I wonder how many of the responses here bifurcate by age. The post resonates with me, but I am now in my early fifties. When I was in my 20's and 30's, I would have happily chased rabbits down all those holes, but now that time seems so brutally finite, I feel that anything encouring me to spend time on stuff other than what really matters is a strong negative. (Where "what matters" includes work, family, friends, and recreation).
When friends start dying within 10 years of your age, it's a hell of a wake up.
"I wish I'd made more throw away apps I never use" ... said no one on their death bed, ever.
I think the bifurcation is between people who want to write code and people who want to have the end product of the code.
People who want to write code hate AI because it's doing the part they wanted to do.
People who want the end product of the code love AI because they want anything that helps them get to the end product faster.
The person who wrote this post feels oddly in neither camp. They like playing with the AI and seeing what comes out the other end. Some of the projects they boast about having built aren't even usable projects, like when they had it mock up a UI of a product and then got bored and moved on to the next before writing a backend.
Well, you are learning something, just the thing you’re learning has an even shorter usable lifespan than programming languages, namely you’re learning what works to get useful responses from ai agents. Whether or not that has value to you is a different matter, but it’s worth bearing in mind something is being learned, even if it’s not engineering or programming.
Well, making throw-away things you never use could also be the definition of a hobby ... and hobbies are arguably good, and differ from "work" in that your goal is not to ship a product, but to scratch an itch. For some people AI assisted vibe coding scratches an itch. For others, not. Of course some hobbies turn into an unhealthy obsession, but that's not a new phenomenon.
Similar age here. And I have similar thoughts, although not about AI specifically. AI helps me get more done and not spend time on trivia and yak shaving, which is great. I do get more projects done, but those are projects I always wanted to do, just never had the time (or, sometimes the motivation, because of yak shaving tasks).
I think the biggest difference is that I no longer care about what people think about me and how I am perceived, so the motivation to publish my work went down to near zero. I used to build open source stuff, I no longer want to spend time on preparing stuff for publishing, making it available, dealing with people who will inevitably want something of me eventually. There just isn't enough time.
I can still be baited into responding on HN for some reason, and I am trying to work on that, because that is the ultimate waste of time.
I'm also in my fifties, and sold my first software (6502 assemblewr) when I was 17!
My younger self was always excited when the latest tech came out, when the latest MSDN arrived, etc. But the last 15 or so years, I totally lost interest. I still love writing code but the desire for the latest and greatest had fade.
These LLMs were dogshit for a while, but I would keep returning to them.
Now I am excited again.
I work on a large web project with lots of legacy that is slowly being rewritten and copilot and codex are helping a lot by first writing tests for the old code, and then converting to the new.
I thought we'd never finish, but now I can see how we can do it.
Agreed with this until the last sentence, haha! I recently have been building throwaway apps and it has helped me scratch a bucket list itch I've had since childhood. Father is a programmer but I could never figure it out until vibe coding.
I don't understand the "deathbed" perspective. Are you going to wish you made more hackernews comments on your deathbed? Probably not. Does this mean you should stop using hackernews?
If you optimized for minimizing deathbed regret perhaps you'd regret that on your deathbed!
If I have cogent thoughts on my deathbed I expect they'll be along the lines of "I wish I wasn't dying" and not regretting the many ways I enjoyed my time on Earth (which includes vibe coding apps nobody uses).
> Young people are always more eager to adapt new tech while older people yell at clouds.
That's not just because young people have time like GP explained, but it is also because young people haven't been through the endless rounds of getting beaten up at work over daring to suggest that the "old ways" of one's superiors might be outdated, inefficient or just plain wrong.
I don't thing the problem is AI, but the mindset and trainning. I have probably as many or more AI projects that this man has but they are extremely useful, even if most of them I won't even sell.
This is like a kid playing videogames instead of studying, you take the console away and force the kid in front of a book and the kid will spend most of his time looking at the wall and dreaming.
I am engineer with very deep programming background that have managed people, with real experience in the real world.
One of the best things about AIs is that you can test crazy ideas and create prototypes very fast. Only one in a hundred will work great in the real world, but you have to create the 100 before to know.
Creating the 100 before AI was extremely expensive, and took so much time.
For me it is liberating and gives me focus because I can spend so little time testing prototypes and spend real time in what is really important and works.
This is something I learned from game developers: If you are going to create a game, you spend a weekend testing the dynamics and the gameplay of your prototype to know if is is fun. You use boxes, no textures, no complex sounds of music.
Then if it works and is is so fun, you create the game! You can spend 2 years creating the game after that.
You don't spend two years doing a Game only to realise later that is not fun, and you either spend 3 more years or abandon it at this moment.
Only mentioning this because the OP did - but for me (also ADHD) it's kind of the opposite. I'm finishing side projects for the first time ever because I can actually get them working before I get bored of them. My projects are more infra-leaning, and not all of them get much use, but some do. Others let me explore certain ideas and then sometimes serve as a reference point later when I run into something that reminds me of that.
Diagnosed with ADHD, ultimately does not change anything for me even through i had the same idea as you. Reason is that i can now start even more stuff in parallel. And some part of them get finished more before i can just prompt more when in focus, but instead of finishing i add more features.
I've definitely spent too many sprints where LLMs told me that something would be easy and they could definitely do it, and then... 2 days later I'm still debugging their crap before it dawns on me... WTF am I doing with my time?!
Overall, I've built a memory safe programming language that solves a lot of problems I personally have - predominately in my spare time over 8 months - and I've learned A TON in the process.
I'm close to a release stage, and on top of that - I've built a lot of good tooling for Ruby that I think other people will find helpful once I polish it (especially if anyone plans to vibe code something non-trivial in Ruby - which I honestly wouldn't recommend).
But... I'm not really sure this is what I actually wanted to do with my time, and I'm constantly questioning how much time I'm sinking into this and why...
It started off as utter amazement of what LLMs can do, and then incredible frustration at what they can't do, and my unending desire to figure out why they're so bad at things so close to what they are exceptionally good at, and if there's anything I can do to bridge that gap.
That's partially what the language is designed for (before I even started using LLMs).
But after all this time... I'm not even sure I've really figured anything out tbh.
I feel the same. I not sure if I have ADHD, but I always have 2 mode of focusing on something, first one is shot burst of focus and other one is real locked in mode where I forgot to eat or drink water.
While second one is much better at delivering value it mostly activated on management/strategy games that I love :D
But with AI assisted coding now I can really work on my side project while having first focus mode. Im just writing or designing the parts I excited about, and then I let AI to handle boring task.
Wow. To me the point of code has always been the crazy ideas and playing around. I love to create just for me and every once in a while for others is ok too. If you only think of code as 'a tool to build useful things' and everything else as wasted then sure, this is the philosophy for you. However, creating a bunch of random not going to follow up on it but I explored and played moments seems like a plus and not a negative to me.
Compilers are orders of magnitude more reliable than LLMs. There's a reason the saying is "it's not a compiler error".
I have a standing challenge to my co-workers that valid compiler errors will be rewarded like a birthday party, with the baked goods, alcohol, or sweets of their choice. It's only been redeemed once, and I've found less than a dozen unreported compiler bugs myself.
I don't really see this as analogous. Yes, you do choose an abstraction level to operate at, I rarely think in terms of transistors, or even gates (which by your logic an assembly programmer should do).
But I often do think across adjacent abstraction levels, because abstractions are (varying levels of) leaky. Modern compilers are after many decades good enough and modern computers fast enough that it is rare that I need to dig into the assembly (but I happens, compiler explorer is in my bookmark bar in Firefox).
Other abstractions are far leakier, it is far more common that I look in wireshark to debug network issues, the application level view is often not enough.
One of the leakiest abstractions currently is LLMs. Maybe in a decade or three they will be good enough, but they aren't yet, that's for sure. At least for the hard realtime systems level programming I do. For code generation they often make enough mistakes that the time spent after review and fixes comes out in the wash, even for simple tools. Their use for bug finding, RAG and similar is however promising.
My lived experience over the past few months is proving you wrong. I started with your position and have since been able to see how good the tools are when properly used. I've also noticed a huge gap in ability among engineers and I think the gap is widening. My theory is that some folks have the premium tools and some don't and the ones that don't are sort of in this weird limbo where they are sort of stubbornly annoyed at the idea of having to pay for these things so they lash out. Understandable but ultimately self defeating. You can in-fact force the LLM to use any pattern you want. I encountered this recently with a hand made framework I wanted to upgrade. It did stuff I didn't like. But well guess what? I provided it new constraints and it started to do what I want. Be as opinionated as you want. That's the whole point. It's basically your intern.
> My theory is that some folks have the premium tools and some don't and the ones that don't are sort of in this weird limbo where they are sort of stubbornly annoyed at the idea of having to pay for these things so they lash out.
At my last job the employer paid for OpenAI access for all of us.
Baby sitting an LLM is not my idea of meaningful use of time. And reviewing code that someone else had an LLM spew out even less so.
I am not lashing out because I don’t have access to LLMs. I had access and I did try it plenty.
Where do you think those bugs reports for gcc and others come from? Some people do look at the assembly coming out of the compilers.
Currently the openbsd mailing list for port is currently going through a clang update and one of the main point is looking at all the packages that failed to build. I even took a long look at the usb stack and the audio subsystem of OpenBSD because of an issue I was having with my DAC.
I literally do packaging for a living and you are misunderstanding my point. Most people just take a binary and run it. There's no analysis of the assembly code. You might profile it and bench it after the fact but no one is sitting there looking at the assembly line by line unless there's a very very good reason and frankly LLMs are better at that type of investigative work. I know because I've been investigating some curious 1 in 100,000 segfaults recently and guess what? It took an LLM to build a tool to let us even hit that bug because it was basically impossible to do by hand and no one in the before times would have sat down to write the tool cause we would not have time so we would have just accepted that 1 in 100,000 requests are segfaulting. At least now I can actually fix the problem.
What's the reliability of compilers this day? How likely for a bug to be in your code and not in the compiler? I think it's close to 99.99...
So when you have a bug and a core dump, you can quickly load it in debugger, see the stack frame and then theorize a model for the bug to happen. If after verifying the source and having complete confidence that it's good, then you start looking at the assembly, most likely while single stepping with the debugger. But you rarely get to that point, because 99.99... it's your code.
That reliability is what AI tooling is lacking. It's exhausting monitoring the output because errors can be as simple as a minus character or the wrong comparison operator.
I'm usually compiling other people's code. Hitting that 1 in 100,000 issue in run time and then having to come up with patch. And then have to make sure it's okay in arm and amd64. The bug I'm thinking of is decidedly a human output and the LLM is cleaning up the slop.
Once you reach a certain skill level you really didn’t ever visit SO anymore. I basically just live in Postgres, Redis, Ruby and Rails documentation. Still do.
I don't think I was ever able to straight copy and paste from SO, everything needs adaptation, and code can often be simplified. And you need to understand your code. SO was useful, but nothing could be used copy pasted.
Maybe this is not the case if you are doing a dozen throwaway websites, but for anything serious that is an absolute requirement. I work in hard realtime safety critical code, think things like brake controllers, medical devices, auto pilots, etc. In my case industrial control systems. You need to have full control and documentation for your development process.
I agree with the notion it's an age thing, but not because I am old, but because the tools are different. When I was learning to program as a kid I blindly 'copied and pasted' from computer magazine. I typed everything in, not understanding what I was doing, and made mistakes. Then came the tedious problem of figuring why the code didn't work. What was the syntax error? Why was it wrong? Why did the computer crash when I poked the wrong memory address?
I learned to debug and built comprehension by typing it in, and built it as a practice. Later in life and career I learned the value of transcription rather than copying and pasting because it at the very least forced me to read and write what I was copying, and built the base and familiarity I needed to learn from what I was copying.
That extends to how I use AI today. I use AI tooling to explore the concept of what I am building, use spec based designs to build solid outlines, and scope individual coding sessions, so that even when I use AI to build it, I have read, edited, and managed the design, and when I run into parts that I don't consider boilerplate I treat it the same way, transcribe what was attempted to understand why it was failing, and make sure I understand what the AI is doing that I haven't done before.
Most code we write is boiler plate nonsense. Writing out React components manually doesn't make me better at writing C. Like it literally is a waste of my time. I could be focusing on "real problems" like the way light diffuses in my simulated atmosphere in my 3D engine. But no I gotta sit there and manually wire up the onclick event for some button or whatever because it doesn't pass the HN sniff test from some pedantic random. Yea trying to fix one of my old job's webpack build for 3 weeks sure did build character. My boss hated how long it took. I hated how long it took. And thank fuck I NEVER have to do it again.
Some things are easier to verify than they are to solve, right?
So if you see an answer on stack overflow, read it, comprehend it, and you can pretty easily mentally verify the correctness to a sufficient degree of confidence…
I don't think those two things are comparable, really.
With SO copy/paste, you still were undertaking the mental exercise (and reward) of thinking through hard problems, researching solutions, and assembling it yourself.
With AI, you literally outsource most or all of that. The way some people "vibe code", they barely are engaged with any of that process, if at all.
I think about it like I do video games: it's a lot of fun to play them, and while it can be interesting to watch someone else play, it's just not the same.
It’s not binary, it’s a plus until it’s not. I agree with the author that the problem is “what happens if code is free” can change the incentives so much you forget why you were there in the first place.
You’re very reasonable response may be “well, why don’t you just do more of what you want to do and less of what you don’t want to do” but that’s not how incentives work.
You could talk about revealed preferences, and how obviously if this person did these things maybe that’s obviously what he wanted to do. And great, feel good about that.
There’s an uncomfortable reality for most of us normies (maybe not popular with the libertarian HN crowd) that an increase in freedom can make it much more difficult to find meaning and purpose. Friction can be good actually.
I do theorize that this is one of the mechanisms by which productivity could be tanked by AI.
Author sounds like they are missing meaning in what they do. If they had a life mission, AI is just an aid in accomplishing that mission, and they wouldn't get sidetracked by all the unfulfilling projects (modulus the ADHD, that has its own bearing on the experience using AI, and is the most interesting part of this post to me).
Perhaps at a population scale AI inhibits people from finding fulfillment.
But on an anecdotal basis, "just go find something meaningful". For some of us that "hate the AI timeline", we are still finding purpose and fulfillment by applying AI toward our personal missions.
I agree with this. AI is a tool and amplifier. If one is already disciplined and strong-minded and has clear meaning and purpose in their life, AI is a very powerful aid in accomplishing missions. But I'm afraid most people in the world are not like that, too.
I've been having the opposite experience; I've been GAINING focus through AI use.
In my day, when there's something that is distracting me from moving my objectives forward, I'm asking "Can AI help me automate this?" The answer is surprisingly often "yes". I call these "rough edges" and have been doing a lot of work over the last few weeks to "file the rough edges down".
I feel like the whole blog and the point can be reversed. If your bottlenecks are meetings and emails, and you make an agent take notes and summarize things for you, you gained focus to work on what you find meaningful.
> He explains that this happens because knowledge work often relies on “pseudo productivity,” where visible busyness is treated as a proxy for real value. Digital tools reinforce this by making people look active: sending more messages, producing more drafts, attending more meetings, and generating more work artifacts. To avoid the trap, he recommends measuring real outcomes, identifying the true bottlenecks in one’s work, and separating deep work from shallow work so that digital tools support meaningful progress instead of consuming attention.
---
Like, you are just as well make the argument that if you replace the pseudo-work, you end up with 8 hours of deep work for things that bring you value.
> your bottlenecks are meetings [...], and you make an agent take notes and summarize things for you.
An agent taking notes and summarizing things is of no use. You are supposed to participate to a meeting, otherwise it is just a memo and the meeting doesn't have to take place. The correct solution is just to not attend it if you know you aren't requested to participate and are just here to grow the numbers and make your company waste money.
> The correct solution is just to not attend it if you know you aren't requested to participate and are just here to grow the numbers and make your company waste money.
If this argument actually worked in practice, the world would be a better place
Same. I focus on the part that matters. As someone with ADHD I feel like AI is a salve for my mind. I used to listen to intense EDM while working. Now I sit in silence and talk to my agents. I maintain inbox zero. I absorb and comment across all relevant projects, even outside my team. I literally feel like I have a support team for the first time.
> In recent times, at least once per month someone sends a screenshot for an awesome tool they are working on. I'm like whoa, that's really something and the sender is obviously proud and enthusiastic. I try not to ask, but am always thinking "and where will you market it?"
What a strange perspective. His dismissal of the long list of projects at the top is also odd.
What's wrong with making something cool and functional (if not "useful"), even if just for yourself, without any profit motive or plan to turn it into some huge business?
I spent the last weekend vibing some plugins for Quod Libet -- a custom bookmark/preview function, a click-to-jump lyrics sidebar, thinking about a search-within-lyrics thing now. It all works beautifully, but I have no illusions about it being some kind of moneymaker -- heck, I doubt it's even worth the time beautifying/minimizing the code to get it acceptable to submit to the Github. But it makes me happy and makes using my library more enjoyable. Isn't that enough? Do they go around asking garage tinkerers and hobby crafters what their marketing plan is, too?
Some people, if not most, will at some point look back (and forward) in their life and wonder if they made anything out of it. And what they are really asking is "how much of an impact on others have I made?"
Sure, if all you did was work on a hobby, that wouldn't deliver satisfaction. But a hobby as a part of a life rich with relationships and people depending on you? Seems like a worthwhile pursuit to me.
The problem is, that working on lots of little random code projects makes you fall into some kind of local minimum of overall joy and satisfaction. You are robbing yourself of the motivation of focusing on something truly substantial, that is still just a hobby, but the end result will leave you far more proud and fulfilled.
At the end of your life, if all you've done are little half baked throwaway projects, you might look back and realize one day you never made anything of any particular significance, just thrashed around building stuff people had already done so many times before that some unthinking, unfeeling LLM can spit it out almost verbatim just so you can say "me too".
This applies to more than just AI, it can be about any type of "side project" really, or any context where you have a wealth of so many possible options that focusing on one intensely forces you to deliberately ignore most of them.
An example for me lately is hackernews. I used to jump around wildy, looking at comments not really even reading articles. I felt like I was learning a lot. But lately I've taken another approach. Instead of clicking a bunch of things, I'm actually determining what is the most interesting article of the day, reading it thoroughly and truly thinking about it, and then after pausing for reflection, forming my own thoughts about it. I have found this to be a far more enriching experience than my previous habit. I think a lot of things in life turn out this way.
The only reason to use AI to build is when you don't really care too much about things, you just want something, anything. An image here, some code there, a ridiculous video. Cheap thrills with no soul required.
Idk, I'd put it in the same category as doing a crossword puzzle, building a LEGO set, doing some DIY task around the house, etc. A nice diversion that's not entirely creative but stimulating enough, and at the end you have something functional/interesting or at least satisfaction that a particular problem has been solved. It won't change the world or your life or make a million bucks, but not everything has to.
There is a difference between learning woodworking as a fun hobby that would allow you to make a chair for yourself vs. doing it in hopes of turning it into a profitable business venture that would make an impact on the world.
By the grandparent comment logic, there is no point in doing anything, unless it can somehow lead you to making an outsized impact on the world. Thus essentially declaring most hobby pursuits (that are done mostly just for the sake of fun and learning) as wasteful.
It puts me in mind of making a jig when woodworking. Make the thing when you need it. You don't need to maintain it, you don't need to sell it. It does it's job and you move on. If it does the job really well maybe you keep it around for next time, maybe you refine it if you use it often.
Never be ashamed of making useless things, the really useful things are hiding amongst them.
If having fun is interfering with your productivity, that isn't necessarily a problem, it is only a problem if it interferes with your livelihood.
If Robots are to take all our jobs, we need to retain our livelihoods. Then we all could perhaps have fun making the things we want to make for the pleasure of making them.
I too have ADHD, perhaps it is different for me because I began medication about the same time the models got good, but I have worked on some individual projects for longer than I could have earlier.
I don't spend all day typing prompts though. It's more of a step in, do a thing, then think about it while doing something else.
Wrote about what I think is the root cause of all the mania the other day [1]: we're addicted to speed, "moving fast," and anchor it all to a vague sense of "productivity."
AI reduces the time cost of making the initial product, bypassing the need for genuine commitment, investment, strong interest, and dedication - which are vital in keeping a project alive.
Every time you need to make an update, you need to bring up the old context, or otherwise get the AI up to speed, which especially if you're using one of the frontier models could be a significant financial drain long term.
You don't get the same dopamine hit too, because you're just making boring updates to something which you threw together in 5 minutes with zero effort. The time and financial cost of building all this stuff may have been better spent on one, good, properly architected project.
Maintaining the project manually also assumes you can quickly understand the codebase which has been produced, otherwise you're completely dependent on Anthropic and them maintaining prices which you can afford. Bearing in mind that as you add new features, the cost of getting the LLM to understand the project increases, right? I might have a naive perspective here.
All that being said - sometimes there really are one-off niche things that are just for personal use that you do continue to use long term. Usually the simpler stuff where you can easily grasp the codebase at a quick glance. It's also great for debugging back and forths.
Personally I just run my local setup with a bunch of MCP stuff and the primary way it helps me is to keep me functional and on task. In some ways it's good if the AI can supervise you as opposed to you supervising it - at least from an ADHD perspective.
It's an interesting idea for sure, I like this article and agree with it.
The lucky normies have work to do, and they use their attention to meet the challenges. Us unlucky different-brained folks operate more like we have a lot of attention, and we have learned to fill it with computer stuff. AI is great for filling it but it’s often ultra processed weirdness and doesn’t seem to leave a trail of learning and productivity.
What if you then use AI to try and maintain only one, a single product into which you’ll put your care and craft to try to make something that’s better than “some dopamine hits”?
That’s how I use it. I might be working on two or three features at a time (iterating, iterating, iterating…), but they’re all scoped and of user value; I don’t feel that I’m just off chasing rabbits.
But I’m also one of those people for whom the “fun” was always solving human problems rather than solving computer problems. I can see how if you are in the latter category AI has already sucked out a lot of joy and how rapidly project switching could be the least-unfun option.
As someone constantly nerd-sniped, the difficulty is that our instincts are still being formed about what this current era of AI tools can and cannot do.
So when a blocker or an idea pops up, it's very easy to use that magic-like tool to solve it quickly and then go back to whatever it's you were doing before.
However, if you care about the quality of your output, that won't be a quick detour. It will pile up with the other "quick" tasks you were doing simultaneously and that's how you end up with 5-10 sessions working on totally unrelated projects.
Sure, but for many folks the distraction is irresistible. It was difficult already to put care and craft into a product, having a slot machine for your attention makes it damn impossible.
"The output was unbridled garbage. Because the effort was removed, so was the commitment, and with the commitment the focus, and with the focus any meaningful product at all. " -- i've noticed that fact first when I started to use linux. It didn't click untill I've built Gentoo from handbook. The commitment: i've asked on Gentoo IRC and it turns out a lot of people had a similar story.
Then, I've built a keyboard for myself and I'm still using it. I liked the process and started to build them basically for giveaway. My hope was that it will help people who eager to switch to ergonomic keyboards but the bar is too high for them to build, to figure out things etc.. But it turned out that people who get it without this effort they just try, fail, and leave it dusting on the shelf. They lack commitment, nothing fuels their enthusiasm.
Your car can drive 100+mph, but it’s probably not a good idea to drive around corners, in town or in your neighborhood at top speed. LLMs are the same. They can go 1,000+MPH, but what you need and what’s safe is 20-80mph. Practice restraint and focus and LLMs will make you more productive and give you some good results.
I think they got most of it right. Whenever there is a tool that helps one super active, it is a one off cliff. Smart ones will figure that out and get back to using it for meaningful purpose. A small percentage weaker ones will get addicted.
This actually really resonates with me, particularly the part about his AI tools for blogging and note taking.
I have zero interest in AI note-taking apps. I write notes for myself to process the meaningful outcomes of a meeting. My notes are short, only capture stuff I actually think I will care about in the future, and after I've written them I have a better mental model of the meeting than I did before.
If I gave the task to an AI, no matter how advanced, it would produce much more unfocused content than the focused notes I am used to writing, and I would lose the process of synthesis that helps me absorb the meeting outcomes. More work product, but actually less productivity.
We're still in the phase where we're having our first reaction to the software development lifecycle with the help of AI. We're quickly starting to realize what AI is making cheap, and where the new bottlenecks are. How most people are currently using AI is rather naive and superficial. One-shotting only takes you so far.
Love the idea of the ADHD amplifier!! It’s so true, being a profound ADHD person i have made many (more than i’m willing to admit) throw away apps with AI.
I must say some are pretty useful and i use on a daily basis, but all could’ve an excel heheheeh
Love and hate AI
> On that last point, this technology is horrific for attention. It's a thermonuclear ADHD amplifier and I have seen the same effect in every single one of my adult friends. Folk running 3 screens simultaneously working on totally unrelated "projects" they have little hope of maintaining, and such little commitment to the outcome that the time is obviously wasted.
This part reminded me of a recent article and it’s interesting that he brings up ADHD because that’s probably the bigger issue then. Because what I got from the article and the related conversation, specifically the top comment:
> > Sometimes, tools don’t move the needle because there’s no needle to move.
> It reminds me of something my old CS mentor, now elderly, had said about LLMs a few months ago: "it's a force multiplier, but there has to be some force to multiply."
The fact that it turned out that “Human Bottlenecks” post was written by the same person who wrote “Notes on Managing ADHD” which I had printed and studied for tips not that long ago made sense.
So, to connect the dots, the fact he made all of those things without them being part of a bigger plan is, I think, the problem. In the framework of the above quote, there’s no needle there, nothing to multiply.
I’ve been trying to think more about whether what I’m doing is going somewhere, or if I can skip it and simplify things.
It seems like the author is overindexing on useful and underindexing on wonderful. He clearly had fun building these products — and in hindsight is disavowing them because they didn’t generate income? An oddly capitalist view of play.
Some really good points on how these bots are incentivized to reward mindless engagement though and the bit about voice transcription not producing useful writing landed. When the barrier to release drops the quality naturally does too.
I think the next stage of us learning to harness these tools is us building the ability to reach for excellence even when we are not required to. To accustom ourselves to going beyond minimum viable bar for functionality and to reach for qualities or standards beyond that which the AI brings to the table unaided. A new kind of engineering rigor.
I move that this was always true and is now only far more so.
In the old days, producing all those things would be tremendous learning opportunity. Today it's a pure waste, not producing income is not a problem, not producing anything is.
If it wasn’t a learning opportunity to build those things, that was the waste. You can learn from an AI far more easily than from a book — only now it’s far more easy not to and many people unconsciously choose that route.
Learning how to use AI effectively was the learning opportunity here, what was created is completely incidental. You're effectively obsessing over programming languages obscuring the machine code that actually runs. "Imagine all the missed learning opportunity of digging into all that machine code!"
Sure, but also, who cares? The machine code is completely incidental for most purposes.
I work with AI everyday, despite what many people suggest there is so little to learn. After a couple of hours you are good to go. You don't even need gstack.
This is patently false. I work with and on AI every day at multiple levels of the stack, and every day I'm learning massive new swathes of information. I'm honestly shocked how deep the field goes and how much more effective you can be with time. The floor is falling and the ceiling is rising and the gap between them is widening every day.
Maybe it depends on the task, but the biggest productivity gains are from boiler plate generation, and there it's as easy as "generate me the boiler plate". Even if you can learn some very specific workflows today they would be model dependent and mostly obsolete within a month or two.
Go off and run a comparison of Qwen 3.6 27B and GLM 5.1 GGUF (https://huggingface.co/ubergarm/GLM-5.1-GGUF) at IQ2_KL 261.988 GiB (2.985 BPW) and let me know if you learn anything.
Or maybe just compare Hermes vs OpenClaw for long-horizon personal agentic tasks. Which one performs better in offline inference personal finance analysis tasks?
Or read up on how the `/code-review` workflow works in Opus 4.8 and give me a guess as to how long it'll take Codex to implement it and which tool would be more appropriate for your engineering team (don't forget to include enterprise API token costs in workflows – it can spin up 100 agents in thirty seconds).
If you can figure out how to secure agents with simultaneous access to personal data and the internet to run unsupervised while avoiding the lethal trifecta (Willison, 2025) let me know.
> Go off and run a comparison of Qwen 3.6 27B and GLM 5.1 GGUF
You may as well ask to run a comparison between gnu libc 2.42 and musl 1.2.5.
> Hermes vs OpenClaw for long-horizon personal agentic tasks. Which one performs better in offline inference personal finance analysis tasks
What are those tasks? This and the paragraph just after seems very much like a XY problem where all the energy is focusing on resolving the Y, not the X. It's like discussing how we can reach the moon using cannons.
> If you can figure out how to secure agents with simultaneous access to personal data and the internet to run unsupervised while avoiding the lethal trifecta (Willison, 2025) let me know.
If you can figure out how to run user submitted JavaScript inside a webpage with access to the internet and other user personal data, you will have your answer. There's a reason we escape user input before rendering it within the browser. The browser is an executing agent and it doesn't differentiate between your markup and other data you choose to embed in it. Same things happens with the processor if you choose to mix input data with executable code.
Even with examples it's still not convincing. I'm working on real products so I don't have time to waste comparing models that won't be relevant next month.
Using AI effectively for long horizon tasks, like maintaining a large codebase, is a wide open field. No single AI is good at it autonomously. That means achieving the right balance of testing, formal specification of pre/post-conditions and invariants and manual review.
It's like having a naive but super knowledgeable junior developer starting under you. It's obvious you'd learn a lot in how to communicate, framing, specifications, and what kind of follow-up you'd need to do to ensure good results.
It is not just about not generating income, it is also about learning very little.
I like to compare AI to GPS navigation. At least my experience of it. With GPS, I enter my destination, follow the direction and get to it. Problem is, I have no idea how I got there, I didn't pay attention to the landmarks, time and orientation, only to the arrow on the screen telling me where I should go, I learned nothing and should I go back, I will need the GPS again. And if the GPS is wrong, maybe because some road closed and it didn't get the update, too bad.
One may argue that using AI is a skill, yeah, sure, as much as following an arrow on a navigation screen is. It is nothing like actual development/navigation.
Personally, I have a terrible sense of direction, so I fully embrace GPS, and importantly, it isn't my job, no one pays me to navigate (they would want their money back anyways :)). But programming is my job, and I believe that if I want to keep it, I have to offer more than mindless vibe coding, that is a part that anyone can do, and practicing is the way to go. And even without the capitalist view, passion is about doing things the hard way because it is more rewarding, the easy way is wonderful at first, but it gets boring quickly.
Now, more specifically for AI, I think it has its uses. It can be a good rapid prototyping tool. I used to write some quick and dirty scripts, but rewrote them completely in a different language, with proper design, once I realized it would grow in complexity and have to be maintained. The first part can be vibe coded, before scrapping everything and doing it over by hand before it starts to grow. It is not an AI problem, it is more like a language problem, plain english simply isn't great for telling computers what to do exactly, in fact it is not good enough for telling other people what to do precisely, that's why many professions evolved their own language, math, chemical diagrams, blueprints, music scores, etc... In fact, that why porting is what AI does best: it already has a precise description of what to do in a programming language, human programmers already did the hard work, the AI just has to translate into another programming language. In the best case scenario, someone even wrote unit test so the AI can go over if it screwed up.
There is pleasure in making something yourself. There is learning. There is pride.
With generative AI you are just stealing other people's work. You are learning nothing. Anything could have generated the same projects. There was no skill involved, just enough disposable income to pay for tokens.
And yes some people develop some weird psychosis and think that they did the thing and not the AI. Everyone else is vibe coding but they got the special sauce, the perfect prompts. They are delusional.
> And I don't read anything indicated they had fun.
Maybe I'm just projecting. I enjoy making things. Maybe they do, maybe they don't. Sounds like you don't.
> There is pleasure in making something yourself. There is learning. There is pride.
You're speaking second person, when you should really be speaking first person. You enjoy making everything yourself, by hand. That is fine. It's also your personal perspective.
> You are learning nothing.
If you really aren't learning anything, you're doing AI wrong.
> Everyone else is vibe coding but they got the special sauce, the perfect prompts. They are delusional.
The delusion here is constructing a strawman out of the worst qualities you can imagine and berating that instead of actually looking at what other people are doing and trying to work out what they're thinking / how they feel. I can guarantee you that virtually nobody thinks they are the only person that can prompt a particular piece of software into existence.
I know this post probably won't land with you, because I'm a little annoyed while I write it (if only because your post comes off emotional and annoyed as well) (and, sorry in advance), but I do encourage you to consider that perhaps there are other worldviews than the clearly embittered and deeply entrenched one you've espoused. And perhaps those other worldviews are more suited to surviving the oncoming storm.
This is not an AI problem. Or rather, AI just made it worse. Focus can be hard. The thing is, AI can help you focus, by making code maintenance easier too.
'Waiting for AI to finish' - even if it's only 1 minute segments, is real, especially if we are delegating. (Maybe I'm interrupted right now!)
But this - it's not the fault of the tool that you're not focused on building something useful, long lasting or material.
That's an entirely different question - and I think if you look into most people's 'experiment' folders, that tendency was always there. Just more code now.
For me the sweet spot has been next suggested edit. I’m still writing code but the autocomplete does make it faster. That’s made coding more fun for me. What’s not fun is prompting then waiting around to find out it’s not what you wanted.
I have a friend who considers himself very humanist. He is really into UBI and more into socialist than me. He is environmentally conscious, all the bells and whistles. He is also Catholic.
He always asked me to help him build this app and that app and thinks his ideas are million dollar ideas. He has ADHD.
Surprisingly, he really loves LLM. He doesn't care that LLM destroys knowledge worker bargains by stealing work without compensating the original authors. He doesn't care that LLM uses a lot of energy. He doesn't care that LLM will concentrate money in the hands of the few. He doesn't care that the Pope has a crusade against LLM. For someone with humanist tendencies this seems to contradict his beliefs.
All he cares is, "I can make apps now and my 5 year old kids are making games by prompting, and we can make money using this, those who don't will be left behind, including you".
article points out a real problem - simplicity is one of the hardest things to achieve. the act of reduction is important.
buts its a refreshing that there is an initial list of half baked projects, i suppose meant to evoke horror at the untidiness and wasted time. but honestly each of those projects sound cool as hell.
not necessarily durable - but who cares. i’d argue there is a skill, one that is different than traditional programming, that the author was building up over that period.
discipline is important. focus is hard. but allowing yourself to play is not a bad thing at all and i dont think building little interesting side projects should be a shameful act.
For 20 years, Google had access to infinite amount of human based Phds, and fresh computer science graduates, and effectively unlimited budgets...and have been "hiring the best" for 20 years straight.
They are a decreasing quality searching engine who shows ads. It has never been about intelligence, or lack of resources. Its about incentives and execution.
Your AI wont save you, or make you rich or increase your productivity.
AI make easy work even easier, at the same time it shortens the attention span making it more difficult to do any difficult work. That's why there is so little real progress despite huge productivity gains.
I think the lack of progress is lack of understanding. If everything is generated and not viewed, does it exist? Like if a tree fell in the forest. Strip the observer and suddenly there is no universe. Strip the engineer and there is no codebase
I have ADHD, and for the last 2+ years, virtually 100% of my AI-assisted coding has gone into one product, which is a SaaS that supports my family. I have no end of ideas for little side projects, things to spin off, components I can open source from what I’ve built, etc. But unlike when I was younger (I’m old now), I’ve been able to resist the siren song of the ADHD side quest, and instead channel that towards the one project I know I should be focused on.
In other words, the issue isn’t the AI subscription, it’s the ADHD.
It seems to me with the rise of astroturfing and lies and deceit being more normalized, the benefit of using these AIs goes disproportionately to the AI companies who get experienced senior engineers training data.
And they get to convince people to pay them to give away their most intimate nontraining data and secret ideas to a for profit entity.
Every time I try to let Claude go off and do stuff on its own it’s always pinging me to approve something. Even in auto mode. Impossible to really run at length without either constantly having your focus broken, or just running it with permissions disabled. I do it in a container from time to time, but then by the time I get back to it sometimes there’s just so much slop it’s impossible to reason.
It’s a way of working that I really despise and if it’s the future of the profession I want nothing to do with it.
> It's a thermonuclear ADHD amplifier and I have seen the same effect in every single one of my adult friends.
You make this sound like a bad thing. ADHD isn't always about attention deficit, although it is right there in the name. It's more about attention dysregulation. For those of us prone to hyperfocus, working with AI can provide the kinds of stimulation we crave. I can hardly remember a time when I've felt more engaged with my work, more productive, and more badass.
I actually enjoy the collaborative programming process, and was pair programming with folks before the term was coined. At the end of the day I have the satisfaction of browsing the pretty, readable, DRY, maintainable code we end up with after rounds of refactoring and back and forth. I have always employed linters and code formatters, and this is no different, and my standards are still the same. I yell at the clanker about code duplication, hard-coded assumptions, tightly coupled logic, and in the end, while I don't understand the details of every algorithm, I really understand what we've built and the architecture we've designed.
Absolutely. I can't tell you how many times I've been in a conversation and halfway through a sentence I need to whip out AI to scratch the mental itch so I can continue with the conversation.
But prior to this I would rabbit hole. I would try desperately to remember some nuance, or I would not be able to move off a point until I got the validation I was looking for.
The worst is when speaking a foreign language and I hit some complex word in my native language that isn't present in my foreign lexicon. My brain just halts. It wants THAT word or phrase, not a 3 minute detour describing a whole concept.
AI has empowered me to move past these unnecessarily difficult speed bumps in my thinking.
> I actually enjoy the collaborative programming process, and was pair programming with folks before the term was coined
Yep, the same here, I'm a long pair programming enjoyer, but I'd like to raise that collaboration is usually meant with a human being in the context of pp, and prompting and agent to execute a task is nothing like that.
Prompting an agent to execute a task assumes you know what the task should be, have done some research on available options, weighed the pros and cons of various approaches, bounced your ideas off a colleague, have written a few test programs to validate your assumptions, considered how the new code will integrate with existing systems, figured out the parts that you should have tests for, and have generally charted a path forward that gives you a reasonable chance of success.
> Except for the SaaS, almost none of this is useful and I don't want to maintain any of it.
So don’t. Nobody’s twisting your arm.
Nobody told the author to sit down and write a bunch of random useless stuff.
This is like blaming your bicycle for enabling you to stop at too many shops that you didn’t mean to go to when you originally meant to ride straight to the grocery store.
I don’t think he’s blaming the technology he’s saying that AI is like crack for people with certain types of ADHD who are always thinking up new projects or going down rabbit holes.
I can relate to this greatly I have started dozens of projects since last summer but have been having a hard time turning these into real value. Not even money but just something that people find useful beyond my own learnings.
> and such little commitment to the outcome that the time is obviously wasted.
Why is it wasted? A powerful new tool was invented, and enthusiasts are exploring ways to harness it. They'll come away with the skill to wield this new tool effectively. The programs they're writing are completely secondary.
AI makes single purpose throw away tools easy to create. This is GREAT. I had to migrate an old Windows 2012 file server share to SharePoint. Microsoft's tools don't work on this old OS. Their SharePoint migration tool running on other machines on the local network constantly failed for nebulous reasons. I finally got fed up and spent a few hours with Gemini Pro and Claude and created a sync tool using C# that does the migration and keeps the network share in sync with SharePoint until we do the final cutover. I don't expect to ever use this tool again, and that's totally fine. I'll still put it on GitHub in case someone has a use for it, but I'm not sure why I should lament the fact that this tool exists and may never see another use or the fact that I won't maintain it.
Don't waste your life playing with shiny new toys, sure, but learning how to use AI by creating things is not a waste of time.
>exploring AI as a lens in Marshall McLuhan-like thinking
I would be wary of using McLuhan-like media analysis of AI. His central argument is that media are tools that extend man's ability. A calculator or a spell checker extend our thinking and writing. AI does not extend those abilities so much as it completely replaces it.
The way in which it does resemble media is insofar as it captures the same urge that McLuhan wrote about to see ourselves extended into the world. McLuhan tied this to the myth of Narcissus. The difference is that where Narcissus falsely believed it wasn't him and fell in love with what he saw, we falsely believe the image we see is ourselves and fall in love with it.
A calculator does replace our ability to do math in a certain sense.
At the grocery store there's countless (no pun) opportunities to do math in the sense of comparing prices and calculating unit costs etc, but most people can't do that math easily in their head because the calculator has made that skill less important.
But people also don't pull out the calculator repeatedly to do this in the grocery store, so the math just doesn't get done.
I wonder how many of the responses here bifurcate by age. The post resonates with me, but I am now in my early fifties. When I was in my 20's and 30's, I would have happily chased rabbits down all those holes, but now that time seems so brutally finite, I feel that anything encouring me to spend time on stuff other than what really matters is a strong negative. (Where "what matters" includes work, family, friends, and recreation).
When friends start dying within 10 years of your age, it's a hell of a wake up.
"I wish I'd made more throw away apps I never use" ... said no one on their death bed, ever.
I think the bifurcation is between people who want to write code and people who want to have the end product of the code.
People who want to write code hate AI because it's doing the part they wanted to do.
People who want the end product of the code love AI because they want anything that helps them get to the end product faster.
The person who wrote this post feels oddly in neither camp. They like playing with the AI and seeing what comes out the other end. Some of the projects they boast about having built aren't even usable projects, like when they had it mock up a UI of a product and then got bored and moved on to the next before writing a backend.
I don’t think this is the full picture. Plenty of people who like writing code are happy to delete code too.
When you write the code you learn from writing the code.
When you have an LLM produce something and then delete, you didn’t learn much.
Well, you are learning something, just the thing you’re learning has an even shorter usable lifespan than programming languages, namely you’re learning what works to get useful responses from ai agents. Whether or not that has value to you is a different matter, but it’s worth bearing in mind something is being learned, even if it’s not engineering or programming.
You're learning to manage idiot savants, which is a very useful skill.
I’m not sure that’s true. All my side projects exist to scratch my own itch, so the appeal to hop straight from design to done is really appealing.
But it’s never really that straightforward.
Well, making throw-away things you never use could also be the definition of a hobby ... and hobbies are arguably good, and differ from "work" in that your goal is not to ship a product, but to scratch an itch. For some people AI assisted vibe coding scratches an itch. For others, not. Of course some hobbies turn into an unhealthy obsession, but that's not a new phenomenon.
Similar age here. And I have similar thoughts, although not about AI specifically. AI helps me get more done and not spend time on trivia and yak shaving, which is great. I do get more projects done, but those are projects I always wanted to do, just never had the time (or, sometimes the motivation, because of yak shaving tasks).
I think the biggest difference is that I no longer care about what people think about me and how I am perceived, so the motivation to publish my work went down to near zero. I used to build open source stuff, I no longer want to spend time on preparing stuff for publishing, making it available, dealing with people who will inevitably want something of me eventually. There just isn't enough time.
I can still be baited into responding on HN for some reason, and I am trying to work on that, because that is the ultimate waste of time.
I'm also in my fifties, and sold my first software (6502 assemblewr) when I was 17!
My younger self was always excited when the latest tech came out, when the latest MSDN arrived, etc. But the last 15 or so years, I totally lost interest. I still love writing code but the desire for the latest and greatest had fade.
These LLMs were dogshit for a while, but I would keep returning to them.
Now I am excited again.
I work on a large web project with lots of legacy that is slowly being rewritten and copilot and codex are helping a lot by first writing tests for the old code, and then converting to the new.
I thought we'd never finish, but now I can see how we can do it.
It's brought a bit of the fun back into the game.
Agreed with this until the last sentence, haha! I recently have been building throwaway apps and it has helped me scratch a bucket list itch I've had since childhood. Father is a programmer but I could never figure it out until vibe coding.
I don't understand the "deathbed" perspective. Are you going to wish you made more hackernews comments on your deathbed? Probably not. Does this mean you should stop using hackernews?
If you optimized for minimizing deathbed regret perhaps you'd regret that on your deathbed!
If I have cogent thoughts on my deathbed I expect they'll be along the lines of "I wish I wasn't dying" and not regretting the many ways I enjoyed my time on Earth (which includes vibe coding apps nobody uses).
Of course it is. Young people are always more eager to adapt new tech while older people yell at clouds. AI usage is MUCH higher among young people.
No, GenZ is AI critical, no matter what a one-liner optimized model trained to discredit comments says.
> Young people are always more eager to adapt new tech while older people yell at clouds.
That's not just because young people have time like GP explained, but it is also because young people haven't been through the endless rounds of getting beaten up at work over daring to suggest that the "old ways" of one's superiors might be outdated, inefficient or just plain wrong.
I don't thing the problem is AI, but the mindset and trainning. I have probably as many or more AI projects that this man has but they are extremely useful, even if most of them I won't even sell.
This is like a kid playing videogames instead of studying, you take the console away and force the kid in front of a book and the kid will spend most of his time looking at the wall and dreaming.
I am engineer with very deep programming background that have managed people, with real experience in the real world.
One of the best things about AIs is that you can test crazy ideas and create prototypes very fast. Only one in a hundred will work great in the real world, but you have to create the 100 before to know.
Creating the 100 before AI was extremely expensive, and took so much time.
For me it is liberating and gives me focus because I can spend so little time testing prototypes and spend real time in what is really important and works.
This is something I learned from game developers: If you are going to create a game, you spend a weekend testing the dynamics and the gameplay of your prototype to know if is is fun. You use boxes, no textures, no complex sounds of music.
Then if it works and is is so fun, you create the game! You can spend 2 years creating the game after that.
You don't spend two years doing a Game only to realise later that is not fun, and you either spend 3 more years or abandon it at this moment.
Only mentioning this because the OP did - but for me (also ADHD) it's kind of the opposite. I'm finishing side projects for the first time ever because I can actually get them working before I get bored of them. My projects are more infra-leaning, and not all of them get much use, but some do. Others let me explore certain ideas and then sometimes serve as a reference point later when I run into something that reminds me of that.
Diagnosed with ADHD, ultimately does not change anything for me even through i had the same idea as you. Reason is that i can now start even more stuff in parallel. And some part of them get finished more before i can just prompt more when in focus, but instead of finishing i add more features.
I guess I'm between you and OP.
I've definitely spent too many sprints where LLMs told me that something would be easy and they could definitely do it, and then... 2 days later I'm still debugging their crap before it dawns on me... WTF am I doing with my time?!
Overall, I've built a memory safe programming language that solves a lot of problems I personally have - predominately in my spare time over 8 months - and I've learned A TON in the process.
I'm close to a release stage, and on top of that - I've built a lot of good tooling for Ruby that I think other people will find helpful once I polish it (especially if anyone plans to vibe code something non-trivial in Ruby - which I honestly wouldn't recommend).
But... I'm not really sure this is what I actually wanted to do with my time, and I'm constantly questioning how much time I'm sinking into this and why...
It started off as utter amazement of what LLMs can do, and then incredible frustration at what they can't do, and my unending desire to figure out why they're so bad at things so close to what they are exceptionally good at, and if there's anything I can do to bridge that gap.
That's partially what the language is designed for (before I even started using LLMs).
But after all this time... I'm not even sure I've really figured anything out tbh.
I feel the same. I not sure if I have ADHD, but I always have 2 mode of focusing on something, first one is shot burst of focus and other one is real locked in mode where I forgot to eat or drink water. While second one is much better at delivering value it mostly activated on management/strategy games that I love :D But with AI assisted coding now I can really work on my side project while having first focus mode. Im just writing or designing the parts I excited about, and then I let AI to handle boring task.
Wow. To me the point of code has always been the crazy ideas and playing around. I love to create just for me and every once in a while for others is ok too. If you only think of code as 'a tool to build useful things' and everything else as wasted then sure, this is the philosophy for you. However, creating a bunch of random not going to follow up on it but I explored and played moments seems like a plus and not a negative to me.
OP paid a machine to have those moments instead of him.
As opposed to the old days when people would just blindly copy/paste random shit from Stackoverflow?
Ya'll need to stop with this cope. It's not a good look.
> As opposed to the old days when people would just blindly copy/paste random shit from Stackoverflow?
Many of the people who are complaining about AI vibecoding today also didn't blindly copy/paste from StackOverflow in the past.
I'm sure they were blindly accepting the Assembly coming out of the compiler.
Compilers are orders of magnitude more reliable than LLMs. There's a reason the saying is "it's not a compiler error".
I have a standing challenge to my co-workers that valid compiler errors will be rewarded like a birthday party, with the baked goods, alcohol, or sweets of their choice. It's only been redeemed once, and I've found less than a dozen unreported compiler bugs myself.
I don't really see this as analogous. Yes, you do choose an abstraction level to operate at, I rarely think in terms of transistors, or even gates (which by your logic an assembly programmer should do).
But I often do think across adjacent abstraction levels, because abstractions are (varying levels of) leaky. Modern compilers are after many decades good enough and modern computers fast enough that it is rare that I need to dig into the assembly (but I happens, compiler explorer is in my bookmark bar in Firefox).
Other abstractions are far leakier, it is far more common that I look in wireshark to debug network issues, the application level view is often not enough.
One of the leakiest abstractions currently is LLMs. Maybe in a decade or three they will be good enough, but they aren't yet, that's for sure. At least for the hard realtime systems level programming I do. For code generation they often make enough mistakes that the time spent after review and fixes comes out in the wash, even for simple tools. Their use for bug finding, RAG and similar is however promising.
My lived experience over the past few months is proving you wrong. I started with your position and have since been able to see how good the tools are when properly used. I've also noticed a huge gap in ability among engineers and I think the gap is widening. My theory is that some folks have the premium tools and some don't and the ones that don't are sort of in this weird limbo where they are sort of stubbornly annoyed at the idea of having to pay for these things so they lash out. Understandable but ultimately self defeating. You can in-fact force the LLM to use any pattern you want. I encountered this recently with a hand made framework I wanted to upgrade. It did stuff I didn't like. But well guess what? I provided it new constraints and it started to do what I want. Be as opinionated as you want. That's the whole point. It's basically your intern.
> My theory is that some folks have the premium tools and some don't and the ones that don't are sort of in this weird limbo where they are sort of stubbornly annoyed at the idea of having to pay for these things so they lash out.
At my last job the employer paid for OpenAI access for all of us.
Baby sitting an LLM is not my idea of meaningful use of time. And reviewing code that someone else had an LLM spew out even less so.
I am not lashing out because I don’t have access to LLMs. I had access and I did try it plenty.
So tldr you don't have it now and have no frame of reference.
Ah yes the “you’re not paying the ai labs enough” argument.
My total usage for this month is less than one days pay.
> out of the compiler
You mean a source that's been tested on billions of PCs over 45+ years?
As opposed to a LLM which outputs code that barely works on my machine™?
If you think that's a good analogy you're in the wronf industry ....
Where do you think those bugs reports for gcc and others come from? Some people do look at the assembly coming out of the compilers.
Currently the openbsd mailing list for port is currently going through a clang update and one of the main point is looking at all the packages that failed to build. I even took a long look at the usb stack and the audio subsystem of OpenBSD because of an issue I was having with my DAC.
I literally do packaging for a living and you are misunderstanding my point. Most people just take a binary and run it. There's no analysis of the assembly code. You might profile it and bench it after the fact but no one is sitting there looking at the assembly line by line unless there's a very very good reason and frankly LLMs are better at that type of investigative work. I know because I've been investigating some curious 1 in 100,000 segfaults recently and guess what? It took an LLM to build a tool to let us even hit that bug because it was basically impossible to do by hand and no one in the before times would have sat down to write the tool cause we would not have time so we would have just accepted that 1 in 100,000 requests are segfaulting. At least now I can actually fix the problem.
What's the reliability of compilers this day? How likely for a bug to be in your code and not in the compiler? I think it's close to 99.99...
So when you have a bug and a core dump, you can quickly load it in debugger, see the stack frame and then theorize a model for the bug to happen. If after verifying the source and having complete confidence that it's good, then you start looking at the assembly, most likely while single stepping with the debugger. But you rarely get to that point, because 99.99... it's your code.
That reliability is what AI tooling is lacking. It's exhausting monitoring the output because errors can be as simple as a minus character or the wrong comparison operator.
I'm usually compiling other people's code. Hitting that 1 in 100,000 issue in run time and then having to come up with patch. And then have to make sure it's okay in arm and amd64. The bug I'm thinking of is decidedly a human output and the LLM is cleaning up the slop.
Once you reach a certain skill level you really didn’t ever visit SO anymore. I basically just live in Postgres, Redis, Ruby and Rails documentation. Still do.
SO was just an example. If you try to tell me you've never copied/pasted code before I know you are lying.
I don't think I was ever able to straight copy and paste from SO, everything needs adaptation, and code can often be simplified. And you need to understand your code. SO was useful, but nothing could be used copy pasted.
Maybe this is not the case if you are doing a dozen throwaway websites, but for anything serious that is an absolute requirement. I work in hard realtime safety critical code, think things like brake controllers, medical devices, auto pilots, etc. In my case industrial control systems. You need to have full control and documentation for your development process.
I agree with the notion it's an age thing, but not because I am old, but because the tools are different. When I was learning to program as a kid I blindly 'copied and pasted' from computer magazine. I typed everything in, not understanding what I was doing, and made mistakes. Then came the tedious problem of figuring why the code didn't work. What was the syntax error? Why was it wrong? Why did the computer crash when I poked the wrong memory address?
I learned to debug and built comprehension by typing it in, and built it as a practice. Later in life and career I learned the value of transcription rather than copying and pasting because it at the very least forced me to read and write what I was copying, and built the base and familiarity I needed to learn from what I was copying.
That extends to how I use AI today. I use AI tooling to explore the concept of what I am building, use spec based designs to build solid outlines, and scope individual coding sessions, so that even when I use AI to build it, I have read, edited, and managed the design, and when I run into parts that I don't consider boilerplate I treat it the same way, transcribe what was attempted to understand why it was failing, and make sure I understand what the AI is doing that I haven't done before.
Most code we write is boiler plate nonsense. Writing out React components manually doesn't make me better at writing C. Like it literally is a waste of my time. I could be focusing on "real problems" like the way light diffuses in my simulated atmosphere in my 3D engine. But no I gotta sit there and manually wire up the onclick event for some button or whatever because it doesn't pass the HN sniff test from some pedantic random. Yea trying to fix one of my old job's webpack build for 3 weeks sure did build character. My boss hated how long it took. I hated how long it took. And thank fuck I NEVER have to do it again.
Anyone worth their salt looked down on copy/paste from Stackoverflow, let alone blindly doing so.
Where does this idea come from that good programmers were ever cool with that?
Some things are easier to verify than they are to solve, right?
So if you see an answer on stack overflow, read it, comprehend it, and you can pretty easily mentally verify the correctness to a sufficient degree of confidence…
I guess I’m not worth my salt.
I don't think those two things are comparable, really.
With SO copy/paste, you still were undertaking the mental exercise (and reward) of thinking through hard problems, researching solutions, and assembling it yourself.
With AI, you literally outsource most or all of that. The way some people "vibe code", they barely are engaged with any of that process, if at all.
I think about it like I do video games: it's a lot of fun to play them, and while it can be interesting to watch someone else play, it's just not the same.
Playing with legos is fine if you can afford them.
It’s not binary, it’s a plus until it’s not. I agree with the author that the problem is “what happens if code is free” can change the incentives so much you forget why you were there in the first place.
You’re very reasonable response may be “well, why don’t you just do more of what you want to do and less of what you don’t want to do” but that’s not how incentives work.
You could talk about revealed preferences, and how obviously if this person did these things maybe that’s obviously what he wanted to do. And great, feel good about that.
There’s an uncomfortable reality for most of us normies (maybe not popular with the libertarian HN crowd) that an increase in freedom can make it much more difficult to find meaning and purpose. Friction can be good actually.
I do theorize that this is one of the mechanisms by which productivity could be tanked by AI.
Author sounds like they are missing meaning in what they do. If they had a life mission, AI is just an aid in accomplishing that mission, and they wouldn't get sidetracked by all the unfulfilling projects (modulus the ADHD, that has its own bearing on the experience using AI, and is the most interesting part of this post to me).
Perhaps at a population scale AI inhibits people from finding fulfillment.
But on an anecdotal basis, "just go find something meaningful". For some of us that "hate the AI timeline", we are still finding purpose and fulfillment by applying AI toward our personal missions.
I agree with this. AI is a tool and amplifier. If one is already disciplined and strong-minded and has clear meaning and purpose in their life, AI is a very powerful aid in accomplishing missions. But I'm afraid most people in the world are not like that, too.
I've been having the opposite experience; I've been GAINING focus through AI use.
In my day, when there's something that is distracting me from moving my objectives forward, I'm asking "Can AI help me automate this?" The answer is surprisingly often "yes". I call these "rough edges" and have been doing a lot of work over the last few weeks to "file the rough edges down".
Has it been part of a longer term shift or just something for the past few weeks? What will happen once all the rough edges are filed down?
I feel like the whole blog and the point can be reversed. If your bottlenecks are meetings and emails, and you make an agent take notes and summarize things for you, you gained focus to work on what you find meaningful.
> He explains that this happens because knowledge work often relies on “pseudo productivity,” where visible busyness is treated as a proxy for real value. Digital tools reinforce this by making people look active: sending more messages, producing more drafts, attending more meetings, and generating more work artifacts. To avoid the trap, he recommends measuring real outcomes, identifying the true bottlenecks in one’s work, and separating deep work from shallow work so that digital tools support meaningful progress instead of consuming attention.
---
Like, you are just as well make the argument that if you replace the pseudo-work, you end up with 8 hours of deep work for things that bring you value.
> your bottlenecks are meetings [...], and you make an agent take notes and summarize things for you.
An agent taking notes and summarizing things is of no use. You are supposed to participate to a meeting, otherwise it is just a memo and the meeting doesn't have to take place. The correct solution is just to not attend it if you know you aren't requested to participate and are just here to grow the numbers and make your company waste money.
> The correct solution is just to not attend it if you know you aren't requested to participate and are just here to grow the numbers and make your company waste money.
If this argument actually worked in practice, the world would be a better place
Same. I focus on the part that matters. As someone with ADHD I feel like AI is a salve for my mind. I used to listen to intense EDM while working. Now I sit in silence and talk to my agents. I maintain inbox zero. I absorb and comment across all relevant projects, even outside my team. I literally feel like I have a support team for the first time.
> In recent times, at least once per month someone sends a screenshot for an awesome tool they are working on. I'm like whoa, that's really something and the sender is obviously proud and enthusiastic. I try not to ask, but am always thinking "and where will you market it?"
What a strange perspective. His dismissal of the long list of projects at the top is also odd.
What's wrong with making something cool and functional (if not "useful"), even if just for yourself, without any profit motive or plan to turn it into some huge business?
I spent the last weekend vibing some plugins for Quod Libet -- a custom bookmark/preview function, a click-to-jump lyrics sidebar, thinking about a search-within-lyrics thing now. It all works beautifully, but I have no illusions about it being some kind of moneymaker -- heck, I doubt it's even worth the time beautifying/minimizing the code to get it acceptable to submit to the Github. But it makes me happy and makes using my library more enjoyable. Isn't that enough? Do they go around asking garage tinkerers and hobby crafters what their marketing plan is, too?
Some people, if not most, will at some point look back (and forward) in their life and wonder if they made anything out of it. And what they are really asking is "how much of an impact on others have I made?"
YMMV
Sure, if all you did was work on a hobby, that wouldn't deliver satisfaction. But a hobby as a part of a life rich with relationships and people depending on you? Seems like a worthwhile pursuit to me.
and marketing something is the answer?
Indeed, if you make something that improved many people’s lives, you‘ll probably see it as a great success.
That often requires marketing it.
The problem is, that working on lots of little random code projects makes you fall into some kind of local minimum of overall joy and satisfaction. You are robbing yourself of the motivation of focusing on something truly substantial, that is still just a hobby, but the end result will leave you far more proud and fulfilled.
At the end of your life, if all you've done are little half baked throwaway projects, you might look back and realize one day you never made anything of any particular significance, just thrashed around building stuff people had already done so many times before that some unthinking, unfeeling LLM can spit it out almost verbatim just so you can say "me too".
This applies to more than just AI, it can be about any type of "side project" really, or any context where you have a wealth of so many possible options that focusing on one intensely forces you to deliberately ignore most of them.
An example for me lately is hackernews. I used to jump around wildy, looking at comments not really even reading articles. I felt like I was learning a lot. But lately I've taken another approach. Instead of clicking a bunch of things, I'm actually determining what is the most interesting article of the day, reading it thoroughly and truly thinking about it, and then after pausing for reflection, forming my own thoughts about it. I have found this to be a far more enriching experience than my previous habit. I think a lot of things in life turn out this way.
The only reason to use AI to build is when you don't really care too much about things, you just want something, anything. An image here, some code there, a ridiculous video. Cheap thrills with no soul required.
Idk, I'd put it in the same category as doing a crossword puzzle, building a LEGO set, doing some DIY task around the house, etc. A nice diversion that's not entirely creative but stimulating enough, and at the end you have something functional/interesting or at least satisfaction that a particular problem has been solved. It won't change the world or your life or make a million bucks, but not everything has to.
Very much agreed.
There is a difference between learning woodworking as a fun hobby that would allow you to make a chair for yourself vs. doing it in hopes of turning it into a profitable business venture that would make an impact on the world.
By the grandparent comment logic, there is no point in doing anything, unless it can somehow lead you to making an outsized impact on the world. Thus essentially declaring most hobby pursuits (that are done mostly just for the sake of fun and learning) as wasteful.
It puts me in mind of making a jig when woodworking. Make the thing when you need it. You don't need to maintain it, you don't need to sell it. It does it's job and you move on. If it does the job really well maybe you keep it around for next time, maybe you refine it if you use it often.
Never be ashamed of making useless things, the really useful things are hiding amongst them.
If having fun is interfering with your productivity, that isn't necessarily a problem, it is only a problem if it interferes with your livelihood.
If Robots are to take all our jobs, we need to retain our livelihoods. Then we all could perhaps have fun making the things we want to make for the pleasure of making them.
I too have ADHD, perhaps it is different for me because I began medication about the same time the models got good, but I have worked on some individual projects for longer than I could have earlier.
I don't spend all day typing prompts though. It's more of a step in, do a thing, then think about it while doing something else.
Wrote about what I think is the root cause of all the mania the other day [1]: we're addicted to speed, "moving fast," and anchor it all to a vague sense of "productivity."
[1] https://news.ycombinator.com/item?id=48326501
AI reduces the time cost of making the initial product, bypassing the need for genuine commitment, investment, strong interest, and dedication - which are vital in keeping a project alive.
Every time you need to make an update, you need to bring up the old context, or otherwise get the AI up to speed, which especially if you're using one of the frontier models could be a significant financial drain long term.
You don't get the same dopamine hit too, because you're just making boring updates to something which you threw together in 5 minutes with zero effort. The time and financial cost of building all this stuff may have been better spent on one, good, properly architected project.
Maintaining the project manually also assumes you can quickly understand the codebase which has been produced, otherwise you're completely dependent on Anthropic and them maintaining prices which you can afford. Bearing in mind that as you add new features, the cost of getting the LLM to understand the project increases, right? I might have a naive perspective here.
All that being said - sometimes there really are one-off niche things that are just for personal use that you do continue to use long term. Usually the simpler stuff where you can easily grasp the codebase at a quick glance. It's also great for debugging back and forths.
Personally I just run my local setup with a bunch of MCP stuff and the primary way it helps me is to keep me functional and on task. In some ways it's good if the AI can supervise you as opposed to you supervising it - at least from an ADHD perspective.
It's an interesting idea for sure, I like this article and agree with it.
The lucky normies have work to do, and they use their attention to meet the challenges. Us unlucky different-brained folks operate more like we have a lot of attention, and we have learned to fill it with computer stuff. AI is great for filling it but it’s often ultra processed weirdness and doesn’t seem to leave a trail of learning and productivity.
It seems like there should be a middle ground where you occasionally write a side project for fun, but not dozens of them just because you can?
Also, if nobody uses them, they don't need to be maintained. You can shut them down with no regrets.
What if you then use AI to try and maintain only one, a single product into which you’ll put your care and craft to try to make something that’s better than “some dopamine hits”?
That’s how I use it. I might be working on two or three features at a time (iterating, iterating, iterating…), but they’re all scoped and of user value; I don’t feel that I’m just off chasing rabbits.
But I’m also one of those people for whom the “fun” was always solving human problems rather than solving computer problems. I can see how if you are in the latter category AI has already sucked out a lot of joy and how rapidly project switching could be the least-unfun option.
As someone constantly nerd-sniped, the difficulty is that our instincts are still being formed about what this current era of AI tools can and cannot do.
So when a blocker or an idea pops up, it's very easy to use that magic-like tool to solve it quickly and then go back to whatever it's you were doing before.
However, if you care about the quality of your output, that won't be a quick detour. It will pile up with the other "quick" tasks you were doing simultaneously and that's how you end up with 5-10 sessions working on totally unrelated projects.
Shades of https://xkcd.com/1319/
Sure, but for many folks the distraction is irresistible. It was difficult already to put care and craft into a product, having a slot machine for your attention makes it damn impossible.
That's funny, that's the exact conclusion I'm starting to come to
Quoting:
"Because the effort was removed, so was the commitment, and with the commitment the focus, and with the focus any meaningful product at all."
This is the truth. Otherwise known as "easy come, easy go".
"The output was unbridled garbage. Because the effort was removed, so was the commitment, and with the commitment the focus, and with the focus any meaningful product at all. " -- i've noticed that fact first when I started to use linux. It didn't click untill I've built Gentoo from handbook. The commitment: i've asked on Gentoo IRC and it turns out a lot of people had a similar story.
Then, I've built a keyboard for myself and I'm still using it. I liked the process and started to build them basically for giveaway. My hope was that it will help people who eager to switch to ergonomic keyboards but the bar is too high for them to build, to figure out things etc.. But it turned out that people who get it without this effort they just try, fail, and leave it dusting on the shelf. They lack commitment, nothing fuels their enthusiasm.
Your car can drive 100+mph, but it’s probably not a good idea to drive around corners, in town or in your neighborhood at top speed. LLMs are the same. They can go 1,000+MPH, but what you need and what’s safe is 20-80mph. Practice restraint and focus and LLMs will make you more productive and give you some good results.
I think they got most of it right. Whenever there is a tool that helps one super active, it is a one off cliff. Smart ones will figure that out and get back to using it for meaningful purpose. A small percentage weaker ones will get addicted.
Nothing different from all innovations.
This actually really resonates with me, particularly the part about his AI tools for blogging and note taking.
I have zero interest in AI note-taking apps. I write notes for myself to process the meaningful outcomes of a meeting. My notes are short, only capture stuff I actually think I will care about in the future, and after I've written them I have a better mental model of the meeting than I did before.
If I gave the task to an AI, no matter how advanced, it would produce much more unfocused content than the focused notes I am used to writing, and I would lose the process of synthesis that helps me absorb the meeting outcomes. More work product, but actually less productivity.
We're still in the phase where we're having our first reaction to the software development lifecycle with the help of AI. We're quickly starting to realize what AI is making cheap, and where the new bottlenecks are. How most people are currently using AI is rather naive and superficial. One-shotting only takes you so far.
It's interesting to look at a man without ai in 2026
Love the idea of the ADHD amplifier!! It’s so true, being a profound ADHD person i have made many (more than i’m willing to admit) throw away apps with AI. I must say some are pretty useful and i use on a daily basis, but all could’ve an excel heheheeh Love and hate AI
It's pretty much impossible for any positive story about AI to get upvoted here, isn't it? Positivity and normalcy doesn't get clicks
> On that last point, this technology is horrific for attention. It's a thermonuclear ADHD amplifier and I have seen the same effect in every single one of my adult friends. Folk running 3 screens simultaneously working on totally unrelated "projects" they have little hope of maintaining, and such little commitment to the outcome that the time is obviously wasted.
This part reminded me of a recent article and it’s interesting that he brings up ADHD because that’s probably the bigger issue then. Because what I got from the article and the related conversation, specifically the top comment:
> > Sometimes, tools don’t move the needle because there’s no needle to move.
> It reminds me of something my old CS mentor, now elderly, had said about LLMs a few months ago: "it's a force multiplier, but there has to be some force to multiply."
From: https://news.ycombinator.com/item?id=48254336
The fact that it turned out that “Human Bottlenecks” post was written by the same person who wrote “Notes on Managing ADHD” which I had printed and studied for tips not that long ago made sense.
So, to connect the dots, the fact he made all of those things without them being part of a bigger plan is, I think, the problem. In the framework of the above quote, there’s no needle there, nothing to multiply.
I’ve been trying to think more about whether what I’m doing is going somewhere, or if I can skip it and simplify things.
It seems like the author is overindexing on useful and underindexing on wonderful. He clearly had fun building these products — and in hindsight is disavowing them because they didn’t generate income? An oddly capitalist view of play.
Some really good points on how these bots are incentivized to reward mindless engagement though and the bit about voice transcription not producing useful writing landed. When the barrier to release drops the quality naturally does too.
I think the next stage of us learning to harness these tools is us building the ability to reach for excellence even when we are not required to. To accustom ourselves to going beyond minimum viable bar for functionality and to reach for qualities or standards beyond that which the AI brings to the table unaided. A new kind of engineering rigor.
I move that this was always true and is now only far more so.
In the old days, producing all those things would be tremendous learning opportunity. Today it's a pure waste, not producing income is not a problem, not producing anything is.
If it wasn’t a learning opportunity to build those things, that was the waste. You can learn from an AI far more easily than from a book — only now it’s far more easy not to and many people unconsciously choose that route.
Learning how to use AI effectively was the learning opportunity here, what was created is completely incidental. You're effectively obsessing over programming languages obscuring the machine code that actually runs. "Imagine all the missed learning opportunity of digging into all that machine code!"
Sure, but also, who cares? The machine code is completely incidental for most purposes.
I work with AI everyday, despite what many people suggest there is so little to learn. After a couple of hours you are good to go. You don't even need gstack.
This is patently false. I work with and on AI every day at multiple levels of the stack, and every day I'm learning massive new swathes of information. I'm honestly shocked how deep the field goes and how much more effective you can be with time. The floor is falling and the ceiling is rising and the gap between them is widening every day.
Maybe it depends on the task, but the biggest productivity gains are from boiler plate generation, and there it's as easy as "generate me the boiler plate". Even if you can learn some very specific workflows today they would be model dependent and mostly obsolete within a month or two.
That would be more convincing if you put up two or more examples of what is there to learn.
Go off and run a comparison of Qwen 3.6 27B and GLM 5.1 GGUF (https://huggingface.co/ubergarm/GLM-5.1-GGUF) at IQ2_KL 261.988 GiB (2.985 BPW) and let me know if you learn anything.
Or maybe just compare Hermes vs OpenClaw for long-horizon personal agentic tasks. Which one performs better in offline inference personal finance analysis tasks?
Or read up on how the `/code-review` workflow works in Opus 4.8 and give me a guess as to how long it'll take Codex to implement it and which tool would be more appropriate for your engineering team (don't forget to include enterprise API token costs in workflows – it can spin up 100 agents in thirty seconds).
If you can figure out how to secure agents with simultaneous access to personal data and the internet to run unsupervised while avoiding the lethal trifecta (Willison, 2025) let me know.
> Go off and run a comparison of Qwen 3.6 27B and GLM 5.1 GGUF
You may as well ask to run a comparison between gnu libc 2.42 and musl 1.2.5.
> Hermes vs OpenClaw for long-horizon personal agentic tasks. Which one performs better in offline inference personal finance analysis tasks
What are those tasks? This and the paragraph just after seems very much like a XY problem where all the energy is focusing on resolving the Y, not the X. It's like discussing how we can reach the moon using cannons.
> If you can figure out how to secure agents with simultaneous access to personal data and the internet to run unsupervised while avoiding the lethal trifecta (Willison, 2025) let me know.
If you can figure out how to run user submitted JavaScript inside a webpage with access to the internet and other user personal data, you will have your answer. There's a reason we escape user input before rendering it within the browser. The browser is an executing agent and it doesn't differentiate between your markup and other data you choose to embed in it. Same things happens with the processor if you choose to mix input data with executable code.
Even with examples it's still not convincing. I'm working on real products so I don't have time to waste comparing models that won't be relevant next month.
Using AI effectively for long horizon tasks, like maintaining a large codebase, is a wide open field. No single AI is good at it autonomously. That means achieving the right balance of testing, formal specification of pre/post-conditions and invariants and manual review.
It's like having a naive but super knowledgeable junior developer starting under you. It's obvious you'd learn a lot in how to communicate, framing, specifications, and what kind of follow-up you'd need to do to ensure good results.
Unless you just happen to work in a domain where the code you generate every day is very common in the AI training data, this isn't true.
It is not just about not generating income, it is also about learning very little.
I like to compare AI to GPS navigation. At least my experience of it. With GPS, I enter my destination, follow the direction and get to it. Problem is, I have no idea how I got there, I didn't pay attention to the landmarks, time and orientation, only to the arrow on the screen telling me where I should go, I learned nothing and should I go back, I will need the GPS again. And if the GPS is wrong, maybe because some road closed and it didn't get the update, too bad.
One may argue that using AI is a skill, yeah, sure, as much as following an arrow on a navigation screen is. It is nothing like actual development/navigation.
Personally, I have a terrible sense of direction, so I fully embrace GPS, and importantly, it isn't my job, no one pays me to navigate (they would want their money back anyways :)). But programming is my job, and I believe that if I want to keep it, I have to offer more than mindless vibe coding, that is a part that anyone can do, and practicing is the way to go. And even without the capitalist view, passion is about doing things the hard way because it is more rewarding, the easy way is wonderful at first, but it gets boring quickly.
Now, more specifically for AI, I think it has its uses. It can be a good rapid prototyping tool. I used to write some quick and dirty scripts, but rewrote them completely in a different language, with proper design, once I realized it would grow in complexity and have to be maintained. The first part can be vibe coded, before scrapping everything and doing it over by hand before it starts to grow. It is not an AI problem, it is more like a language problem, plain english simply isn't great for telling computers what to do exactly, in fact it is not good enough for telling other people what to do precisely, that's why many professions evolved their own language, math, chemical diagrams, blueprints, music scores, etc... In fact, that why porting is what AI does best: it already has a precise description of what to do in a programming language, human programmers already did the hard work, the AI just has to translate into another programming language. In the best case scenario, someone even wrote unit test so the AI can go over if it screwed up.
> He clearly had fun building these products
The author did not build those products. AI did.
And I don't read anything indicated they had fun.
There is pleasure in making something yourself. There is learning. There is pride.
With generative AI you are just stealing other people's work. You are learning nothing. Anything could have generated the same projects. There was no skill involved, just enough disposable income to pay for tokens.
And yes some people develop some weird psychosis and think that they did the thing and not the AI. Everyone else is vibe coding but they got the special sauce, the perfect prompts. They are delusional.
> And I don't read anything indicated they had fun.
Maybe I'm just projecting. I enjoy making things. Maybe they do, maybe they don't. Sounds like you don't.
> There is pleasure in making something yourself. There is learning. There is pride.
You're speaking second person, when you should really be speaking first person. You enjoy making everything yourself, by hand. That is fine. It's also your personal perspective.
> You are learning nothing.
If you really aren't learning anything, you're doing AI wrong.
> Everyone else is vibe coding but they got the special sauce, the perfect prompts. They are delusional.
The delusion here is constructing a strawman out of the worst qualities you can imagine and berating that instead of actually looking at what other people are doing and trying to work out what they're thinking / how they feel. I can guarantee you that virtually nobody thinks they are the only person that can prompt a particular piece of software into existence.
I know this post probably won't land with you, because I'm a little annoyed while I write it (if only because your post comes off emotional and annoyed as well) (and, sorry in advance), but I do encourage you to consider that perhaps there are other worldviews than the clearly embittered and deeply entrenched one you've espoused. And perhaps those other worldviews are more suited to surviving the oncoming storm.
This is not an AI problem. Or rather, AI just made it worse. Focus can be hard. The thing is, AI can help you focus, by making code maintenance easier too.
AI makes code maintenance harder
The point about interruptions is valid.
'Waiting for AI to finish' - even if it's only 1 minute segments, is real, especially if we are delegating. (Maybe I'm interrupted right now!)
But this - it's not the fault of the tool that you're not focused on building something useful, long lasting or material.
That's an entirely different question - and I think if you look into most people's 'experiment' folders, that tendency was always there. Just more code now.
That's on us.
AI makes me far more productive, but I’ve lost quite a bit too. There’s less fun in coding these days, and it leaves me feeling adrift at times.
For me the sweet spot has been next suggested edit. I’m still writing code but the autocomplete does make it faster. That’s made coding more fun for me. What’s not fun is prompting then waiting around to find out it’s not what you wanted.
Coding has engaging parts, and plenty of drudgery. AI is generally good at the latter, and you don't need to use it for the former.
I wrote about this a little bit today too. You’re up against a dopamine machine that writes code for you.
https://www.tyleo.com/blog/the-terminal-star
A lot of good comes out but it can be hard to separate from the parts that just take advantage of your brain.
I have a friend who considers himself very humanist. He is really into UBI and more into socialist than me. He is environmentally conscious, all the bells and whistles. He is also Catholic.
He always asked me to help him build this app and that app and thinks his ideas are million dollar ideas. He has ADHD.
Surprisingly, he really loves LLM. He doesn't care that LLM destroys knowledge worker bargains by stealing work without compensating the original authors. He doesn't care that LLM uses a lot of energy. He doesn't care that LLM will concentrate money in the hands of the few. He doesn't care that the Pope has a crusade against LLM. For someone with humanist tendencies this seems to contradict his beliefs.
All he cares is, "I can make apps now and my 5 year old kids are making games by prompting, and we can make money using this, those who don't will be left behind, including you".
article points out a real problem - simplicity is one of the hardest things to achieve. the act of reduction is important.
buts its a refreshing that there is an initial list of half baked projects, i suppose meant to evoke horror at the untidiness and wasted time. but honestly each of those projects sound cool as hell. not necessarily durable - but who cares. i’d argue there is a skill, one that is different than traditional programming, that the author was building up over that period.
discipline is important. focus is hard. but allowing yourself to play is not a bad thing at all and i dont think building little interesting side projects should be a shameful act.
For 20 years, Google had access to infinite amount of human based Phds, and fresh computer science graduates, and effectively unlimited budgets...and have been "hiring the best" for 20 years straight.
This what they have been spending their human tokens on: https://killedbygoogle.com/
They are a decreasing quality searching engine who shows ads. It has never been about intelligence, or lack of resources. Its about incentives and execution.
Your AI wont save you, or make you rich or increase your productivity.
The author has a problem with spending too much time at computer.
AI make easy work even easier, at the same time it shortens the attention span making it more difficult to do any difficult work. That's why there is so little real progress despite huge productivity gains.
I think the lack of progress is lack of understanding. If everything is generated and not viewed, does it exist? Like if a tree fell in the forest. Strip the observer and suddenly there is no universe. Strip the engineer and there is no codebase
I have ADHD, and for the last 2+ years, virtually 100% of my AI-assisted coding has gone into one product, which is a SaaS that supports my family. I have no end of ideas for little side projects, things to spin off, components I can open source from what I’ve built, etc. But unlike when I was younger (I’m old now), I’ve been able to resist the siren song of the ADHD side quest, and instead channel that towards the one project I know I should be focused on.
In other words, the issue isn’t the AI subscription, it’s the ADHD.
It seems to me with the rise of astroturfing and lies and deceit being more normalized, the benefit of using these AIs goes disproportionately to the AI companies who get experienced senior engineers training data.
And they get to convince people to pay them to give away their most intimate nontraining data and secret ideas to a for profit entity.
Every time I try to let Claude go off and do stuff on its own it’s always pinging me to approve something. Even in auto mode. Impossible to really run at length without either constantly having your focus broken, or just running it with permissions disabled. I do it in a container from time to time, but then by the time I get back to it sometimes there’s just so much slop it’s impossible to reason.
It’s a way of working that I really despise and if it’s the future of the profession I want nothing to do with it.
> It's a thermonuclear ADHD amplifier and I have seen the same effect in every single one of my adult friends.
You make this sound like a bad thing. ADHD isn't always about attention deficit, although it is right there in the name. It's more about attention dysregulation. For those of us prone to hyperfocus, working with AI can provide the kinds of stimulation we crave. I can hardly remember a time when I've felt more engaged with my work, more productive, and more badass.
I actually enjoy the collaborative programming process, and was pair programming with folks before the term was coined. At the end of the day I have the satisfaction of browsing the pretty, readable, DRY, maintainable code we end up with after rounds of refactoring and back and forth. I have always employed linters and code formatters, and this is no different, and my standards are still the same. I yell at the clanker about code duplication, hard-coded assumptions, tightly coupled logic, and in the end, while I don't understand the details of every algorithm, I really understand what we've built and the architecture we've designed.
For me it's been useful as an idea categorizer: "oh well, that turned out to be a crap idea."
It's allowed me to clear out some long-standing brush on the forest floor. And burn it down once or twice.
Absolutely. I can't tell you how many times I've been in a conversation and halfway through a sentence I need to whip out AI to scratch the mental itch so I can continue with the conversation.
But prior to this I would rabbit hole. I would try desperately to remember some nuance, or I would not be able to move off a point until I got the validation I was looking for.
The worst is when speaking a foreign language and I hit some complex word in my native language that isn't present in my foreign lexicon. My brain just halts. It wants THAT word or phrase, not a 3 minute detour describing a whole concept.
AI has empowered me to move past these unnecessarily difficult speed bumps in my thinking.
> I actually enjoy the collaborative programming process, and was pair programming with folks before the term was coined
Yep, the same here, I'm a long pair programming enjoyer, but I'd like to raise that collaboration is usually meant with a human being in the context of pp, and prompting and agent to execute a task is nothing like that.
Prompting an agent to execute a task assumes you know what the task should be, have done some research on available options, weighed the pros and cons of various approaches, bounced your ideas off a colleague, have written a few test programs to validate your assumptions, considered how the new code will integrate with existing systems, figured out the parts that you should have tests for, and have generally charted a path forward that gives you a reasonable chance of success.
AI is great for programs but every product ever kinda sucks if you don't understand a lot of things computers are pretty bad at, generally.
I think this blames the technology way too much.
> Except for the SaaS, almost none of this is useful and I don't want to maintain any of it.
So don’t. Nobody’s twisting your arm.
Nobody told the author to sit down and write a bunch of random useless stuff.
This is like blaming your bicycle for enabling you to stop at too many shops that you didn’t mean to go to when you originally meant to ride straight to the grocery store.
I don’t think he’s blaming the technology he’s saying that AI is like crack for people with certain types of ADHD who are always thinking up new projects or going down rabbit holes.
I can relate to this greatly I have started dozens of projects since last summer but have been having a hard time turning these into real value. Not even money but just something that people find useful beyond my own learnings.
> and such little commitment to the outcome that the time is obviously wasted.
Why is it wasted? A powerful new tool was invented, and enthusiasts are exploring ways to harness it. They'll come away with the skill to wield this new tool effectively. The programs they're writing are completely secondary.
AI makes single purpose throw away tools easy to create. This is GREAT. I had to migrate an old Windows 2012 file server share to SharePoint. Microsoft's tools don't work on this old OS. Their SharePoint migration tool running on other machines on the local network constantly failed for nebulous reasons. I finally got fed up and spent a few hours with Gemini Pro and Claude and created a sync tool using C# that does the migration and keeps the network share in sync with SharePoint until we do the final cutover. I don't expect to ever use this tool again, and that's totally fine. I'll still put it on GitHub in case someone has a use for it, but I'm not sure why I should lament the fact that this tool exists and may never see another use or the fact that I won't maintain it.
Don't waste your life playing with shiny new toys, sure, but learning how to use AI by creating things is not a waste of time.
what if…
we just used ai to improve products and services
instead of all this wanking off showing how you go through 1 billion tokens a month (not really that impressive)
what would be way more cool is
i made something that reliably saves others 8 hours a month of busywork
>exploring AI as a lens in Marshall McLuhan-like thinking
I would be wary of using McLuhan-like media analysis of AI. His central argument is that media are tools that extend man's ability. A calculator or a spell checker extend our thinking and writing. AI does not extend those abilities so much as it completely replaces it.
The way in which it does resemble media is insofar as it captures the same urge that McLuhan wrote about to see ourselves extended into the world. McLuhan tied this to the myth of Narcissus. The difference is that where Narcissus falsely believed it wasn't him and fell in love with what he saw, we falsely believe the image we see is ourselves and fall in love with it.
> AI does not extend those abilities so much as it completely replaces it.
These two elements (extend/replace) are not mutually opposed according to McLuhan's tetrad.
A calculator does replace our ability to do math in a certain sense.
At the grocery store there's countless (no pun) opportunities to do math in the sense of comparing prices and calculating unit costs etc, but most people can't do that math easily in their head because the calculator has made that skill less important.
But people also don't pull out the calculator repeatedly to do this in the grocery store, so the math just doesn't get done.
Cal Newport is a grifter whose one and only output nowadays is posting anti-AI rhetoric.