Ep. 361: Do You Need an App Blocker?  ## Metadata - Author: [[Deep Questions with Cal Newport]] - Full Title: Ep. 361: Do You Need an App Blocker? - Category: #podcasts - URL: https://share.snipd.com/episode/dae47cfe-4eb3-406c-88d9-25010955b9a2 ## Document Notes ## Highlights - **Validating Phone Use Concerns** - Feeling trapped by your phone is valid, not a moral panic or weakness. - Expert dismissals often make people doubt their genuine concerns about phone overuse. Transcript: Cal Newport All right, I want to start with this first question. This is often missed, but it's the question of do I really need to use these apps? Now, what do I mean by the fact that this question is often missed? I want to step back here a little bit because I think this is something that a lot of professional commentators and technology critics miss when they think and talk about smartphone Use. What professionals like me don't always realize is that for the average user, this topic of smartphone and distraction and addiction can carry with it a lot of self-doubt and anxiety. A lot of people feel maybe like, hey, am I being overly concerned? Am I being overly sensitive? Is it dumb that I'm considering having to bring in technology to stop me from using my technology? Right? We've been perhaps bullied a little bit into thinking that, hey, if you're really worried about your phone, maybe you're falling into a moral panic. Maybe you're just being weak. It's worth taking a moment to address this because I think it's important. And it's something that we let slide. So I want to give you an idea of where these type of doubts come from. So, you know, if you're unfamiliar with this idea of people worrying about their troublesome phone use actually not really being a problem, I want to give you a sense where these type Of self-doubts come from. I picked up an article here that was in the BBC last year. They published this article. It had a paternalistic title. You're not actually addicted to your smartphone. In this article, the BBC interviewed an expert psychology professor. And what I want to do is give three quotes. These are three specific quotes from this article that I think do a good job of capturing the general type of ambivalence that a lot of professional commentators and critics seem to Put off about people's concerns about their phones. Okay. So here's the first quote from that article. We talk about things like, oh, I played that game or watched that show the other day and it was so addictive. What we mean there is that we really liked it or that we played it or used it a lot. In some cases, we say it with a bit of a negative connotation. And what we mean is that we've binge watched the TV show and we didn't feel so great afterwards. That's how we use the word addiction in day-to terms. Obviously, addiction has a different meaning, a very strict clinical definition to describe things like drug addictions or gambling addictions. All right, let's step back and look briefly at that quote. The move there in that quote, which we hear all the time from professional commentators, is to belittle the experience of feeling addicted to your phone. It's basically the expert in this article arguing, hey, addiction is for people with a drug problem or people who can't stop gambling away the mortgage money at the blackjack tables. You're using that term because you're unhappy about a guilty pleasure or you're just meaning that to me. You know, I like something and I do it a lot. So that's one move we get is a sort of professional belittling of the experience of the personal experience of addiction. All right, let me read you a second quote from this article from the expert here. So is there such a thing as healthy and unhealthy screen use? Yes, maybe. But I wonder whether we're asking the wrong sorts of questions or rather we're framing the questions in the wrong sort of way. We are fundamentally social animals, right? And social media, the clues in the title, it's a social connecting experience. All right. So in this quote, we see another common response from professional commentators. They're basically saying it's not a problem to use these apps a lot. They're actually good for you. You're wrong that you feel worried about it. They do lots of stuff that's good for you, and it would be a problem or a shame to miss out on those. All right, here's the third quote. It's understandable where those worries come from, because experientially, we all feel as though we've had bad experiences with our devices, and there's a need to hold the industry To account. This is the final move we often hear from professional commentators on these issues. This idea that, look, you're concerned you're voicing about your phone isn't really about your phone. It's a front for something else, like the fact that you dislike the tech industry. Maybe you had a bad experience or you're upset about something that a tech company did. So now you're going to get back at them by saying, oh, the technology they produce is bad, but the technology is actually fine. This is just a misplaced feeling of frustration. So if you're worried about your phone use and thinking about, hey, do I need something like an app blocker? And then you run into these sort of typical responses you sometimes get from the expert class. It can be really discouraging. It can make you feel at best like you're overreacting and worse like you're disingenuous or maybe just weak-willed. But here's the thing, and this is something I argued in my email newsletter that went out two weeks ago. By the way, if you don't subscribe to my weekly newsletter, you should probably do so at KellNewport.com. The newsletter that came out two weeks ago was titled, Trust Your Moral Intuition About Phones. And the core argument of that is your instincts aren't wrong. If you feel really worried about your phones, if this is why you're interested in app blockers, then you're likely right to be worried, even if that professor in the BBC tells you, no, That's not how you really feel. So to try to emphasize this, to sort of validate your concerns about your phone and your need to use things like app blockers, I'm going to read you a different quote. This one comes from the poet Lisa Wells in a New York Times op-ed she wrote about the experience of quitting social media and how it reminded her about her battle to quit smoking. Here's Lisa. About a year ago, I noticed a distressing tendency in myself to drift off while the people I loved were talking. It didn't matter if they were talking about a book they had read or recent health problems or crushing grief or revelations from therapy. Never before had I struggled to listen, but now I couldn't help checking out. Several times in the last year, my husband has had to ask the middle of a conversation, where did you go? Where did I go? Nowhere good. Usually my mind returned me to the small computer in my pocket, to an unanswered email, to a like or a retweet, to a comment I found threatening or flattering. Suffice to say, I went away. In giving my attention to the device, I withheld it from the people I valued most. And there were some other troubling symptoms. It was hard to read or write for sustained periods, which is concerning because that is my job. I was forcing myself to push through a handful of pages before reaching for the phone as a reward, orienting toward the activities I loved as if they were chores and toward the objects As a source of pleasure. It was more often a source of anxiety. I hadn't deliberately chosen to worship my smartphone, but when you repeatedly bow your head to something, stroking it thousands of times a day, it begins to shine like an idol. All right, so this is my answer to our first question. Do you need an app blocker? If you feel trapped by your phone, like it's becoming your idol, and that the people that you love and the activities that you used to cherish have now become obstacles that you have to Ignore, just try to grind through before you get back to your next digital fix, you're not wrong to want help. You're not falling into a moral panic. You're ([Time 0:05:28](https://share.snipd.com/snip/95a8e64c-3bd1-4d3a-9085-0f0f76448672)) - **Use Screen Time Passcode Lock** - Use built-in tools like iPhone Screen Time for light app blocking. - Add a passcode and involve someone to set it to prevent you from disabling blocks. Transcript: Cal Newport All right. Question number two, which app blocker should you use? This question is a little bit trickier than you might imagine as you're going to see, at least from a technological perspective. As I looked into this more, it turned out to be more complicated than I realized. But let me tell you how I'm going to approach this question because the first thing that became clear when I thought about app blocking is that there's different categories of solutions. They differ by the intensity of their intervention. Not everyone needs the same level of intensity of intervention. It depends really where you are, your honest assessment of how much trouble you're having. So what I'm going to do here is organize our discussion of app blockers into three levels of intensity of intervention, and we'll go from the most moderate to the most severe. All right, the first level of intensity of intervention would be to use, for example, on your iPhone, just the built-in screen time, oops, to use the built-in screen time app, So there's A built-in service in iOS, there's something similar in Android plus apps, it could do something similar that allows you to have pretty fine-grained control over the apps that you Use. So an Apple screen time app, for example, you can begin by just getting a report of the total amount of time you're spending on different apps or categories of apps each week. Like for example, I might say you're using four hours of Instagram a day on average. When you're in the screen time app on iOS, you can also set limits on the total use of specific apps per day. Like I don't want to use more than one hour per day of Instagram, or you can create windows of time in which you're allowed to use a given app. So for example, you might say the Substack app is something I'm only allowed to access between six o'clock and eight o'clock. So for example, after work hours, but not after bedtime. All right. So you can set those directly. There's other apps. If you don't like the interface for screen time, there's other apps that allow you to do similar type of restrictions or windows. Freedom is a popular one. Some people like its interface better than the screen time interface. Here is the shortcoming and why this is a moderate level of intensity. Here's the shortcoming of this approach. These blocks can be easily turned off. So yes, you can set a limit for yourself in screen time. And it will enforce it like, hey, you've overcome your limit for the day, but you can just go in the settings on your phone and turn off screen time. Similarly, Freedom uses a VPN to block apps. So it actually runs all of your traffic through a Freedom VPN. So it can just throw out packets that are aimed at the apps you say you want to block. But you could just go in and turn off VPNs in your phone settings and the block is turned off. People get really good at this. You do it so fast that it can seem almost automatic. So who is this level of intervention for? I think it's for people that just need a little bit of nudges in the right direction. So for some people, it is enough just to see the statistics. You're like, oh man, I didn't realize I was using Instagram that much. All right, enough of that. Other people, they respond well to some minor friction, right? It says, oh no, you're outside of your window for using Instagram. Yes, they could go turn that off, but that friction is enough of a reminder for them like, oh, yeah, I didn't want to use this so much. Great. I'm not going to use the app right now. So for people that are using their phone too much, but with a little bit of reminder can make changes, that level of intervention is right for you. All right. Level two is getting into more serious app blocking. So here's the question, right? And this is the question I had. If screen time on your phone, for example, is easy to turn off, can't you use a stronger blocker that's harder to turn off? This is where the answer turns out to be dependent on the platform you're using. So if we're talking about a like a desktop machine or a laptop machine or an Android phone, then yes, there are stronger app blockers that are much harder to circumvent. An example I want to focus on is Cold Turkey, which has a super strong app blocking feature. They've really thought about the ways that people try to circumvent blockers and they've made it impossible to use these, right? So you can't easily turn it off in the settings. You can't go to, as some people will do with other blockers, to your task manager and your computer and try to shut down the process running it because it works lower level in the operating System. Rebooting the machine isn't gonna help either. It turns out the only way to circumvent something like a cold turkey app block is to do something like booting your computer from an external hard drive to get around where cold turkey Has been integrated into your operating system. It's such complicated hacker stuff that really it's a block that you can't really get around. But here is the thing. Most people, when they think about app problems, they're thinking about their phones. And for a lot of people, that phone is going to be an iPhone. You can't use an app like cold turkey in that way on your iPhone. Here's what happened. This is a story that I think a lot of people have missed. From the very beginning, Apple, which tries to tightly control their app experience in the app store, was worried about the idea of apps that could control other apps. To them, that was a violation of their controlled experience goal. If I could download some random app that starts messing around with other people's apps. So they made it difficult for that to happen. Now, there's workarounds. Companies like Freedom figured out workarounds like using VPNs. But then in 2018, Apple said, no, no, we're going to be very specific about it. You cannot have an app in the App Store that stops or limits uses of other apps. If you worry about distraction or phone overuse, the only game in town is screen time. So that is when they introduced screen time. And they said, this is the only way you're allowed to control your access to other apps. Third-party apps can't do this. Now, of course, the problem is, as we talked about in level one, screen time is easy to get around. You can just go to settings and turn it right off. So are you out of luck if you use iOS but need a higher level of intensity of intervention to your app usage? Not necessarily. So this is what the internet seems to think is the the right way to do this. You can set a passcode for your screen time settings. Now this is meant for kids. I'm giving a phone to a I want to control their screen time. I'm going to have a passcode for those settings that the kid doesn't know, so they can't turn it off. That is actually a very strong passcode. If that is in place, there is no easy way to circumvent the passcode to change or alter your screen time settings. ([Time 0:13:14](https://share.snipd.com/snip/d0ea18de-d18a-473e-99c7-df5bbd6f8bdb)) - **Cold Turkey for Severe Overuse** - Consider deleting social media accounts or switching to a dumb phone for severe phone overuse. - This cold turkey approach suits those with major app-related issues and little need for those apps. Transcript: Cal Newport This is to actually go cold turkey, not the app cold turkey, just go cold turkey. Delete any social media accounts you have. There's nothing to log into or switch from a smartphone to a dumb phone. So forget worrying about passcodes or who's going to help you or what your settings are. Just take the activity that you're worried about out of your life, throw out the Marlboros. Now, who would this final level of intensity be for? People who really struggle with certain apps and don't really have a great reason to be using them. Like there's not some scenarios where it's really important that they use them. And so they say, you know what, why don't I just take this out of my life altogether? That second level intervention will work well if I'm using this too much, but I do sometimes need to use them. The third level might just be the solution for, I have a real problem here, and let's be honest, unlike what that professor from the BBC article said, I don't really get a lot of value out Of these apps. All right, so those are the three levels of intervention. All right, so this brings us to our third question we're going to address. What's the end game here? Is the idea just to have app blockers on your phone for the rest of your life? ([Time 0:21:51](https://share.snipd.com/snip/d3d152ca-961a-411b-a9b7-502cee63cd18)) - **App Blockers Rewire Urges** - App blockers often quickly reduce urges to check distracting apps. - Using blockers temporarily retrains your brain, allowing long-term behavior change after removal. Transcript: Cal Newport Is the idea just to have app blockers on your phone for the rest of your life? Or is there a different outcome that we're hoping for? Well, here I have good news to share. When it comes to these app blockers, they tend to do an effective job relatively quickly of eliminating the urge to check the apps that you're blocking. In other words, it's more common than not that when you've had a blocker like this on for a moderate amount of time, you can then take it off and you won't feel immediately re-compelled To go back and check those apps that you were blocking. I want to return to Lisa Wells, who we quoted earlier about her phone addiction. Here's what she wrote about the experience of, in her case, she did a level three intervention and replaced her smartphone with a dumb phone. It took about 72 hours to teach my body that we had gone back to the old ways. Though I had assumed it would take much longer, the change was almost instantaneous. Moderation requires effort and willpower, but when the device is gone, there's nothing to resist. I can read a book for hours in a sitting, and when my loved ones speak, I hear the story that they're telling, which is to say I am free again to enjoy the things I've always loved, to worship The God I chose. So as you can see here, in just a few days, her mind began to adjust back to not looking at the phone or being that interested in it. If she left those blockers on, or in this case, use a dumb phone for, let's say, a full month, I would argue that even if she went back to a smartphone, it would not have that same pull over Her that it did before. So that's the good news about app blocking. It's not a permanent configuration. It is a training tool that can recalibrate your mind back in a useful way. All right, so this is what we got. That's my discussion of app blocking. It is okay if you feel like you have a problem. You're not broken or in a moral panic or just trying to lash out at tech companies. It's neither is it a quixotic pursuit either. App blocking works. You find the right level of intensity. Some are more complicated than others. You'll feel a change almost immediately. And over time, those deep grooves that get you rolling towards trying to use your phone more than you want, they will fill in. ([Time 0:22:53](https://share.snipd.com/snip/7b8bc73a-494e-4855-b229-0f0f52f6378e)) - **Set Boundaries on Admin Work** - Schedule regular, fixed times for admin tasks to avoid constant intrusion. - Use office hours and communication protocols to manage requests; don't strive to excel excessively at admin work. Transcript: Cal Newport The May exam cycle is just finished and I'm feeling burned out. Please, can you suggest ways in which I can put boundaries on my administrative work so that it doesn't encroach all waking hours? It wants to do that as well. I know your pain, V, having my own off and off administrative positions in academia. It is a hard beast. It's like a hydra. You try to cut off one head, three more grow back out, and all the heads are shaped like email requests for you to jump on calls. I do have four things to suggest here that I think help you corral administrative work in a position like this. All right. One autopilot scheduling, regular times on regular days for regularly occurring administrative work. So it doesn't just exist. Otherwise, if it just exists out there, your mind will say, if there's any work that people need me to do administratively, like I should probably work on it. Your mind is going to be like, my goal is to get to a place where there's nothing I need to do for this piece of my job before I'll work on anything else. But that never comes because as you finish some things, new things come in. And this is how you get trapped in the cycle of always sort of just reacting to the new stuff. So with an autopilot schedule, these are the times I work on the major things. It's on these days and these times and these locations. And that's when I do it. And if I'm not in one of those times, I don't think about the work. And if you build your autopilot schedule correctly, you keep up with things, but you're freeing yourself from this impossible goal of finish everything before I can move on. All right. Second idea is office hours. You really need to try to avoid unscheduled back and forth communication through email and Slack as your main way of collaborating with people. As much as possible, if there's questions or queries that can't be answered with a single message, stop by my office hours. You probably need to make those office hours every day, which is different than what a normal professor does. You need to defer people to it. Hey, just call me during my office hours next time you can. More and more defer things to quick synchronous conversation in these fixed times. It will save you from like constantly having to be in your inboxes. Number three, for everything else, have communication protocols. So if there's certain administrative work you have to do all the time, figure out this is how we're going to communicate about it. Here's the process. You put the form I need to sign into this shared document that I've linked to from the website, shared folder rather, and there's two days a week when I go to that shared folder and I process The things and then I move them into this other folder for you to take and the review and then you have to get back to me on Friday if you have any issues. And then I will pass it off to be submitted or whatever it is. But you have protocols for how common work unfolds to try to save you from having to just have like people emailing you attachments and files or saying, oh my God, this is due tomorrow. Can you do it? So you put together communication protocols. My book, A World Without Email, gets into more detail about how to design those. Finally, my fourth suggestion, you might have to have a touch of the Feynman. What I mean by that, and I wrote about this in a few of my books, Richard Feynman famously said his strategy for not drowning in admin work as he became a higher and higher level professor At Cornell was to just not be very good at it. He said, I actively cultivated a reputation as being irresponsible, and then people left me alone. They want to give you as much stuff to do. Now, you don't have to be as much of a malcontent as Feynman attempted to be, but it's okay not to be great at the admin work. By what I mean by it's not dropping the ball and not by doing it poorly, but also not being like you're doing it on your own schedule. I have autopilot. I have office hours. It gets done. I'm not the fastest. It gets done well. But it's not – I'm not like amazing from other people's perspective, which usually means you're like very responsive and you're on things all the time. I do a passable job on things that aren't core to my success as a professor, but I don't try to do a great job on them. And like I say, no more often, and people kind of get upset at me and I get, you know, some evil looks sometimes, but I do an amount of this work that I think is useful. Like it's okay. You don't have to please everyone. Your goal as a professor is not to be good at admin work. You're not rewarded for being good at admin work. So you don't have to put your considerable intellectual gifts and skills towards that goal, save them for things that matter more. ([Time 0:34:09](https://share.snipd.com/snip/c156da50-8cc7-405c-bfbf-4b685a2c9204)) - **Read Newsletters Like Magazines** - Curate newsletters and collect them in one folder or label. - Read newsletters periodically like a magazine to enjoy a focused, curated content experience. Transcript: Cal Newport Next up is Tim. I received several weekly newsletters, including your freshly revamped one. I find it difficult to read them. How do you read newsletters? Jesse, have you seen the revamped looking newsletter of mine? Oh, yeah. Yeah, we got. For those who haven't followed up, I've been trying to take my newsletter more seriously. It's once a week. The type of ideas we talk about on the show. More – I would say my newsletter has more editorializing, kind of more op-eds, more like commentary and critics, more so critiques than I have on the show. But I'm doing it every week now, and the new newsletter director, Nate, has cleaned up the act literally. So the newsletter itself actually looks better. For those who haven't seen it, and Jesse can attest this is true, the newsletter is now a giant image of my face. And then as you click on different parts of my face, there's a flash animation that brings across like text that's related to it. So what am I seeing this week? You click on my eyes and then like a bunch of flashing text goes by with something about that. And what am I listening to? You click on my ears. It does play like really disturbing sounds when you click the buttons. And it is unfortunately configured in such a way that you can't mute your phone or stop the sound. So it can be a little bit embarrassing to use in public. But I think it's better. It cost me $300,000 to program that too. So I think it was worth it. No, but it looks nice. So I'm, I'm a writer. I realized newsletter used to be a big part of my life. I want it to be a big part of life. I'm again, so calnewport.com to subscribe. All right, let's get back to the question. How do you read newsletters? Here's what I think you should do. Curate your newsletters, right? Like here's newsletters where I like these writers. We're in a golden age of newsletters right now where really good writers are writing really good newsletters that are either free or cost a small amount of money. I think this is a great thing. So you can have a curated selection of newsletters you really like. I think you should consume newsletters like a magazine, like the best magazine you've ever subscribed to because it is super bespoke to exactly your interest. So what does that mean? Have a folder or a label for newsletters. When they come in, you can just throw them into that folder or apply the label to if you're using Gmail. And then on a semi-regular basis, like once or twice a week, you can go through there and read those newsletters like you're flipping through a magazine. You can print them out. You can put them on an iPad and bring them to a coffee shop. Bring your laptop to a bench outside your office. Like I'm going to read through three or four of these at a time. And you see it as like this is a magazine that a bunch of writers I like all are contributing to every week. So I think that's the right way to do. Consolidate and read like a magazine. ([Time 0:41:11](https://share.snipd.com/snip/09e333c3-a6fe-4e74-b720-72d5a9640069)) - **Avoid Polyphasic Sleep Hacks** - Sleep as your body naturally needs, avoiding polyphasic sleep hacks. - Focus on doing fewer things if time is tight, rather than hacking sleep. Transcript: Cal Newport What are your thoughts on polyphasic sleep to prepare for deep work? It's basically multiple sleeps in a 24-hour period. Look, people have to sleep. Different people need more sleep than others, you can't really hack it. Just sleep. I mean people, especially in Silicon Valley, they love this idea, especially in the 90s. They love this idea that Leonardo da Vinci was a polyphasic sleeper and there's a way to sleep like a couple hours at a time and never like more than that. And you'll have all these more waking hours or whatever. And in the end, you get really tired and upset because we are evolved. Our species is evolved. They have a period of sleep and a period of wakefulness and all of our hormones and chemicals are oriented around it. And that's just what you have to do. And you have to try to get as much sleep as your body typically needs. And if you're finding a hard time doing this, like you don't have enough time to get everything done you want to do, do fewer things. But sleep is not something, it's like trying to get around eating. You kind of just have to do it the way we're evolved to do it. ([Time 0:44:23](https://share.snipd.com/snip/8b2b4f7f-58a8-47ee-b52f-db8a7f9c1c6f)) - **Engineer’s Daily Shutdown Ritual** - A software engineer created a daily shutdown ritual by logging work progress before leaving. - This ritual gave peace of mind and helped start the day with a clear focus despite micromanagement. Transcript: Cal Newport These are where people write in to talk about how they've applied the type of advice we talk about on the show in their own situations. Like today's case study comes from RGJ, who says, I work as a software engineer at a tech startup. About five or six months ago, my boss started requiring us to post updates on our work in a Slack channel once a day. We use a ticketing system similar to JIRA to track work, so the updates are just a couple of bullet points about our progress on our self-assigned tickets. While I do think this is a little heavy-handed and smells like micromanagement, I managed to spin this into a positive in my working life with some of your deep work principles. Thanks to your advice, I recognize an opportunity to set up some habits that have helped me excel in my work. I spend the last 5 to 10 minutes of each workday documenting what I did that day in a little journal note on my computer. This has become a clear shutdown ritual for me, something I'd never been able to maintain before. I've been pleasantly surprised with how much peace of mind it gives me on my commute home and in the evening. Also, each morning, I copy what I wrote in the journal, edit it a little, and post it to the Slack channel. This morning ritual gives me something to do right as I start work and gives my brain a little bit of time to warm up before jumping into the day's deep work session. The only downside is that I have to open slack in the morning, which can at times be distracting. At the bare minimum, I plan on carrying the shutdown ritual wherever I work, regardless of whether there's a mandate or not. Thank you again for codifying and sharing your heart and wisdom with the rest of us. Well, I appreciate that case study because shutdown rituals are fantastic. Giving your brain some closure each day on what you did or didn't do and making sure you're not leaving any open loops or urgent things you're not dealing with gives you peace. You get so much more relaxation and recharging, but also just enjoyment out of your non-work time if you take 10 minutes to really shut down your work time properly. So I'm sorry you have some micromanagement in your job, but I'm glad that it led as a side effect to having a shutdown ritual inserted into your life because it seems like it's doing a lot Of good. ([Time 0:45:25](https://share.snipd.com/snip/b95417c3-ee7d-4306-9389-e4b8e790ccb8)) - **Limits of AI Vibe Coding** - AI coding tools help mostly by completing boilerplate code and offering autocompletion. - Fully coding complex projects via AI remains unreliable and requires human knowledge and debugging skills. Transcript: Cal Newport I actually wanted to share, this is maybe a first for this segment, two tweets that were brought to my attention. I didn't discover them myself, but they were brought to my attention. I think they present an interesting, I wouldn't say challenge, but more of like a bit of a course correction on a current theme that's getting a lot of coverage in the tech media. And that is using AI, the computer program. Vibe coding, that you can just tell an AI model what you want a program to do, and it will produce the code to do what you want it to do without you really having to learn to code or use the code Is a topic that's being talked about all the time in the media. I mentioned this in a recent episode. There's a lot of things that aren't due to AI that are being ascribed to AI because of this narrative. So, you know, we have the post-pandemic tech industry contraction, which is leading to some layoffs and reduction in jobs at tech companies and computer science majors going down. A lot of reporters are credulously just jumping in and saying, this is because AI is taking those jobs because AI can computer program now. It's not true. That's not why those jobs are going away. We got into that, but this is a really popular theme right now that we're moving up the ladder of complexity of computer programs that AI can just do. This brings me to these two tweets that were recently brought to my attention. I'll give a hat tip here to Gary Marcus on helping to identify these. All right. So the first tweet, This comes from Agilia, and let me read it. Alas, just the other day, I tried vibe coding a non-trivial Unity shader, and after hours of trying to make it work, I went back to human-made tutorials and started writing it again from Scratch to get the desired result. I would love to just vibe code, but it doesn't work past hello world examples. LLM-enabled autocomplete is helpful for boilerplate to avoid rewriting the same thing 100 times. So what Agilia here is saying is she had a relatively straightforward thing she wanted. A Unity shader is like a bit of code that helps impact how Unity does shading. It's a 3D engine for making video games. And just asking AI models work didn't work. But she makes the point at the end, LMs are still useful for coding, but in the autocomplete setting of like they help you fill in boilerplate code so you don't have to remember like exactly What function calls are, what parameters functions have, or how to access different libraries. But in her experience, creating code from scratch wasn't working well. Let me read another tweet. This one comes from Stuart Winterter. Actually, no, he's retweeting this from someone else. This is actually coming from a Reddit thread. All right, so let me read this Reddit thread. I am giving up. I just want to say that I'm giving up on creating anything anymore. I was trying to create my little project, but every time there are more and more errors, and I'm sick of it. I'm working on it for about three months. I do not have any experience with coding and was doing everything through AI, but every time I want to change a little thing, I kill four days debugging and other things go south. So I don't have any more energy in me to work on this. It is hopeless. AI is just so stupid and will fix one thing but destroy 10 other things in your code. I am really sad because I was enjoying it in the beginning, but now it is just pain and rage. Hat down for those people who can create something and it's working without coding knowledge. All right, so here's the point I want to make. How is AI helping computer coding? There's two ways. One, if you're a serious software developer, there's tools like Cursor or Github Copilot that help you create your code by preventing you from having to go look things up. Library in your operating system and you don't quite know what the call is or what the parameters are, it'll just kind of give you a sample code for how to do that. And then you can edit it as needed. Or you want to create a new function for an event handler and you forget exactly what the format is, it can fill that in for you right away. So it is a very useful aid to your coding on the scale of when Google became big. It was a big quantum leap for computer programmers because now they could Google things from forums like Stack Overflow instead of having to go to reference books. Then there's this thing called vibe coding, which is where you basically just ask AI to create a program from scratch. This is really cool, but the things that I think these tweets make clear is that it has a variable success rate. If you're careful and you know a little something about coding and the thing you're coding is at the right level of complexity, it can really help you save some time. A friend of mine, for example, was telling me about how for his kid's school, they were having some sort of fundraising charity auction. And he was thinking like, hey, if we had like a little website and we could do this to a web interface, it'd be much easier than people like filling stuff into a shared spreadsheet or just Emailing their bids. He told me like it took a few hours. He's not a great coder and he used AI and it kind of helped him figure out what services to use and write the code. He's like, this is cool. And I don't think I would have done this if I had to go out and do it from scratch. It's a great example of vibe coding working. There's other examples where I think Kevin Roos had an example where he was able to like get a Tetris game to work right away, just asking AI. That's less impressive because just those games exist. And so it was basically just taking code that already existed. But as seen in these tweets, it's not a magic wand. And if you don't know enough about coding or what you're doing is complicated enough, or just the AI models on the cranky side of the bed that day, it can be really frustrating, right? It is more fickle than people think. So on one hand, we have these conversations about whole software industries are disappearing. Every job loss or lost computer science major anywhere is due to AI. We have this idea that it can already handle coding up to this level completely, and now it's going to get to the next level. The reality is more complicated. It sometimes works, sometimes doesn't. It depends on your comfort and what you're trying to do. A lot of people are struggling with it. My point here is I think coding is one of the areas, like I talked about in my big article on this last month, where AI is going to be very useful and has been very useful, but it is not yet a Magic wand or panacea. So even in this area where there is perhaps no better of target for the capabilities of a large language model, we're still in a sort of uncertain phase in terms of the rewards and success We're having. So progress is being made. There's these two tweets show, be very wary about people that are overly confident in talking about how much AI can do from a coding perspective. It really depends on the person and the context, and it is not yet a general solution to a whole class of type of programming efforts or jobs. ([Time 0:51:06](https://share.snipd.com/snip/13c33154-0a5c-4c54-8117-0d34d17d597d))