I've been vibecoding a ton for the past 9 months, built a bunch of cool little apps for myself with AI, ran experiments, built an entire SDLC on skills, did the agent orchestration harness thing, etc. In the past few weeks I've hit a wall where I'm just tired of it. Each model becomes more independent but also harder to direct in detail. They produce massive, tedious, sloppy text outputs with very little input. They're bad at socializing knowledge and communicating design forks.
The places where I've seen unequivocal wins with AI are repetitive tech debt tasks that apply the same transformation across a large amount of code or refactor under a pre-existing test suite with good coverage. It's great for initial research, brainstorming, and can be good (despite the sycophancy) as a rubber duck conversation partner. I use AI constantly, for work and in my personal time, but we've hit a ceiling where I no longer find it helpful for the models to absorb more of the intellectual labor. They get things wrong more aggressively, and more elaborately. They're inadequately curious. I cannot keep up with the endless bad technical writing, and it makes it harder to spot factual errors and bad reasoning.
Here's what I want: I want AI as an assistant that helps me make decisions, and ensures that I'm in the driver's seat. AI as an over-confident prodigy on speed is what we're getting lately, and it's losing me.
> as an assistant that helps me make decisions, and ensures that I'm in the driver's seat
This... quite describes what the current AI feels like to me. So I really don't see a problem. If AI does too much for you, you should, uh, just prompt it to do less.
Everyone talks about how their AI escapes sandbox and virtualization or delete their entire repos or something. Perhaps they're using a skynet I'm not aware of. From my experience when I added "discuss; don't edit any file yet" at the end of the prompt I've never seen the LLM go against my will and edit the files anyway.
Creating products with AI is like creating a tightly packed ball of wool. Then, to try to understand what happened you have to somehow pick a strand and pull it out to have a look. Very difficult with tightly packed wool.
I've run some experiments in the past month that get me closer to being able to work with this opaque bundle.
It's the main thing I think about every day, how to solve this problem.
> I want AI as an assistant that helps me make decisions, and ensures that I'm in the driver's seat.
Right now we’re still in the benchmaxxing phase, I feel like most AI models of combination of models and harness are not really a “product” yet in the sense of having a particular and opinionated UX engineered to be helpful at a specific task, there much more still in a race to be the smartest.
There’s a lot of room for diversification of the actual character of this interaction, for example in the direction you mention. I think once that starts we’ll get into the real “useful” phase, right now we’re really just showing off and exploring how much it can do, even if it’s incidentally really useful in some cases.
Personally I agree with the parent post, I hate it when AI tries to lead me “would you like me to ...” I’d much prefer something awaiting my instructions than wasting tokens tryitn or steer me to doing something inevitably dumb.
You develop a taste as to what details you can skim and what you need to dive on. We are all trained on this due to doing 10000 wax on wax off movements, called a "PR Review". I used to scrutinize. Now I think "yep that bit looks good and tests will catch errors plus I can manually test. This bit over here looks scary will spend time more deeply understanding".
With AI you don't need to understand every line in depth but it does need good judgement to decide which.
Yeah, I think the peak expectation for AI is that the user gets to be the manager of a very smart team. But we've all seen that. If the manager is clueless about the technology, the results are disappointing.
My AI dream (that I’m living happily) is getting to focus on the details that I find interesting and ignoring all the boilerplate details that modern software requires.
You said what I was thinking. AI can do the things I always found mind numbing and tedious, but because I did it already early in my career, I know when to intervene because the agent is not doing it correctly. I think that is the OP's point.
I think about LLMs in software today as attempting to provide a similar role as compilers in terms of translating high-level thoughts into low-level details.
The only problem is that I actually trust compilers.
I hope we can save “jobs aren’t supposed to be fun” for stuff like undertaker and grief counselor.
My feeling is making things should be mentally stimulating, and as many tedious parts that can be automated should be, so we can spend precious human thought on thinking instead of recitation.
I don’t know what you mean by the last part. Do you mean programmers depend on busy work to eat or the users depend on the human touch on the boring details?
Some of us just wouldn't work if it wasn't fun. I'd let myself starve in the streets. (I have some experience with starvation from an eating disorder phase in my twenties, it's only unbearable if you have to work).
Every word of this seems objectively false. AI is more than capable of handling the details. I have generated countless tools for myself without needing to know or care about the details.
The pro argument is always quantity based, yet we never get to see the actual gain in quality. It's always "I'm doing SO MUCH more". Never "I'm doing better, with less effort."
It's also what we see in the wild. There are so many things going on, so many news about "AI". Are things getting better in any meaningful way? Why are we getting so much from "AI" yet things keep getting worse? The only real, objective and verifiable gain has been the stock price of a handful of companies, most of them heavily into infrastructure and manufacturing hardware.
Seems like a pretty clear pattern. There's so much output, but everything is worse. And the only argument we hear for why this is actually a good thing is just how much more output there is. Hmmm
I mostly work on UI (for an internal, science-focused app) and it used be that features would take a sprint or two to grind out, and would always be an MVP to meet just the most essential user requirements, because that's all we could afford, time-wise, to do. And if I figured out while implementing there could be a nicer UI approach to something, it was often too late to change the approach.
Now, I can get to that MVP in under a day. And I can experiment with ten radically different approaches, or change my mind about something significant at any stage in the implementation. I can afford to add refinements and user-pleasing extras that I could never have done before. I'm not delivering features at 10x the speed. There's a limit to the amount of code I can (and should) be submitting for review. But the features I'm building are _better_, and the effort is less. I'm not wasting mental bandwidth typing out a dozen React components full of the same-old state management, form processing code, etc. I'm not trawling through charting library docs, looking for the precise combination of settings to make my chart's x-axis ticks labels rotate 45 degrees. My mental energy is focussed on figuring out what solution is actually _right_, not the minutiae of how to persuade the computer to do it.
> The pro argument is always quantity based, yet we never get to see the actual gain in quality. It's always "I'm doing SO MUCH more". Never "I'm doing better, with less effort."
For me, it's "I can do this now; yesterday I could not."
I am not a developer - I am a hack with some experience in a smattering of languages who hits very real obstacles very quickly. AI has allowed me to overcome those obstacles and build things that literally I could not do before.
Yesterday I could not differentiate quake 3 matches within a single demo file.
Today I can split them into separate files and have the server stream them to me so I can watch them on demand.
Tools? Yes. Reliable foundations? Eh, mostly not. Maybe I'm holding it wrong, but I'm not impressed. Not in the least because in order to get a good system you must have a good image of the system in your head. And if AI builds the system, how will you ever get that system in your head?
It doesn't handle anything, he reproduce statistical patterns. If it seems it cares about the details it's because he's been trained to regurgitate stuff. Another thing is being expert in something and being able to prompt the clanker, and those who are, usually underestimate how much their knowledge is steering the machine in the right direction
Everyone feels like they are now the ceo of their own empire of ai flunkies who do the actual work guided by their above-all-that vision, and they luuuuuuv that.
But useless ceos generate bad output just like useless direct coders.
It's like everyone is adopting the once only-for-the-rich mindset where artisan is actually a derogatory term. Where Micheal Angelo is the same as the cow stall poop shoveler, because they actually do something directly themselves.
it depends on what you mean by power and details and handing off and whom you are handing them to and what you are getting handed in return and your counterparty's relationship to these things and the extent to which you measure things in the same way as one another.
We're living in the glorious future where software engineers don't have to worry about nitty-gritty stuff like actually making software and can focus on the really important work: administrative and managerial tasks!
This same argument could be applied to anything. Layers of abstraction exist for a reason, because at a certain point, we can only deal with so many things at once. We have to be able to delegate "the details" to others -- be that a person, a company, or an AI model.
Does "not getting into the details" mean you have to understand how the GCC compiler works when you write C code? Do you need to be an expert in machine code, or how SSE and pipelined instruction caches work to write your little bit of code? Do you need to understand how Ethernet frames work to write an API route for a web server?
Knowing how these things work can be helpful in a broader sense, and perhaps when encountering weird edge cases or dealing with exotic implementation but are generally not required to get the job done. The details, simply put, don't matter because someone else has already thought through the problem and solved it in a way that is good enough for the vast majority of use cases.
The same goes with AI. It's helpful to know how things work, but as the models continue to get better and better, it doesn't matter. As long as they are trained properly by someone who does know the details, that's a far better place to be than training a million different people on it who will each have their own biases, levels of understanding and misconceptions.
You're confusing translation and interpretation. A compiler is merely translating your code into a lower-level form. (There is a degree of heuristic-based interpretation that goes into generating good code, but the code's observable behavior is not allowed to change.)
But an LLM has to interpret your words and make judgment calls about the logical details, which is a very different thing. An LLM sufficiently intelligent to make acceptable judgment calls in all situations (acceptable enough to ship to production without human review of those details) would likely be indistinguishable from AGI.
Your vouching for centralization of knowledge, and even understanding, which I think is a very bad idea. It’s not like the rest of us are becoming specialized at something else - knowledge, in the broad sense, is the foundation of all skills. Without it we have nothing to contribute
I didn't interpret it as an argument about layers of abstraction. You need to be in the detail at the level you're writing (and it's certainly advantageous to know some of the detail of the abstractions below) because you're building abstractions on top of that, because the details subtly change at the base level you're building at, impacting the level that you're building to.
I've been vibecoding a ton for the past 9 months, built a bunch of cool little apps for myself with AI, ran experiments, built an entire SDLC on skills, did the agent orchestration harness thing, etc. In the past few weeks I've hit a wall where I'm just tired of it. Each model becomes more independent but also harder to direct in detail. They produce massive, tedious, sloppy text outputs with very little input. They're bad at socializing knowledge and communicating design forks.
The places where I've seen unequivocal wins with AI are repetitive tech debt tasks that apply the same transformation across a large amount of code or refactor under a pre-existing test suite with good coverage. It's great for initial research, brainstorming, and can be good (despite the sycophancy) as a rubber duck conversation partner. I use AI constantly, for work and in my personal time, but we've hit a ceiling where I no longer find it helpful for the models to absorb more of the intellectual labor. They get things wrong more aggressively, and more elaborately. They're inadequately curious. I cannot keep up with the endless bad technical writing, and it makes it harder to spot factual errors and bad reasoning.
Here's what I want: I want AI as an assistant that helps me make decisions, and ensures that I'm in the driver's seat. AI as an over-confident prodigy on speed is what we're getting lately, and it's losing me.
> as an assistant that helps me make decisions, and ensures that I'm in the driver's seat
This... quite describes what the current AI feels like to me. So I really don't see a problem. If AI does too much for you, you should, uh, just prompt it to do less.
Everyone talks about how their AI escapes sandbox and virtualization or delete their entire repos or something. Perhaps they're using a skynet I'm not aware of. From my experience when I added "discuss; don't edit any file yet" at the end of the prompt I've never seen the LLM go against my will and edit the files anyway.
Agree.
Creating products with AI is like creating a tightly packed ball of wool. Then, to try to understand what happened you have to somehow pick a strand and pull it out to have a look. Very difficult with tightly packed wool.
I've run some experiments in the past month that get me closer to being able to work with this opaque bundle.
It's the main thing I think about every day, how to solve this problem.
> I want AI as an assistant that helps me make decisions, and ensures that I'm in the driver's seat.
Right now we’re still in the benchmaxxing phase, I feel like most AI models of combination of models and harness are not really a “product” yet in the sense of having a particular and opinionated UX engineered to be helpful at a specific task, there much more still in a race to be the smartest.
There’s a lot of room for diversification of the actual character of this interaction, for example in the direction you mention. I think once that starts we’ll get into the real “useful” phase, right now we’re really just showing off and exploring how much it can do, even if it’s incidentally really useful in some cases.
Personally I agree with the parent post, I hate it when AI tries to lead me “would you like me to ...” I’d much prefer something awaiting my instructions than wasting tokens tryitn or steer me to doing something inevitably dumb.
> Here's what I want: I want AI as an assistant that helps me make decisions, and ensures that I'm in the driver's seat.
You aren't going to get that. Why would they even offer you that? You're going to be the meatbag peripheral to an AI.
So don't look to "them" to offer it, look elsewhere.
Plus it would reduce token counts. Careful use is almost directly opposed to their financial goals.
You develop a taste as to what details you can skim and what you need to dive on. We are all trained on this due to doing 10000 wax on wax off movements, called a "PR Review". I used to scrutinize. Now I think "yep that bit looks good and tests will catch errors plus I can manually test. This bit over here looks scary will spend time more deeply understanding".
With AI you don't need to understand every line in depth but it does need good judgement to decide which.
Yeah, I think the peak expectation for AI is that the user gets to be the manager of a very smart team. But we've all seen that. If the manager is clueless about the technology, the results are disappointing.
All details are not created equal.
Some details are boring.
My AI dream (that I’m living happily) is getting to focus on the details that I find interesting and ignoring all the boilerplate details that modern software requires.
You said what I was thinking. AI can do the things I always found mind numbing and tedious, but because I did it already early in my career, I know when to intervene because the agent is not doing it correctly. I think that is the OP's point.
I think about LLMs in software today as attempting to provide a similar role as compilers in terms of translating high-level thoughts into low-level details.
The only problem is that I actually trust compilers.
How quickly would you drop a non-deterministic compiler for the source code of a project?
So what? Jobs aren't supposed to be fun. Lots of people depend on those boring details.
I hope we can save “jobs aren’t supposed to be fun” for stuff like undertaker and grief counselor.
My feeling is making things should be mentally stimulating, and as many tedious parts that can be automated should be, so we can spend precious human thought on thinking instead of recitation.
I don’t know what you mean by the last part. Do you mean programmers depend on busy work to eat or the users depend on the human touch on the boring details?
Some of us just wouldn't work if it wasn't fun. I'd let myself starve in the streets. (I have some experience with starvation from an eating disorder phase in my twenties, it's only unbearable if you have to work).
It is all about abstraction, and basic English is a quite bad abstraction
Every word of this seems objectively false. AI is more than capable of handling the details. I have generated countless tools for myself without needing to know or care about the details.
The pro argument is always quantity based, yet we never get to see the actual gain in quality. It's always "I'm doing SO MUCH more". Never "I'm doing better, with less effort."
It's also what we see in the wild. There are so many things going on, so many news about "AI". Are things getting better in any meaningful way? Why are we getting so much from "AI" yet things keep getting worse? The only real, objective and verifiable gain has been the stock price of a handful of companies, most of them heavily into infrastructure and manufacturing hardware.
Seems like a pretty clear pattern. There's so much output, but everything is worse. And the only argument we hear for why this is actually a good thing is just how much more output there is. Hmmm
I'm doing better, with less effort.
I mostly work on UI (for an internal, science-focused app) and it used be that features would take a sprint or two to grind out, and would always be an MVP to meet just the most essential user requirements, because that's all we could afford, time-wise, to do. And if I figured out while implementing there could be a nicer UI approach to something, it was often too late to change the approach.
Now, I can get to that MVP in under a day. And I can experiment with ten radically different approaches, or change my mind about something significant at any stage in the implementation. I can afford to add refinements and user-pleasing extras that I could never have done before. I'm not delivering features at 10x the speed. There's a limit to the amount of code I can (and should) be submitting for review. But the features I'm building are _better_, and the effort is less. I'm not wasting mental bandwidth typing out a dozen React components full of the same-old state management, form processing code, etc. I'm not trawling through charting library docs, looking for the precise combination of settings to make my chart's x-axis ticks labels rotate 45 degrees. My mental energy is focussed on figuring out what solution is actually _right_, not the minutiae of how to persuade the computer to do it.
> The pro argument is always quantity based, yet we never get to see the actual gain in quality. It's always "I'm doing SO MUCH more". Never "I'm doing better, with less effort."
For me, it's "I can do this now; yesterday I could not."
I am not a developer - I am a hack with some experience in a smattering of languages who hits very real obstacles very quickly. AI has allowed me to overcome those obstacles and build things that literally I could not do before.
Yesterday I could not differentiate quake 3 matches within a single demo file.
Today I can split them into separate files and have the server stream them to me so I can watch them on demand.
I think the author means ecomonically valuable not valuable to you.
Tools? Yes. Reliable foundations? Eh, mostly not. Maybe I'm holding it wrong, but I'm not impressed. Not in the least because in order to get a good system you must have a good image of the system in your head. And if AI builds the system, how will you ever get that system in your head?
How would you do it if you hired an employee in your company to do it? You'd talk to the employee and read the docs they write. Same here.
It doesn't handle anything, he reproduce statistical patterns. If it seems it cares about the details it's because he's been trained to regurgitate stuff. Another thing is being expert in something and being able to prompt the clanker, and those who are, usually underestimate how much their knowledge is steering the machine in the right direction
It’s always tools tho, and AI tools especially. Why not anything else?
Everyone feels like they are now the ceo of their own empire of ai flunkies who do the actual work guided by their above-all-that vision, and they luuuuuuv that.
But useless ceos generate bad output just like useless direct coders.
It's like everyone is adopting the once only-for-the-rich mindset where artisan is actually a derogatory term. Where Micheal Angelo is the same as the cow stall poop shoveler, because they actually do something directly themselves.
> to become good at the thing in the first place requires a complete reversal of the mindset that would lead one to having wanted to hand it off
Is this true? I can be good at something and be happy to not have to do it anymore I feel
Though not the other way around. You don't become good at something by handing it off. In that case you either are already good, or never become good.
CEOs, serial entrepreneurs, managers etc all seem to find it pretty empowering.
I treat my Ais like employees, pizza party and all.
it depends on what you mean by power and details and handing off and whom you are handing them to and what you are getting handed in return and your counterparty's relationship to these things and the extent to which you measure things in the same way as one another.
We're living in the glorious future where software engineers don't have to worry about nitty-gritty stuff like actually making software and can focus on the really important work: administrative and managerial tasks!
AI can do administrative tasks.
This same argument could be applied to anything. Layers of abstraction exist for a reason, because at a certain point, we can only deal with so many things at once. We have to be able to delegate "the details" to others -- be that a person, a company, or an AI model.
Does "not getting into the details" mean you have to understand how the GCC compiler works when you write C code? Do you need to be an expert in machine code, or how SSE and pipelined instruction caches work to write your little bit of code? Do you need to understand how Ethernet frames work to write an API route for a web server?
Knowing how these things work can be helpful in a broader sense, and perhaps when encountering weird edge cases or dealing with exotic implementation but are generally not required to get the job done. The details, simply put, don't matter because someone else has already thought through the problem and solved it in a way that is good enough for the vast majority of use cases.
The same goes with AI. It's helpful to know how things work, but as the models continue to get better and better, it doesn't matter. As long as they are trained properly by someone who does know the details, that's a far better place to be than training a million different people on it who will each have their own biases, levels of understanding and misconceptions.
You're confusing translation and interpretation. A compiler is merely translating your code into a lower-level form. (There is a degree of heuristic-based interpretation that goes into generating good code, but the code's observable behavior is not allowed to change.)
But an LLM has to interpret your words and make judgment calls about the logical details, which is a very different thing. An LLM sufficiently intelligent to make acceptable judgment calls in all situations (acceptable enough to ship to production without human review of those details) would likely be indistinguishable from AGI.
Your vouching for centralization of knowledge, and even understanding, which I think is a very bad idea. It’s not like the rest of us are becoming specialized at something else - knowledge, in the broad sense, is the foundation of all skills. Without it we have nothing to contribute
I didn't interpret it as an argument about layers of abstraction. You need to be in the detail at the level you're writing (and it's certainly advantageous to know some of the detail of the abstractions below) because you're building abstractions on top of that, because the details subtly change at the base level you're building at, impacting the level that you're building to.