code bases because they can grow at such a faster rate now. Like when you make a bad decision here, it costs so much more. It used to only cost a little bit because you'd be like, "Okay, I think this is the right way to go. We all agree on this. Let's start going." And then within 1 month, you're like, "Dude, we produced like 30,000 lines of code together. And yeah, this is not good. What we've done is not good." Now, it's like, "Okay, what we produced is like 500,000 lines of code and it's an absolute disaster." Coding is roughly 20% of our time as software engineers. Typing is cheap. Wisdom is expensive. Wisdom is super expensive. Yes, I also am very curious about this, which is what happens like how does effect get created today? I feel like we're going to see less new libraries because so many so many people don't really concern themselves with the interfaces anymore. Right? There's so many people that have like that would have been great at creating something, but instead be like, "Oh, I just don't have to think about it anymore." Right? We got Jai. Yes, Gary Tan's blog is a very interesting example. He took Ruby on Rails, which is a language designed to be a blog and somehow wrote 250,000 lines of code plus 250,000 lines of test to have a blog that's generated in a language written for a blog. Crazy. Absolutely crazy. What's up, baby? Anyways, just wanted to yap about that for a little bit. It's kind of interesting. Kind of weird. I'm curious about all this. I'm curious where it's all going to go. What's going to happen in, you know, in 5 years? I have no idea. Oh, what's going to happen really in 5 months? I have no idea. I just assume that things are going to be nuts., Industry Wide AI Psychosis, YouTube, 01ks86amm40j178rm97grxek07.