Rendered at 01:41:23 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
tekacs 8 hours ago [-]
I've said for a long time that composability in software is a bit like playing Tetris: the lines have to clear.
I feel like that gives an even more literal tower-rising metaphor, and that's what it feels like people using agents naively (and software engineers of lower skill or earlier-career), end up violating.
Agents are getting better at folding things into themselves, especially if you direct them to... but unfortunately I've found that the architectural instincts, even of Fable and 5.6 Sol, are still wildly behind what I reflexively achieve, say.
For sure there is an ability to have agents go back over work and try to fold it into better and better abstractions until it's sort of annealed into something good. I've done something similar on codebases that I have, but the 'high reaches' of architecture with great _prediction of how the software will evolve in the future_ in _subtle_ ways – those are, for now, out of reach of agents.
There is a part of me that wonders if it's partly just how much they can hold in their head right now, though. Even with the greatest articulation and high density of feeding them, the current setups don't allow them to hold a high-quality, sparse, 'zoomable' model of the world in their head that well yet, which we can do pretty well.
But the fact that I'm talking about it in terms of that kind of subtlety is itself promising, I guess?
Animats 8 hours ago [-]
The upper bound on program complexity used to be the power of the human mind.
"Vibe coding" can break through that barrier. But not because the problem being solved needs that complexity. Because the process does not drive itself towards compact abstractions. It's the AI-powered version of the scaling problem Brooks described back in "The Mythical Man-Month". The combinatoric problems get worse with scale. Concretely, multiple similar implementations of roughly the same thing appear in different parts of the project. This is a known problem of vibe coding now.
We need some way to make AI-driven coding strive for parsimony.
ColdStream 1 hours ago [-]
The question is that if it can go beyond the upper limit of human mind complexity, would we be able to understand it or use it?
I cannot remember who it was but there was an author who was traveling with their dog. They noticed that their dog would always pee on various tree to mark them as their territory. On their travels they ended up need some giant Red wood trees and figured, "I want to give my dog the ultimate claim of territory."
So he took his dog up to the Red wood tree and it did nothing, instead the dog wandered over to a smaller sapling and peed on that instead. The problem was the Red wood was so big and alien to the dog, it didn't recognize it as a tree.
I do wonder how many things are like that in our universe, that even if we could see them, we just wouldn't be able to understand it because it just goes beyond what we are capable of understanding. We think we have a grasp of the universe and use models to codify it but that is no guarantee that we can truly 'get it'.
Could higher level AI code be like that, would we know when we see it?
jambalaya8 57 seconds ago [-]
Excellent post.
conartist6 8 hours ago [-]
Why would it? It has optimized what it was built to optimize: this is the token-selling industry. Take note that the people hawking the dream of a gold rush are not actually mining but selling shovels
ashdksnndck 7 hours ago [-]
Same issue happens in models trained by organizations who aren’t selling tokens. I believe it’s because being parsimonious is simply harder. Achieving the task at hand independently and declaring the job done is easier than building an abstraction and reconciling between every use case.
radarsat1 4 hours ago [-]
Agreed. I've noticed this happens without fail in any project driven primarily by "tickets" and "sprints". Finishing the immediate task ends up taking precedence over future concerns that cover a wider scope. On the other end of the spectrum it's easy to spend too much time at the beginning humming and hawing about the best, most future proof design, because you just know you're not going to be able to easily fix it later. As is often the case, the right approach is somewhere inbetween, which is partly why it's a hard process to automate. you're constantly making judgement calls.
tekacs 5 hours ago [-]
Labs are trying to make long-horizon work. Even if you're a coding agent, adding more and more surface area is distracting to that goal. There is reason that RL over long traces should, at least in principle, optimize for building in ways that help the result fit in the model's context window.
A meaningful risk of course is that the tools available to the model (ripgrep + fancier semantic approaches) allow it to do a good job of reasoning over things much larger than its context window, and so it doesn't pay the penalty sufficiently to fix it.
conartist6 5 hours ago [-]
Does that not sound a little silly to you when you say it? Should I invest in becoming a memory athlete as a way of becoming a better software engineer? ...or should I learn how to build and use tools?
gnatolf 4 hours ago [-]
While I don't disagree, memory certainly was more of a restrictions on us humans than it is on llms. Therefore, the answer may not be as obvious as it seems. We build abstractions to reduce (memory) footprint of features, right?
chongli 3 hours ago [-]
Humans built codebases many millions of lines long, well before LLMs existed. Human memory has not been a restriction on us in a long time.
Look at all the libraries full of books we've built. It's useful for more than mere training sets.
conrs 3 hours ago [-]
I think the trick here is plural; I guarantee no single human knows all 1 million lines. Note this is different than knowing how to orient yourself in a million line codebase quickly.
The limit here I think the ancestor comments are getting at is cognitive load, which is real and measured. We only have so much memory to devote to a "stack" when executing, and it's usually quite constrained.
chongli 3 hours ago [-]
Note this is different than knowing how to orient yourself in a million line codebase quickly.
Hence my library mention. Humans have been doing this for millennia: orienting ourselves within a library (the physical kind, full of books) and calling upon its information resources as needed to accomplish tasks (research). Ultimately, it's all just one big cache hierarchy. Your short term memory, your long term memory, the book in your hands, the desk at the library, the nearby shelves, the card catalogue, the stacks, the inter-library loan system.
To manage it all, we humans have developed our abilities for abstraction. When we build clean, tight abstractions we reduce our cognitive load. Perhaps the best abstraction we've built so far is the TCP/IP and web stack. We don't need to care at all about the hardware details of a server in order to talk to it. It's such a powerful and airtight abstraction that we take it for granted.
I'd like to hear from more people who have spent a lot of time building with LLMs, because so far what people are saying is that these models do not have the ability to reason about and build the kind of marvellous abstractions us humans have built.
conrs 3 hours ago [-]
Fair, my mistake.
I've built a lot with LLM's, my experience sort of but not really tracks that. I've had to course correct a few bad abstractions but the larger the code base becomes the better it seems to be at reusing things. Maybe this is because of types, or spec-first development (with OpenAPI), or black box integration testing - but also maybe not. But generally I have to think about the abstractions and let the LLM fill in the details with rare exception.
conartist6 2 hours ago [-]
I built a web-OS, a graphical IDE, and a version control system to replace git, all in about 40,000 lines of highly abstracted Javascript. If you're thinking about how important it is to be able to maintain million-line codebases, I suspect you might have substituted a metric for the actual end goal.
nomel 5 hours ago [-]
All the open weight models, that are given away for free, across orgs and even nations, are using the same methods to achieve high performance.
What's more profitable, optimizing for inference time or optimizing to increase inference time by increasing token count?
petra 1 hours ago [-]
That's an interesting question.
What happens if we take the most abstract libraries in any given field - and:
1. Bound to the llm to only use those as building blocks. Does it affect his reasoning ? Will it think more abstractly ?
2. Train the llm on those, so maybe it will get a feel for abstraction ?
fooqux 5 hours ago [-]
> The upper bound on program complexity used to be the power of the human mind.
Maybe for simple one-person projects. We've long since developed methods and models to allow us to make things bigger than ourselves. Linux, SAP, etc. These software projects are not held in the mind of a single developer. But we use structure, rules, and other tools so that the pieces still fit together.
sroerick 4 hours ago [-]
The Allegorical Agent-Aeon?
TacticalCoder 5 hours ago [-]
> "Vibe coding" can break through that barrier. But not because the problem being solved needs that complexity. Because the process does not drive itself towards compact abstractions.
It's the infinite AI monkeys at a computer keyboard phenomenon.
Or the car on the highway that bumps left and right on the guardrails until, eventually, it arrives at its destination and nearly everybody is amazed at that great success.
The AI kool-aid drinkers are going to answer: "but that's how human code too".
And I'm really not sure about that.
bluefirebrand 5 hours ago [-]
It's perhaps how some humans code but frankly if you have those people employed to build software for you, you have big problems
pishpash 5 hours ago [-]
That's some idealistic nostalgia. Software is generally poorly built today, and it's evidently not big enough a problem to fix.
munk-a 4 hours ago [-]
Large companies that can keep themselves alive with regulatory capture - absolutely. For smaller companies that need to compete the software quality and ongoing cost of maintaining that software is a real consideration.
That isn't to say software is perfectly built, but it's usually pragmatically built to balance costs of development and correctness - well chosen abstractions let us push up both qualities at once.
tibco 4 hours ago [-]
I think every type of company has ways to workaround software quality. B2B can tend to define the bug as the feature or too low priority.. B2C can often decide backwards compatibility simply doesn't matter and just replace things with other things with only some actual feature overlap, etc.
arkmm 5 hours ago [-]
I think part of the problem is the context windows for humans are actually much smaller than what an LLM can keep track of today. The small context window of humans is a feature that forces modularity and abstraction in software engineering so that you can decompose what you're working on into something that can fit into your head. But since LLMs can fit so much more in their head, so to speak, they don't have this same incentive, and you get the unorganized mess of spaghetti code that current agents often produce.
_superposition_ 37 minutes ago [-]
I can't prove it but I have strong beliefs that the logic and intuition required for abstracting for future changes is not possible in a stream of predicted tokens. Mental models aren't built from text. Something's missing that can't be measured.
segmondy 4 hours ago [-]
Just a matter of time. Go download gpt2 or llama2 and be shocked at how bad they are compared to today. They were entirely "useless" yet we marveled at them. Go examine GPT3.5/gpt4 out which was all the rage and then marvel at how a qwen27b or gemma31b model mops the floor today. My point is that the models will eventually learn to have a great model of software system in their head, just a matter of time and proper RL.
VMG 5 hours ago [-]
Isn't this just an effect of what the LLMs are RL'ed for? Solving short-horizon tasks.
I assume one can't benchmaxx multi-year long efforts, clean architecture, taste etc as easily as these "make tests pass" tasks
throwaway27448 6 hours ago [-]
> the lines have to clear.
Sorry, the lines have to clear what? Surely there must be some kind of constraint on "lines" that they have to overcome.
edoceo 5 hours ago [-]
The lines (rows) in Tetris have to become complete and then disappear to make room for the new falling pieces.
In code the thing has to become stable, can't just keep packing more and more noise onto it.
tekacs 5 hours ago [-]
Absolutely this: and it needs to ideally become the kind of set of abstractions that mean that every new thing added uses less net-new surface area than it would without them.
5 hours ago [-]
tekacs 5 hours ago [-]
I mean that at the bottom of the Tetris board, the lines need to vanish so that the Tetris board keeps moving downward and doesn't grow unbounded.
stavros 7 hours ago [-]
Agreed, and ever since LLMs started being able to write competent code, I've noticed a massive difference in quality on codebases where I knew the technology, and ones I didn't. This is because I can much more efficiently steer the LLM on e.g. backend code, which is my expertise, vs yoloing everything on mobile, where I have no idea.
The codebases using technologies I have no idea about tend to quickly become unmaintainable and buggy, because the LLM still doesn't make good architectural choices, but the codebases that use technologies I'm familiar with basically never devolve into unmaintainability.
The difference between the two is massive, and that's why I think that a competent engineer steering an LLM in their area of expertise gets two orders of magnitude more productive, whereas someone steering an LLM in an area they know nothing about are basically producing tech debt at the speed of thought.
maest 6 hours ago [-]
> two orders of magnitude more productive
Shipping 100x more features per day?
Cthulhu_ 5 hours ago [-]
100x faster towards becoming an email client
stavros 5 hours ago [-]
Yes.
Avicebron 5 hours ago [-]
Do you have specifics? It would be interesting to see what kind of improvements are possible.
stavros 5 hours ago [-]
I just see in my usage that I can release tens of features a day, whereas I'd be able to release one or two a day usually. I don't know if it's 100x, but it's definitely more than 10x.
The biggest thing to get right is to let the LLMs do what they're great at (code implementation from very detailed specs, and code review), and you do what humans are great (architecture and making sure the high level of the implementation is sane). That way, you get the best of both worlds, and a lot of speed at high quality.
wasabi991011 3 hours ago [-]
Thanks for the detailed explanation of your process, I think I'll try it out.
Cthulhu_ 5 hours ago [-]
I have a theory (armchair take here lmao) that AIs are trained on public code, but the biggest codebases are not public.
Although I suspect models from Google, Facebook and Microsoft can be trained on their massive internal codebases. Whether they are is another question.
mbreese 4 hours ago [-]
I’m not sure you’d see that big of a difference in quality. There is quite a bit of cruft that can accumulate when you know the code will never be public.
But, you would probably see a difference of scale and architecture. Larger projects that need better organization are probably more likely to be in private codebases (Linux excluded). So you might be right about the lack of private code in LLM being an issue.
aleph_minus_one 5 minutes ago [-]
I have another thought on this:
At least in the past (before LLM-based code contributions got socially acceptable in some circles), in open-source project you often got very direct comments on code that was of bad quality. Yes, this was always a little bit abrasive, but it did a lot for the code quality.
For internal applications used at companies, such an abrasive behaviour is typically not accepted ("not a team player" (as if this is something bad), offended snowflakes, "not socially adept" etc.). Thus the code quality suffers quite a lot for internal applications.
Have you ever thought why salespeople have such an easy time selling some LLM for coding to big companies? Because the code quality of many internal applications is so bad, which makes even shitty LLM slop code often better than what is there.
4b11b4 4 hours ago [-]
Hah Tetris that really matches my "everyone has to draw their own line constantly".
Gud 7 hours ago [-]
Do you believe "micro services" can make a comeback?
local daemons with an exposed API, each daemon vibe coded?
jdlshore 6 hours ago [-]
Microservices don’t reduce complexity, they just move it to the interactions between services. You have the same fundamental design problem.
In other words, if you can’t design a modular monolith, you can’t design a set of microservices.
aleph_minus_one 14 minutes ago [-]
> if you can’t design a modular monolith, you can’t design a set of microservices.
I somewhat disagree:
Yes, it is possible to design a modular monolith, but thinking about the system in terms of a "minimum viable service" (but keep an eye on "viable", otherwise you can easily get into the "interaction problem") makes it much easier.
This is very similar to how you can write programs with no implicit state in a imperative programming language, but doing this in a pure functional programming language such as Haskell is much easier.
swiftcoder 7 hours ago [-]
Unless we are planning to deploy them all individually to an expensive serverless platform like Lambda, the coordination challenges and overprovisioning are going to more than outweigh whatever architectural benefit you reap (in human-centred development, micro services are solving an entirely different problem - Conway's Law)
Cthulhu_ 5 hours ago [-]
What problem would it solve? They're still part of a larger system ultimately. Sure, smaller codebases with more focused scope can be good for e.g. human individuals and LLMs, but there's multiple ways to achieve that that don't require a network boundary.
layer8 5 hours ago [-]
Microservices are about separate deployment. Regarding separating the development/maintenance of components, you can achieve that in a monolith by composing it out of corresponding modules/libraries with defined APIs. That’s good practice anyway.
CuriouslyC 3 hours ago [-]
Not microservices, but something more akin to FaaS into a mesh, with a backing domain logic library.
throwaway27448 6 hours ago [-]
Sure, why not? The same reasons they succeeded originally will work just as fine now.
stavros 7 hours ago [-]
Please no.
jambalaya8 7 hours ago [-]
Lemme guess, you were in favour of monolithic systemd too?
ssivark 7 hours ago [-]
The core thesis of this essay is reminiscent of the Lisp Curse [1] / Bipolar Lisp Programmer [2].
It's been a few years since I read these, but if I recall the argument there, it was that Lisp makes it so easy to build stuff and scratch exactly your own itch, that there's no real strong push for lisp programmers to come together and collaborate to build non-trivial and general purpose artifacts. And that is why the landscape of public lisp software is poorer as a result, compared to languages which demand much more effort to get anything substantial done.
Armin seems to be making a very similar point about AI coding.
Assembly programmers made the same argument. It seems that we revisit this same trope each time the practice of software engineering undergoes a paradigm shift.
Some have a harder time with the transition than others.
segmondy 4 hours ago [-]
Is it really the same argument? I think there's enough of people that will argue that there will be no need to come together to collaborate, just rally a bunch of agents and you can build whatever non-trivial stuff you can imagine.
baq 6 hours ago [-]
Au contraire if typescript and rust didn’t steal the whole show it’d be a great time to be a lisp LLM pilot: agents can explain pretty much everything without any confabulations nowadays, so the understanding problem essentially goes away, if you care, which is exactly the point of the article if you ask me.
ares623 5 hours ago [-]
But at that point why use Lisp (which LLMs have been so far still to struggle to get matching parens every now and then)
What made Lisp cool and powerful goes away when you do it through an LLM.
sroerick 4 hours ago [-]
I have a web framework running with a lisp interpreter built in, and I think it unlocks a LOT for the LLMs.
baq 5 hours ago [-]
lisp can do things that make the currently accepted as standard best practice software engineering processes redundant, but it needs a different set of processes which I’m not sure are written about anywhere since apparently nobody is running with e.g. hot patching of application code
sroerick 4 hours ago [-]
I'm doing fully interpreted Lisp with the AST stored in Postgres. I can access the REPL remotely and I honestly dont know if I'm ever going to build an API or CLI again
Aeolos 3 hours ago [-]
This sounds super interesting. Do you have any blog or pointers where I could read more about this?
zephen 6 hours ago [-]
> compared to languages which demand much more effort to get anything substantial done.
It is not clear at all to me that other languages "demand much more effort" for the same end result.
It is clear that many non-lisp programmers value syntax, and many lisp programmers don't. Even many people who programmed enough lisp to have their minds blown and expanded still prefer not to program in lisp. I'm still awaiting psychological studies on this, but the rift is so large, I think there may be some significantly different brain processing going on between the two groups.
To your point, yes, it is also clear that, to the extent that lisp can match the productivity of other languages, whether it exceeds them or not, one of the tools that is needed to achieve this productivity boost in lisp is heavy usage of homoiconicity, and this results in every serious lisp program being a collection of DSLs, each of which is only understood by one person or very few people.
kazinator 29 minutes ago [-]
> Even many people who programmed enough lisp to have their minds blown and expanded still prefer not to program in lisp.
I can guarantee you that most of these simply didn't go far enough; they had their "minds blown" with too little. They latched onto a highlight or two, found a way of sort doing that in C++ or Javascript and moved on.
There are people who have had their "minds blown" by time/mass dilation of special relativity, or phenomena in quantum physics, who don't actually know anything about physics; they couldn't work out how far a canon ball will land fired at a 45 degree angle with a certain velocity, though they can crack jokes about Schrödinger's cat.
For those of us sticking with the Lisp program, in one way or another, it was more of a sequence of small revelations in a progression of topics. Oh, that's a good way of representing that; that is pretty nicely thought out; that fits well with that; they had that how many decades ago? Sheesh; ...
Then once you start grappling with the Lisp issues, and kick the rock farther down the road, even in some small way, you are invested.
nico 3 hours ago [-]
> Even many people who programmed enough lisp to have their minds blown and expanded still prefer not to program in lisp
For me, the answer to this is economics. Even if you love lisp, there are way more companies hiring for stacks that don’t include lisp
If you then optimize for employability, which I assume most developers do (not all, but a large percentage), you might end up with not that many people practicing lisp regularly
sixtyj 8 hours ago [-]
> There is the appealing idea that AI-assisted programming means better tools which lets us build more ambitious software. That is certainly true at the level of the individual and without doubt a developer with an agent will be dramatically more capable of changing a codebase. But large software projects have never been limited only by how quickly an individual can produce code. They are limited by how well people can coordinate their understanding of the system they are changing.
So true.
Since Nov 30, 2022 everything has become… more complex.
overgard 7 hours ago [-]
I feel like with software, things have gotten way too complicated (just layer's upon layers upon layers). But to deal with that complexity, now we're using something that just creates WAY more complexity. I've been coding for a while, and I remember the 90s and early 00s where people could make pretty powerful applications with like visual basic or php with essentially no formal training. Those technologies weren't great, but they were really simple and easy to pick up. In contrast, if you try to pick up web development or desktop app development today, it's absolutely overwhelming. Like, something like React is useful but the amount of things you need to know to use it properly is pretty high.
I think introducing AI to deal with this is overall a mistake though. We're just adding more complexity on top of the existing complexity. At best, it's a massive waste of hardware. At worst, we'll probably have agents introducing as many bugs as they fix as they also drown in complexity, and a lot of stuff built using these techniques are going to be fragile garbage while the overall skillset of humanity diminishes because people aren't learning the skills anymore.
Fundamentally, software does not need to be this complicated and it's a solvable problem, but it does require people that care about craftsmanship.
throwaway219450 5 hours ago [-]
I had a discussion with folks at work about what information is worth retaining in the face of AI doing everything for us. A lot of what we have in our heads to qualify as "domain experts" is pretty esoteric. How to invoke command line tools, gotchas because library A uses one convention over library B, AWS vs GCP; so much is specific to a tool rather than a method. There are also a lot of entrenched tools that are effectively unfixable due to the risk of breaking changes, so you have to shrug and accept + learn that's how it works.
Catch-22 is it's still important to know the fundamentals so you know what to ask for, but if you don't know the esoterica, the model is eventually going to make an assumption and screw things up. And the models don't have much taste either in prose, or in coding/comment style.
galaxyLogic 2 hours ago [-]
Yes I think the problem is not that AI "has all the answers". It is that we don't have "all the questions". If we don't understand the whole of what AI produces for us, we can't ask it to make parts of it different. We are not in control any more of what exactly is produced, and we don't want to give AI all the control, because then it would not be doing what we want it to do.
Therefore it is critical that whatever AI produces is understandable to us humans. That is why we must demand that AI tools and agents produce "well designed" well-structured software. That's the bottlenexk to progress I think. Even AI can't deal with exponential complexity explosion.
SoftTalker 6 hours ago [-]
And what's ironic is that a lot of those layers and complexity were added with the stated goal of making it easier for average developers to build applications.
sixtyj 6 hours ago [-]
Amen.
Drowning in complexity. Paralysis of choice.
I read a comment (joke) that if you want to follow all LLM development you should have to be unemployed.
lelanthran 7 hours ago [-]
> They are limited by how well people can coordinate their understanding of the system they are changing.
It's not really news, though. Programming as Theory Building (Peter Naur) was published in the 80s, I think?
Maybe the younger entrants to this field never came across it, but even if you never came across it, it was common knowledge amongst experienced devs that understanding of the system you are about to change is crucial.
sixtyj 6 hours ago [-]
The complexity of coordinating a project involving more than one entity is, of course, an issue across all industrial sectors—just look at the construction industry.
Thanks for mentioning Peter Naur’s Programming as Theory Building (1985).
I would add Fred Brooks and his The Mythical Man-Month.
jgalt212 6 hours ago [-]
> It's not really news, though. Programming as Theory Building (Peter Naur) was published in the 80s, I think?
The news is that Agentic Programming has made this always challenging task even more challenging.
sixtyj 3 hours ago [-]
Challenging. Adventurous. And tiring.
It’s global madness fired up by continuous stream of news from LLM providers. It’s like The Verge that is almost about FAANG only, but multiplied as it is “magic” for most people, as vibe coding is “so easy” and dopamine-producing activity that it is similar to runners that don’t want to stop as it stimulates them (and in this case ofc it is healthy :)
StackOptimist 1 hours ago [-]
[flagged]
calvinmorrison 8 hours ago [-]
I don't know. some stuff has gotten less. Major databases now ship effective HA tooling, microservices seem on their way out, structured databases seem to be back in instead of NoSQL.
HTML and pre-rendering are back in, HTMx, liveview
The degaussing of CSS and the hacks we did, hell i was trying to explain how we debugged web pages in IE6 to a younger staff member today.
Some things are more complex, some things got good enough to make them less complex.
endorphine 7 hours ago [-]
A database with built-in HA is a significantly more complex system than one without it.
hilariously 7 hours ago [-]
And one built without it and not coordinating with it is often much harder to reason about when you bolt it on later.
paulryanrogers 8 hours ago [-]
> Major databases now ship effective HA tooling
Which ones? PostgreSQL doesn't have HA in core.
calvinmorrison 8 hours ago [-]
MySQL 8, but upon review that was 2018. 5.7 had some but it's certainly improved overall since then as well
pixl97 8 hours ago [-]
>Since Nov 30, 2022 BC everything has become… more complex.
FTFY
Increasing complexity is the story of mankind. It's the story of civilization.
Someone from 20,000 BC would wander around the earth trying to find food, trying not to freeze, and trying not to get eaten. Someone from 5,000 BC would be trying to grow food, hoping it rains, and hoping disease didn't wipe out the village. The second one increases the complexity from all the systems required to manage people and keep the land growing. Today the vast majority of people on earth don't grow their own food at all, and instead are busy in some way managing the complexity of a large society.
Someone from 1970-80 would think our software from pre-llm days was vastly more complex. They'd just code directly to the hardware with no abstraction layer. Now almost no one does that. We abstracted the hardware away in most cases. With cryptography libraries for the vast majority of people it's complexity is abstracted away and mostly people are told "don't try to write your own crypto because you will fuck it up".
The question now becomes, how quickly will LLMs be able to coordinate their understanding of the system they are changing?
qsera 7 hours ago [-]
>LLMs be able to coordinate their understanding
I think the next time I see "LLMs" and "Understanding" in the same sentence, I am going to lose it....
pixl97 6 hours ago [-]
[flagged]
irishcoffee 6 hours ago [-]
I'll meet you in the middle: an LLM "understands" words in the same way a toddler understands the phrases they say. "My want cookie!" The toddler has zero comprehension of what any of those words mean, but they know that saying them in that order might result in something desirable.
An LLM has zero understanding of "my", "want", or "cookie" because an LLM has no id/ego, has never felt desire, and has never eaten a cookie.
pixl97 6 hours ago [-]
I believe you've made a category error in understanding, um, understanding. You've tied emotion into it. This to me are entirely different concepts where both happen to be wrapped up in meaty flesh that drives us humans. Now, these concepts are very important in sociology and human understanding of how we behave, but they also may have zero importance for the domain that encompases all understanding.
HN would commonly recommend reading the book Blindsight here.
Moreso, all you've done is recreate the Searle Chinese Room thought experiment which gets bounced around with no means of deciding if it reflects reality or not.
irishcoffee 4 hours ago [-]
No I didn’t. You just don’t like the analogy so you rejected it with a lot of empty words.
The_Blade 4 hours ago [-]
the shy one of my shelter cats that figured out how to open and close the cabinet below the one with wet food at 3am without meowing has infinite more understanding than any LLM of how the world works
CamperBob2 3 hours ago [-]
How'd your cat do at IMO last year?
CamperBob2 5 hours ago [-]
I'll meet you in the middle: an LLM "understands" words in the same way a toddler understands the phrases they say.
How'd your toddler do at IMO last year?
irishcoffee 3 hours ago [-]
Ask an LLM for the difference between a definite and indefinite article, and then pronouns. Maybe just a whole run down of basic grammar. And then read the comment again.
a2dam 6 hours ago [-]
You should be less upset over semantics that everyone else has usefully settled on. LLMs understand things fine.
sokka_h2otribe 4 hours ago [-]
For what it's worth, nomadism and agricultural systems are not necessarily more complex - that is, the complexity of a nomadic lifestyle can be quite high. Randomization, varied food sources, hybrid lifestyles, etc.
The hierarchy is certainly higher in agriculture societies it seems, but the complexity is up for some debate
> The shared language of a software project is not English or Python but it is the common understanding of what its concepts mean, where the boundaries are, which invariants matter, who owns what, and why the system has the shape it does. This language is rarely written down in one place. It lives partly in documentation and code, but also in code review, conversations, arguments, and the experience of having to explain a change to somebody else.
This is so true. I am a big fan of Christopher Alexander’s “Pattern Language” concept, which addresses this exact problem! In fact he recommends developing your own pattern languages for your own domains (which of course led to the famous GoF Design Patterns book).
I have been experimenting with a “Pattern Language” skill which instructs the AI to maintain 3 pattern languages for every project. One in the business domain, one in the product domain, and one in the technical domain. It is working really well. It is always super cool to see it reference the pattern languages during planning and curate them during implementation and review.
I credit using it with keeping my 100% ai-coded projects well organized, aligned across domains, and easy to work on.
ucyo 5 hours ago [-]
Would be interesting to see an example. I heard a couple times that code maturance is very similar on how a city grows naturally, but never seen an example.
whattheheckheck 7 hours ago [-]
Do you have any github examples publically available?
HiPhish 6 hours ago [-]
> But it’s not the biblical story. At Babel, the loss of common language stops construction whereas in AI-assisted engineering, construction can continue after shared understanding has already collapsed. The lack of an immediate failure is what makes it curious and a bit disorienting. The tower does not fall, and so we do not notice what was lost. It just keeps rising.
I don't know whether the author thinks this is a good or a bad thing, but in my eyes it's clearly a bad thing. Intelligence is knowing that a tomato is a fruit, wisdom is knowing not to put it in a fruit salad. AI is the the ultimate form of intelligence with zero wisdom. Actually, it's not even intelligence, it's an illusion of intelligence. If there is no human who can understand what the AI is doing it's time to stop and accept that we do not have the wisdom to contain what we are building.
The_Blade 4 hours ago [-]
> I don't know whether the author thinks this is a good or a bad thing, but in my eyes it's clearly a bad thing.
I like that the author lets the image do the work, rather than preach at me even if I were to agree. History never repeats itself but it always rhymes.
robocat 5 hours ago [-]
> Intelligence is knowing that a tomato is a fruit
Intelligence is learning to avoid using childish cliches, unless your intention was to mislead. Categorisation and understanding dependencies are hard enough problems already.
At the supermarket or taxation (Nix v. Hedden) tomatoes are a vegetable.
godwinson__4-8 4 hours ago [-]
Speaking of "categorisation and understanding dependencies" and "intelligence" and "fruit":
Interestingly, the Bible does not specify what kind of fruit was on the Tree of Knowledge. The association between the forbidden fruit and an apple was a derivative of later translations.
Imagine the consternation if the claim had become Adam and Eve ate of the forbidden tomato. Then this incessant controversy regarding the tomato's status would take on a biblical dimension... though then perhaps the Mormon claim that the Garden of Eden is to be found in the New World would actually have some credibility. Indeed, answers to some questions are not a matter of intelligence at all. Sometimes it seems to me intelligence is more about knowing what questions to ask in the first place.
robocat 4 hours ago [-]
> knowing what questions to ask in the first place
That's my 2nd hardest skill when trying to use AI. 3rd hardest is better filtering.
I keep having flashbacks to learning how to use Google effectively.
I'm an arthiest, but I'm in a historically christian society so bible stuff comes up and I often delve in.
Whenever I do look into the bible, it amazes me just how extremely bullshitty vague the bible actually is. I don't understand why people argue about it, or pay attention to past arguments (many famous examples). If I wanted to argue about vague shit I'd be into philosophy or psychology or I could choose any of the many other branches of argument. Humans are not very logical. At least capitalism has a clearer scorecard.
Last nerdsniped look was about Angels due to Scott Alexander:
(The Bible describes very clearly what angels look like. [snip] This is the highest-grade antimeme I feel comfortable using as an example; if you don’t see the fnords they can’t eat you.)
Edit: Which just meta-nerdsniped me finishing up on the beaut Hispanic word "Marianismo".
godwinson__4-8 3 hours ago [-]
I am also not religious, but yes the Bible is everywhere. So like you, I acquaint myself for similar reasons.
Though I take issue with one statement you made - "If I wanted to argue about vague shit I'd be into philosophy..." though I do sympathize.
Modern analytic philosophy in particular, if it is good, is all about trying to be exceedingly precise, often in domains where objects like the Bible (as you noted) are exceedingly vague. Philosophy has some catching up to do, but ultimately one should not disparage it so cavalierly. After all, the best philosophers were usually also the members of society most against the sort of uncritical religiosity that seems to appeal to so many, as it does now and even more severely through the ages, before the efforts of such thinkers to enlighten their fellow human were realized.
If there is a subject that seeks to answer the question of "how to ask questions" it is surely philosophy. The best mathematicians were often well acquainted. In the most famous times of "human flourishing" the philosophers and the mathematicians were one and the same.
But yes, there is a lot of bad philosophy. And there is a lot of religion pretending to be science. Sad, but it has always been true.
Mundus vult decipi, ergo decipiatur
geraneum 5 hours ago [-]
I have this funny feeling that someone’s probably gonna ask their favorite frontier LLM about the fruit salad thing to refute your point.
alwa 8 hours ago [-]
I come back to Babel and the Bruegel image too, although taking from it a little less optimism.
I feel these systems rising and sprawling with wee myopic agents developing out their little corners of this unknowably vast whole… a tower with 50 parapets on one side and some wacky cantilevered maiden tower on the other, and a very serviceable adobe roof over some patio for god-knows-why, and thatch over the landing next to it…
Some grotesque fatberg of designs that make sense at the level of individual design efforts, but that lack the fractal sort of levels of policy and judgment that unify the overall enterprise.
The overall language, as it were.
And language takes discipline to establish and maintain through any sufficiently large group of people—witness the company-speak or army-speak of pretty much any successful organization.
We feel like we’ve conquered the problem of talking the same language as our “Gastown Mayors” (who in turn are talking the same language as their “polecats” and so on all the way down the chain of golems)… but it’s only when it’s all built that the good Lord will humble us… that we’ll realize the understanding we thought we’d transmitted perfectly from our thrones wasn’t quite so shared as we’d imagined.
zemo 7 hours ago [-]
Anakin: "a developer with an agent will be dramatically more capable of changing a codebase"
I increasingly feel that the reason vibe coding specifically prohibits reading the generated code is that it's impossible to forget the horror that lurks in these python files.
The code will do what you asked for in a broad sense but wherever choices arise on how to accomplish the goal it'll have made those choices incoherently.
There'll be strange validations applied inconsistently to some user inputs but not others. Data will be repeatedly sorted or converted to lower case for no reason at 3 stages. The column labels are all hard coded strings even though they are the first row of the input CSV. It'll create global functions taking data classes half the time and classes the other half.
If you look, and if you think you might one day need to update and maintain the code manually, then it's very hard to resist fixing this sort of thing, but fixing this sort of thing cancels a lot of the time saving out.
jeffreyrogers 8 hours ago [-]
My comment is not directly responding to the essay, but it got me thinking about about how agentic programming is much more akin to management than it is to actual programming. Managers generally only have a high level idea of what ICs are working on and often don't have the time, bandwidth, and in some cases ability to understand everything the ICs they're supervising are doing. As more and more software gets written agentically the role of software engineer becomes less technical and more managerial.
segmondy 4 hours ago [-]
We can say that about programmers, most ICs don't understand what's going on in the layers beneath were they work. Most have no idea what's going on with libraries, frameworks, remote APIs, it's all abstractions. Most people can't tell you how system calls are implemented or function. They don't have the time, bandwidth to understand it all, they just operate at their own layer to get the job done.
snarf21 8 hours ago [-]
It feels to me like I'm stuck doing code reviews for a junior dev all day so I use it as little as possible and mostly to look for things I may have missed.
gwbas1c 7 hours ago [-]
It's great for "mechanical" changes.
For example, yesterday I came across some unit tests that didn't have error messages in their assertions. Normally, it takes me ~10 minutes to fix a handful of tests in this situation. In this case, I gave a 2-3 sentence prompt, went to the bathroom, and reviewed the result after I washed my hands. Saved me a bunch of time!
I encourage you to accept a feeling of "imposter syndrome" when using it, and keep trying new things with it. Don't feel like you have to be hands off, except when you're confident that you can be. (IE, if you think you need to spend 30+ minutes on mindless refactoring, see if you can explain it to an agent and then look at HN while it runs. You might get a good result, otherwise, it probably was time for a break anyway.)
BTW: It's important to try different models. The Claude 5.0 models are slow and give me bad results, so I'm sticking with 4.x for now.
BatFastard 6 hours ago [-]
Use the force Luke!
I finally learned to let go of the code. I dont even run my C++ editor anymore.
I run frequent code and architectural reviews. Its awesome.
prymitive 8 hours ago [-]
It used to be that you need a good reason to make huge refactorings, because it’s often so much work. Now agent can rewrite half of your code if your prompt is vague enough and you don’t actual try to review it all. And so the “soul” of a program can change dramatically every single day. It’s both great and very much not so.
Xirdus 8 hours ago [-]
The biggest obstacle to huge refactoring has always been minimizing the risk of bugs, not losing any features, and ensuring compatibility with the existing ecosystem. The reason it's become easier in the age of AI is because we stopped caring about these things.
sarchertech 7 hours ago [-]
Yep. That’s what people are forgetting. If you have an application that many people depend on to do real work, to make money, you won’t survive if you allow AI to constantly make huge changes.
Your test suite doesn’t cover all workflows. It doesn’t cover every combination of actions a user can take. So every big AI refactor while change some of those.
If this is happening frequently, your software will feel like a janky piece of unusable crap.
8 hours ago [-]
phoneafriend 3 hours ago [-]
Is communication at risk? Perhaps more demanding of one's attention, more insistent in its attempt at inclusion and contribution?
Large Language Models are the most powerful communication tools to have ever existed - probably behind the written and spoken word and marginally behind the internet, maybe more important than phones and texting.
With understanding now on tap in this new form of infrastructure, the tower (codebase) and our coordination within it are both very customizable - up, down, sideways - per the will of the team and maintainers.
The sample size of your architectural taste, intuition, or expertise at abstraction may never be large enough for a LLM to match it. But if you want it to, it too can keep getting better and closer at this as it seeks alignment with your acceptance.
Once your tower is built, what's next?
softwaredoug 1 hours ago [-]
One day we'll get to the plateau of productivity of this hype cycle and wonder WTF were we thinking.
Not that we won't have most code written by agents. I truly beleive that to be true. But that we got so far down the "never look at code" rabbit hole that our abstractions will become so divorced from the reality that is the actual.. freaking... code. We will truly have spaghetti code, bloated towers of babble.
We'll tut-tut these times and think "if you only spent half a day thinking about the big picture, looking at code, dare I say write some code you could have saved the project a whole mess of trouble".
In the meantime we're going to be generate 10K in spaghetti what should probably be 500 reasonably understandable LoC.
Spending a day reading some code can save you $100s of tokens, and weeks of headaches.
cadamsdotcom 4 hours ago [-]
Software construction needs to become like Mission Control - everyone can see everyone else's screen. Maybe that's a monitor bank for passive monitoring, and letting anyone hop into any tmux and steer any agent - and an agent pool works on tasks from the backlog, rather than engineers being assigned tasks and using agents to do them. It needs to be more like an orchestra, where things are being directed in real time - designating a conductor who steers things. Maybe it's like a ship's crew, with roles, a chain of command, and lots of overlap. So each day someone gets to be Captain.
It was all good when change was slow and needed deep contemplation. Things sped up this year. Chaos is harder to defeat now. It really does need a different paradigm than "people facing into corners each doing today's task".
_pdp_ 4 hours ago [-]
Vibe-coded software require a new type of discipline that has not been invented yet.
cheschire 5 hours ago [-]
Douglas Adams basically predicted Wikipedia on smartphones, and I suspect we are only a few years away from babelfish ear buds.
Tagbert 5 hours ago [-]
Apple AirPods and Google buds both can do automatic translation. It's still a little too slow and is a little clumsy to use but it's getting better fast.
bobro 3 hours ago [-]
Really surprised not to see Pope Leo's recent encyclical on AI/tech mentioned here.
I had the same thought, glad someone else noticed this.
Comparing AI adoption to the Tower of Babel is a central theme to Magnifica Humanitas, but written by someone with a much deeper understanding of the Biblical narrative.
trjordan 7 hours ago [-]
The agent will always fill in the gaps in your understanding. It's not a compiler. It's categorically different from any of the other ways we've built software.
I'm not sure reading code is coming back. The ritual of reading code must come back, because that's the only way to build products that don't collapse under their own incoherence, both technically and visibly.
"just ask Claude" is fine, but it's not the end state
conartist6 8 hours ago [-]
Does it really keep rising? Many of my fondest memories of technology come from times past...
CobrastanJorji 8 hours ago [-]
I interpret "keeps rising" negatively. Changes keep getting made, certainly. The AIs will perhaps never fail to fulfill your feature request. But there's no overall plan. It's just undirected, cancerous growth. It's Homer Simpson telling a team of automotive engineers to add feature after feature.
a2dam 7 hours ago [-]
This isn't really a good way to judge things. In the future, the fondest memories someone else has about technology will be about the present. The past is not better, you're just nostalgic for it.
jambalaya8 6 hours ago [-]
Every time I think the past was better, I think about how terrible ksh scripting was in 1995. And look at how great peoples' bash scripting is now compared to when we though bash had reached its apex in like 2009.
archonis 4 hours ago [-]
Conversely, Rexx scripting on the desktop was glorious in 1995.
jambalaya8 11 minutes ago [-]
lol I was actually kidding. The past was way better, almost always (it isn't just nostalgia). I think my enjoyment of tech actually used to sort follow a wavelike pattern. Then, it sort of became a straight line a few years before COVID, and I worry that AI is cementing people into an always-forward mindset that removes the enjoyment from just about everything.
I mean, yeah, hobby coding is not going away, but the feeling of exploration for me is totally gone. I don't dislike tech, I just don't see anyone being the way they used to be. People in tech are different and too many people are in tech.
Probably only part of the joy will ever be there, now. Which is weird because I did my own thing with computers til college, and don't consider myself a super sociable person. I just used to know that it was there, and now it isn't.
lstodd 3 hours ago [-]
ARexx on Amiga was way before 1995. And it was glorious.
jambalaya8 5 minutes ago [-]
I used to want an Amiga. they are probably still for sale on ebay. Like in 2010. :(
GlickWick 8 hours ago [-]
The tower is not about fondness, its about growth
conartist6 8 hours ago [-]
Is growth enough if technology makes our lives worse? Is a tower the pride of the civilization if a strong gust of wind could bring it down? It is before the gust, when all that matters is that the tower is tall rather than strong. After the gust, things are a bit more nuanced. Fingers are pointed.
The tricky part here is that you can't tell if a once-topmost part of the tower is sturdy until a great deal more tower is resting on it. Well, now a lot the economy is resting on little other than AI dreams. Your move, rational people.
archonis 4 hours ago [-]
Cancer is also growth.
BigTTYGothGF 6 hours ago [-]
> Many of my fondest memories of technology come from times past...
Is that because of the technology or because of who you were at the time?
sigbottle 5 hours ago [-]
If humans are still around by the end of the century: previously shunned philosophical movements such as pragmatism will actually become mainstream again.
The most interesting problems arise when you don't try and force one shared standard upon everybody yet still try and play nice.
Alternatively, power could concentrate and the winners get to decide what is valuable and not, thus cutting down the space of possible complexity by construction.
luciana1u 5 hours ago [-]
the tower keeps rising but the elevator pitch for every new floor is 'it solves the problems created by the floor below it'
thadk 6 hours ago [-]
Three or so years ago, Omar, the creator of DSpy pointed out on Twitter that ~LLMs get better most by better internal collaboration. Wish I could find it.
It seems to me that LLMs and particularly chatbots have already allowed for bigger scale collaboration within the LLM companies versus what was possible within the prior cohort of big platform companies.
Has the result just been taller towers, or actually a change of what is possible?
guessbest 7 hours ago [-]
I don't know why people hold on to all this extra software and features when with the tools its easier than ever to strip that out and refactor the end product in to a much more compact deliverable. Maybe once upon a time it was useful to keep legacy parts of the software solution around, but it can be recreated with fresh eyes if needed given the power of the new LLM models. My philosophy is if its not needed, it needs to be removed.
ddp26 5 hours ago [-]
> I can ask an agent to add OAuth, you can ask one to add caching, and somebody else can ask one to rebuild the database from first principles and make the UI pink. Each change can be reasonable in isolation.
But this is just bad vibecoding? This would be bad if humans did it too. With agents or humans, you need to coordinate.
zawaideh 5 hours ago [-]
This is exactly what Marx meant by labourers being alienated from their work because none of them understood anymore how the repetitive task they did factored into the product sold.
We are going through a transition from a guild based software production with primitive division of labour to a machinery based one where AI is the steam engine and the job of the engineer is to build the production line, be the mechanic fixing the line, and also the assembly line worker.
aaron_m04 6 hours ago [-]
This could've been a much better article without the strained Tower of Babel article.
aaron_m04 3 hours ago [-]
Err, last word is "analogy"
overfeed 6 hours ago [-]
AI replaces a single tower with millions of 5-over-1s[1]. The aggregate height, and speed of construction mind-boggling, but when each building is considered individually, not very impressive.
1. Perhaps with a handful of skyscrapers sprinkled in.
abraxas 5 hours ago [-]
so the post-AI landscape is software as a shanty town? That's not a bad analogy. We won't like the looks of it and it will barely function but like real life shanty towns it will function nonetheless.
archonis 4 hours ago [-]
Kowloon Walled City.
__0x01 7 hours ago [-]
Agents are very good at making us think the tower is rising, when in fact it is falling beneath our feet.
jzer0cool 6 hours ago [-]
> the people is one, ... one language, ..., nothing will be restrained from them
Why being one (I see as collaborative) was it not desired? Interpretations? Why is it seemed *more* harmful rather than good?
zoobaloo 34 minutes ago [-]
The Pope discusses this exact question at length in his recent Encyclical:
at one point - future generations - will look at people who designed unix like tools - tools that do one thing well & compose with other tools as demigods.
ares623 5 hours ago [-]
lol no. They'll "rediscover" it and claim themselves as demigods.
overgard 7 hours ago [-]
I feel like this is missing the ending of "until gravity wins"
lstodd 3 hours ago [-]
not gravity, enthropy.
jadbox 6 hours ago [-]
This reminds me of Ted Chiang's "Tower of Babylon". You really should read it (and all of TC's works)!
beardyw 7 hours ago [-]
No, the story of the tower of Babel was:
"we can, so we should".
It ended badly.
fantasizr 7 hours ago [-]
ai eliminating friction is eliminating learning and understanding. this is felt with more severe consequences in K-12 writing and music.
alkyon 3 hours ago [-]
Ultimately, AI models will need to ingest the slop ziggurat they created. This means, other things being equal, that it won't be so easy as it is now to achieve incremental gains from the sheer fact of gobbling up more code.
Invictus0 42 minutes ago [-]
this is just a software engineer's ego trying to protect itself in flowery language
jdw64 5 hours ago [-]
Since GPT 5.2, AI has been writing code much better than I can. In 5.6, it beat Tourlist in competitive coding. After seeing that, I only practice about an hour a day just to keep the feel of coding.
Honestly, rather than pointless debates about whether human coding is bad or AI coding is bad, I just think it's good to build tools that help me understand the world. I don't really care whether it's hand-coded or bad code.
Because most of my career has been spent as an on-site programmer. Staying at factories, visiting public institutions, deploying services for financial companies. My career is short, but I was lucky enough to work in various places.
When AI first came out, I thought I still wrote better code. But after the GPT 5 series, I've completely switched to AI and I'm now thinking about how to avoid errors and maintain larger codebases.
In the world I work in, it's common to see functions with 10,000 lines. Many people don't consider coupling or cohesion. So these days, instead of focusing on programming syntax, I'm studying programming theory and thinking about how to handle code when it becomes massive and turns into a black box through vibe coding. And I think this approach is right, because I believe I need to get used to using AI, so I keep coding with it.
But due to my cognitive limits, I've restricted myself to C# and TypeScript, which I'm comfortable with. C++ has too much to memorize and is hard to keep up with. In my region, there are very few C++ jobs, and those that exist are either extremely high-paying or garbage-tier jobs. There's nothing in between. So I stick with C# and TypeScript.
In practice, when building large programs, I often just set external configuration values I don't fully understand and code based on heuristics. I don't know the internals of Kafka, RabbitMQ, or PostgreSQL. I just know how to use them. And yet they work fine.
I feel the same way about AI code. Even if AI code is messy, if it runs, I use it. When bugs appear or performance is off, I just plug in debuggers or print statements and fix the necessary parts, like working with legacy code. Programming is so complex that if you try to understand everything, you can only design very small parts. Do the people who wrote Linux understand the entire codebase? They trust people they can rely on.
I've also reached an internal agreement to trust AI code. To support that, I'm spending time on creating rules for how to get good code from AI. Things like adding gates or CI, and seeing if that improves the code.
The problem is, I know this means no one will want to use other people's work or collaborate. The middle layer will disappear. There will be only highly admired projects or personal projects. In the past, even mid-sized projects had humans helping each other. But now mid-sized projects barely need human help. So I think projects will become increasingly polarized and become a zero-sum game.
Brooks divided complexity into two types in The Mythical Man-Month: Essential Complexity and Accidental Complexity. Personally, I think AI has greatly reduced Accidental Complexity. However, the essential difficulty, the problem of modeling, still needs to be done by humans. Because AI has no physical embodiment, it's inherently hard for it to understand domains the way humans do. Learning about something is different from experiencing it.
So I've decided to believe that vibe coding is also a valid approach. Supporters talk about compilers being deterministic, but LLMs are not deterministic. Critics say AI only produces garbage code, but I've seen that with high-quality prompts, the output becomes much better. Math PhDs say AI is good at things like theorem proving, and most of coding is similar to theorem proving.
It's not about good or bad. I've decided to believe it's just another approach. Yes, this is just a religion. My religion.
No matter how much people say vibe coding is bad, those who use it well do use it well. And there's no reason to criticize those who don't use it. I've just decided to treat this programming approach as a religion. Arguing about what's right or wrong is pointless anyway. Everyone has different values based on their environment, and convincing others is a waste of time.
People in open source communities might feel like AI code is destroying their communities. The code they used to communicate with, and the time they spent on it.
But for someone like me, who's been in delivery and on-site work, it feels like an escape hatch. It freed me from the hell of dealing with difficult people. So I've decided to rationalize it to myself: AI coding is just one way of doing things
I just think new methodologies will emerge. Instead of dividing code by functions or methods, people will think about how to divide things at a larger scale.
I'm just living to adapt to this era. I have nothing to lose anyway. I'm just waiting for the new era.
archonis 4 hours ago [-]
Indeed. This is an awkward era where we're using these artificial language tools to target abstractions made for humans to manually write.
At some point the path from human language to machine code is not going to require an interstitial step writing to higher level languages as we know them.
lowbloodsugar 2 hours ago [-]
Domain Driven Design
4b11b4 4 hours ago [-]
Really liked the framing on this
jagged-chisel 8 hours ago [-]
... and narrowing.
Where the "tower" was once a company (or team?) of human devs, it can now be a single dev and their agents.
The right engineer can likely replace non-technical co-founders with a couple LLMs. Geez, I can't wait to write that article...
m3kw9 8 hours ago [-]
You use a shared agents.md and an auto updated architecture doc but that is the one that needs to be heavily scrutinized and everyone gets a turn to review it.
jagenabler2 7 hours ago [-]
this doesn't work in any truly complex system. If the entire organization's shared understanding could be captured in a few documents, software engineering would've been a solved problem ages ago.
I feel like that gives an even more literal tower-rising metaphor, and that's what it feels like people using agents naively (and software engineers of lower skill or earlier-career), end up violating.
Agents are getting better at folding things into themselves, especially if you direct them to... but unfortunately I've found that the architectural instincts, even of Fable and 5.6 Sol, are still wildly behind what I reflexively achieve, say.
For sure there is an ability to have agents go back over work and try to fold it into better and better abstractions until it's sort of annealed into something good. I've done something similar on codebases that I have, but the 'high reaches' of architecture with great _prediction of how the software will evolve in the future_ in _subtle_ ways – those are, for now, out of reach of agents.
There is a part of me that wonders if it's partly just how much they can hold in their head right now, though. Even with the greatest articulation and high density of feeding them, the current setups don't allow them to hold a high-quality, sparse, 'zoomable' model of the world in their head that well yet, which we can do pretty well.
But the fact that I'm talking about it in terms of that kind of subtlety is itself promising, I guess?
We need some way to make AI-driven coding strive for parsimony.
I cannot remember who it was but there was an author who was traveling with their dog. They noticed that their dog would always pee on various tree to mark them as their territory. On their travels they ended up need some giant Red wood trees and figured, "I want to give my dog the ultimate claim of territory."
So he took his dog up to the Red wood tree and it did nothing, instead the dog wandered over to a smaller sapling and peed on that instead. The problem was the Red wood was so big and alien to the dog, it didn't recognize it as a tree.
I do wonder how many things are like that in our universe, that even if we could see them, we just wouldn't be able to understand it because it just goes beyond what we are capable of understanding. We think we have a grasp of the universe and use models to codify it but that is no guarantee that we can truly 'get it'.
Could higher level AI code be like that, would we know when we see it?
A meaningful risk of course is that the tools available to the model (ripgrep + fancier semantic approaches) allow it to do a good job of reasoning over things much larger than its context window, and so it doesn't pay the penalty sufficiently to fix it.
Look at all the libraries full of books we've built. It's useful for more than mere training sets.
The limit here I think the ancestor comments are getting at is cognitive load, which is real and measured. We only have so much memory to devote to a "stack" when executing, and it's usually quite constrained.
Hence my library mention. Humans have been doing this for millennia: orienting ourselves within a library (the physical kind, full of books) and calling upon its information resources as needed to accomplish tasks (research). Ultimately, it's all just one big cache hierarchy. Your short term memory, your long term memory, the book in your hands, the desk at the library, the nearby shelves, the card catalogue, the stacks, the inter-library loan system.
To manage it all, we humans have developed our abilities for abstraction. When we build clean, tight abstractions we reduce our cognitive load. Perhaps the best abstraction we've built so far is the TCP/IP and web stack. We don't need to care at all about the hardware details of a server in order to talk to it. It's such a powerful and airtight abstraction that we take it for granted.
I'd like to hear from more people who have spent a lot of time building with LLMs, because so far what people are saying is that these models do not have the ability to reason about and build the kind of marvellous abstractions us humans have built.
I've built a lot with LLM's, my experience sort of but not really tracks that. I've had to course correct a few bad abstractions but the larger the code base becomes the better it seems to be at reusing things. Maybe this is because of types, or spec-first development (with OpenAPI), or black box integration testing - but also maybe not. But generally I have to think about the abstractions and let the LLM fill in the details with rare exception.
What's more profitable, optimizing for inference time or optimizing to increase inference time by increasing token count?
What happens if we take the most abstract libraries in any given field - and:
1. Bound to the llm to only use those as building blocks. Does it affect his reasoning ? Will it think more abstractly ?
2. Train the llm on those, so maybe it will get a feel for abstraction ?
Maybe for simple one-person projects. We've long since developed methods and models to allow us to make things bigger than ourselves. Linux, SAP, etc. These software projects are not held in the mind of a single developer. But we use structure, rules, and other tools so that the pieces still fit together.
It's the infinite AI monkeys at a computer keyboard phenomenon.
Or the car on the highway that bumps left and right on the guardrails until, eventually, it arrives at its destination and nearly everybody is amazed at that great success.
The AI kool-aid drinkers are going to answer: "but that's how human code too".
And I'm really not sure about that.
That isn't to say software is perfectly built, but it's usually pragmatically built to balance costs of development and correctness - well chosen abstractions let us push up both qualities at once.
I assume one can't benchmaxx multi-year long efforts, clean architecture, taste etc as easily as these "make tests pass" tasks
Sorry, the lines have to clear what? Surely there must be some kind of constraint on "lines" that they have to overcome.
In code the thing has to become stable, can't just keep packing more and more noise onto it.
The codebases using technologies I have no idea about tend to quickly become unmaintainable and buggy, because the LLM still doesn't make good architectural choices, but the codebases that use technologies I'm familiar with basically never devolve into unmaintainability.
The difference between the two is massive, and that's why I think that a competent engineer steering an LLM in their area of expertise gets two orders of magnitude more productive, whereas someone steering an LLM in an area they know nothing about are basically producing tech debt at the speed of thought.
Shipping 100x more features per day?
I've written up my process here:
https://www.stavros.io/posts/how-i-write-software-with-llms/
The biggest thing to get right is to let the LLMs do what they're great at (code implementation from very detailed specs, and code review), and you do what humans are great (architecture and making sure the high level of the implementation is sane). That way, you get the best of both worlds, and a lot of speed at high quality.
Although I suspect models from Google, Facebook and Microsoft can be trained on their massive internal codebases. Whether they are is another question.
But, you would probably see a difference of scale and architecture. Larger projects that need better organization are probably more likely to be in private codebases (Linux excluded). So you might be right about the lack of private code in LLM being an issue.
At least in the past (before LLM-based code contributions got socially acceptable in some circles), in open-source project you often got very direct comments on code that was of bad quality. Yes, this was always a little bit abrasive, but it did a lot for the code quality.
For internal applications used at companies, such an abrasive behaviour is typically not accepted ("not a team player" (as if this is something bad), offended snowflakes, "not socially adept" etc.). Thus the code quality suffers quite a lot for internal applications.
Have you ever thought why salespeople have such an easy time selling some LLM for coding to big companies? Because the code quality of many internal applications is so bad, which makes even shitty LLM slop code often better than what is there.
In other words, if you can’t design a modular monolith, you can’t design a set of microservices.
I somewhat disagree:
Yes, it is possible to design a modular monolith, but thinking about the system in terms of a "minimum viable service" (but keep an eye on "viable", otherwise you can easily get into the "interaction problem") makes it much easier.
This is very similar to how you can write programs with no implicit state in a imperative programming language, but doing this in a pure functional programming language such as Haskell is much easier.
It's been a few years since I read these, but if I recall the argument there, it was that Lisp makes it so easy to build stuff and scratch exactly your own itch, that there's no real strong push for lisp programmers to come together and collaborate to build non-trivial and general purpose artifacts. And that is why the landscape of public lisp software is poorer as a result, compared to languages which demand much more effort to get anything substantial done.
Armin seems to be making a very similar point about AI coding.
[1] https://www.winestockwebdesign.com/Essays/Lisp_Curse.html
[2] https://www.marktarver.com/bipolar.html
Some have a harder time with the transition than others.
What made Lisp cool and powerful goes away when you do it through an LLM.
It is not clear at all to me that other languages "demand much more effort" for the same end result.
It is clear that many non-lisp programmers value syntax, and many lisp programmers don't. Even many people who programmed enough lisp to have their minds blown and expanded still prefer not to program in lisp. I'm still awaiting psychological studies on this, but the rift is so large, I think there may be some significantly different brain processing going on between the two groups.
To your point, yes, it is also clear that, to the extent that lisp can match the productivity of other languages, whether it exceeds them or not, one of the tools that is needed to achieve this productivity boost in lisp is heavy usage of homoiconicity, and this results in every serious lisp program being a collection of DSLs, each of which is only understood by one person or very few people.
I can guarantee you that most of these simply didn't go far enough; they had their "minds blown" with too little. They latched onto a highlight or two, found a way of sort doing that in C++ or Javascript and moved on.
There are people who have had their "minds blown" by time/mass dilation of special relativity, or phenomena in quantum physics, who don't actually know anything about physics; they couldn't work out how far a canon ball will land fired at a 45 degree angle with a certain velocity, though they can crack jokes about Schrödinger's cat.
For those of us sticking with the Lisp program, in one way or another, it was more of a sequence of small revelations in a progression of topics. Oh, that's a good way of representing that; that is pretty nicely thought out; that fits well with that; they had that how many decades ago? Sheesh; ...
Then once you start grappling with the Lisp issues, and kick the rock farther down the road, even in some small way, you are invested.
For me, the answer to this is economics. Even if you love lisp, there are way more companies hiring for stacks that don’t include lisp
If you then optimize for employability, which I assume most developers do (not all, but a large percentage), you might end up with not that many people practicing lisp regularly
So true.
Since Nov 30, 2022 everything has become… more complex.
I think introducing AI to deal with this is overall a mistake though. We're just adding more complexity on top of the existing complexity. At best, it's a massive waste of hardware. At worst, we'll probably have agents introducing as many bugs as they fix as they also drown in complexity, and a lot of stuff built using these techniques are going to be fragile garbage while the overall skillset of humanity diminishes because people aren't learning the skills anymore.
Fundamentally, software does not need to be this complicated and it's a solvable problem, but it does require people that care about craftsmanship.
Catch-22 is it's still important to know the fundamentals so you know what to ask for, but if you don't know the esoterica, the model is eventually going to make an assumption and screw things up. And the models don't have much taste either in prose, or in coding/comment style.
Therefore it is critical that whatever AI produces is understandable to us humans. That is why we must demand that AI tools and agents produce "well designed" well-structured software. That's the bottlenexk to progress I think. Even AI can't deal with exponential complexity explosion.
Drowning in complexity. Paralysis of choice.
I read a comment (joke) that if you want to follow all LLM development you should have to be unemployed.
It's not really news, though. Programming as Theory Building (Peter Naur) was published in the 80s, I think?
Maybe the younger entrants to this field never came across it, but even if you never came across it, it was common knowledge amongst experienced devs that understanding of the system you are about to change is crucial.
Thanks for mentioning Peter Naur’s Programming as Theory Building (1985).
I would add Fred Brooks and his The Mythical Man-Month.
The news is that Agentic Programming has made this always challenging task even more challenging.
It’s global madness fired up by continuous stream of news from LLM providers. It’s like The Verge that is almost about FAANG only, but multiplied as it is “magic” for most people, as vibe coding is “so easy” and dopamine-producing activity that it is similar to runners that don’t want to stop as it stimulates them (and in this case ofc it is healthy :)
HTML and pre-rendering are back in, HTMx, liveview
The degaussing of CSS and the hacks we did, hell i was trying to explain how we debugged web pages in IE6 to a younger staff member today.
Some things are more complex, some things got good enough to make them less complex.
Which ones? PostgreSQL doesn't have HA in core.
FTFY
Increasing complexity is the story of mankind. It's the story of civilization.
Someone from 20,000 BC would wander around the earth trying to find food, trying not to freeze, and trying not to get eaten. Someone from 5,000 BC would be trying to grow food, hoping it rains, and hoping disease didn't wipe out the village. The second one increases the complexity from all the systems required to manage people and keep the land growing. Today the vast majority of people on earth don't grow their own food at all, and instead are busy in some way managing the complexity of a large society.
Someone from 1970-80 would think our software from pre-llm days was vastly more complex. They'd just code directly to the hardware with no abstraction layer. Now almost no one does that. We abstracted the hardware away in most cases. With cryptography libraries for the vast majority of people it's complexity is abstracted away and mostly people are told "don't try to write your own crypto because you will fuck it up".
The question now becomes, how quickly will LLMs be able to coordinate their understanding of the system they are changing?
I think the next time I see "LLMs" and "Understanding" in the same sentence, I am going to lose it....
An LLM has zero understanding of "my", "want", or "cookie" because an LLM has no id/ego, has never felt desire, and has never eaten a cookie.
HN would commonly recommend reading the book Blindsight here.
Moreso, all you've done is recreate the Searle Chinese Room thought experiment which gets bounced around with no means of deciding if it reflects reality or not.
How'd your toddler do at IMO last year?
The hierarchy is certainly higher in agriculture societies it seems, but the complexity is up for some debate
https://en.wikipedia.org/wiki/VM_(operating_system)
This is so true. I am a big fan of Christopher Alexander’s “Pattern Language” concept, which addresses this exact problem! In fact he recommends developing your own pattern languages for your own domains (which of course led to the famous GoF Design Patterns book).
I have been experimenting with a “Pattern Language” skill which instructs the AI to maintain 3 pattern languages for every project. One in the business domain, one in the product domain, and one in the technical domain. It is working really well. It is always super cool to see it reference the pattern languages during planning and curate them during implementation and review.
I credit using it with keeping my 100% ai-coded projects well organized, aligned across domains, and easy to work on.
I don't know whether the author thinks this is a good or a bad thing, but in my eyes it's clearly a bad thing. Intelligence is knowing that a tomato is a fruit, wisdom is knowing not to put it in a fruit salad. AI is the the ultimate form of intelligence with zero wisdom. Actually, it's not even intelligence, it's an illusion of intelligence. If there is no human who can understand what the AI is doing it's time to stop and accept that we do not have the wisdom to contain what we are building.
I like that the author lets the image do the work, rather than preach at me even if I were to agree. History never repeats itself but it always rhymes.
Intelligence is learning to avoid using childish cliches, unless your intention was to mislead. Categorisation and understanding dependencies are hard enough problems already.
At the supermarket or taxation (Nix v. Hedden) tomatoes are a vegetable.
Interestingly, the Bible does not specify what kind of fruit was on the Tree of Knowledge. The association between the forbidden fruit and an apple was a derivative of later translations.
Imagine the consternation if the claim had become Adam and Eve ate of the forbidden tomato. Then this incessant controversy regarding the tomato's status would take on a biblical dimension... though then perhaps the Mormon claim that the Garden of Eden is to be found in the New World would actually have some credibility. Indeed, answers to some questions are not a matter of intelligence at all. Sometimes it seems to me intelligence is more about knowing what questions to ask in the first place.
That's my 2nd hardest skill when trying to use AI. 3rd hardest is better filtering.
I keep having flashbacks to learning how to use Google effectively.
I'm an arthiest, but I'm in a historically christian society so bible stuff comes up and I often delve in.
Whenever I do look into the bible, it amazes me just how extremely bullshitty vague the bible actually is. I don't understand why people argue about it, or pay attention to past arguments (many famous examples). If I wanted to argue about vague shit I'd be into philosophy or psychology or I could choose any of the many other branches of argument. Humans are not very logical. At least capitalism has a clearer scorecard.
Last nerdsniped look was about Angels due to Scott Alexander:
Edit: Which just meta-nerdsniped me finishing up on the beaut Hispanic word "Marianismo".Though I take issue with one statement you made - "If I wanted to argue about vague shit I'd be into philosophy..." though I do sympathize.
Modern analytic philosophy in particular, if it is good, is all about trying to be exceedingly precise, often in domains where objects like the Bible (as you noted) are exceedingly vague. Philosophy has some catching up to do, but ultimately one should not disparage it so cavalierly. After all, the best philosophers were usually also the members of society most against the sort of uncritical religiosity that seems to appeal to so many, as it does now and even more severely through the ages, before the efforts of such thinkers to enlighten their fellow human were realized.
If there is a subject that seeks to answer the question of "how to ask questions" it is surely philosophy. The best mathematicians were often well acquainted. In the most famous times of "human flourishing" the philosophers and the mathematicians were one and the same.
But yes, there is a lot of bad philosophy. And there is a lot of religion pretending to be science. Sad, but it has always been true.
Mundus vult decipi, ergo decipiatur
I feel these systems rising and sprawling with wee myopic agents developing out their little corners of this unknowably vast whole… a tower with 50 parapets on one side and some wacky cantilevered maiden tower on the other, and a very serviceable adobe roof over some patio for god-knows-why, and thatch over the landing next to it…
Some grotesque fatberg of designs that make sense at the level of individual design efforts, but that lack the fractal sort of levels of policy and judgment that unify the overall enterprise.
The overall language, as it were.
And language takes discipline to establish and maintain through any sufficiently large group of people—witness the company-speak or army-speak of pretty much any successful organization.
We feel like we’ve conquered the problem of talking the same language as our “Gastown Mayors” (who in turn are talking the same language as their “polecats” and so on all the way down the chain of golems)… but it’s only when it’s all built that the good Lord will humble us… that we’ll realize the understanding we thought we’d transmitted perfectly from our thrones wasn’t quite so shared as we’d imagined.
Padmé: "For the better, right?"
Anakin: (gazes in silence)
Padmé: "For the better, right?"
The code will do what you asked for in a broad sense but wherever choices arise on how to accomplish the goal it'll have made those choices incoherently.
There'll be strange validations applied inconsistently to some user inputs but not others. Data will be repeatedly sorted or converted to lower case for no reason at 3 stages. The column labels are all hard coded strings even though they are the first row of the input CSV. It'll create global functions taking data classes half the time and classes the other half.
If you look, and if you think you might one day need to update and maintain the code manually, then it's very hard to resist fixing this sort of thing, but fixing this sort of thing cancels a lot of the time saving out.
For example, yesterday I came across some unit tests that didn't have error messages in their assertions. Normally, it takes me ~10 minutes to fix a handful of tests in this situation. In this case, I gave a 2-3 sentence prompt, went to the bathroom, and reviewed the result after I washed my hands. Saved me a bunch of time!
I encourage you to accept a feeling of "imposter syndrome" when using it, and keep trying new things with it. Don't feel like you have to be hands off, except when you're confident that you can be. (IE, if you think you need to spend 30+ minutes on mindless refactoring, see if you can explain it to an agent and then look at HN while it runs. You might get a good result, otherwise, it probably was time for a break anyway.)
BTW: It's important to try different models. The Claude 5.0 models are slow and give me bad results, so I'm sticking with 4.x for now.
I finally learned to let go of the code. I dont even run my C++ editor anymore.
I run frequent code and architectural reviews. Its awesome.
Your test suite doesn’t cover all workflows. It doesn’t cover every combination of actions a user can take. So every big AI refactor while change some of those.
If this is happening frequently, your software will feel like a janky piece of unusable crap.
Large Language Models are the most powerful communication tools to have ever existed - probably behind the written and spoken word and marginally behind the internet, maybe more important than phones and texting.
With understanding now on tap in this new form of infrastructure, the tower (codebase) and our coordination within it are both very customizable - up, down, sideways - per the will of the team and maintainers.
The sample size of your architectural taste, intuition, or expertise at abstraction may never be large enough for a LLM to match it. But if you want it to, it too can keep getting better and closer at this as it seeks alignment with your acceptance.
Once your tower is built, what's next?
Not that we won't have most code written by agents. I truly beleive that to be true. But that we got so far down the "never look at code" rabbit hole that our abstractions will become so divorced from the reality that is the actual.. freaking... code. We will truly have spaghetti code, bloated towers of babble.
We'll tut-tut these times and think "if you only spent half a day thinking about the big picture, looking at code, dare I say write some code you could have saved the project a whole mess of trouble".
In the meantime we're going to be generate 10K in spaghetti what should probably be 500 reasonably understandable LoC.
Spending a day reading some code can save you $100s of tokens, and weeks of headaches.
It was all good when change was slow and needed deep contemplation. Things sped up this year. Chaos is harder to defeat now. It really does need a different paradigm than "people facing into corners each doing today's task".
https://www.vatican.va/content/leo-xiv/en/encyclicals/docume...
Comparing AI adoption to the Tower of Babel is a central theme to Magnifica Humanitas, but written by someone with a much deeper understanding of the Biblical narrative.
I'm not sure reading code is coming back. The ritual of reading code must come back, because that's the only way to build products that don't collapse under their own incoherence, both technically and visibly.
"just ask Claude" is fine, but it's not the end state
I mean, yeah, hobby coding is not going away, but the feeling of exploration for me is totally gone. I don't dislike tech, I just don't see anyone being the way they used to be. People in tech are different and too many people are in tech.
Probably only part of the joy will ever be there, now. Which is weird because I did my own thing with computers til college, and don't consider myself a super sociable person. I just used to know that it was there, and now it isn't.
The tricky part here is that you can't tell if a once-topmost part of the tower is sturdy until a great deal more tower is resting on it. Well, now a lot the economy is resting on little other than AI dreams. Your move, rational people.
Is that because of the technology or because of who you were at the time?
The most interesting problems arise when you don't try and force one shared standard upon everybody yet still try and play nice.
Alternatively, power could concentrate and the winners get to decide what is valuable and not, thus cutting down the space of possible complexity by construction.
It seems to me that LLMs and particularly chatbots have already allowed for bigger scale collaboration within the LLM companies versus what was possible within the prior cohort of big platform companies.
Has the result just been taller towers, or actually a change of what is possible?
But this is just bad vibecoding? This would be bad if humans did it too. With agents or humans, you need to coordinate.
We are going through a transition from a guild based software production with primitive division of labour to a machinery based one where AI is the steam engine and the job of the engineer is to build the production line, be the mechanic fixing the line, and also the assembly line worker.
1. Perhaps with a handful of skyscrapers sprinkled in.
Why being one (I see as collaborative) was it not desired? Interpretations? Why is it seemed *more* harmful rather than good?
https://www.vatican.va/content/leo-xiv/en/encyclicals/docume...
It's odd that TFA doesn't acknowledge this.
"we can, so we should".
It ended badly.
Honestly, rather than pointless debates about whether human coding is bad or AI coding is bad, I just think it's good to build tools that help me understand the world. I don't really care whether it's hand-coded or bad code.
Because most of my career has been spent as an on-site programmer. Staying at factories, visiting public institutions, deploying services for financial companies. My career is short, but I was lucky enough to work in various places.
When AI first came out, I thought I still wrote better code. But after the GPT 5 series, I've completely switched to AI and I'm now thinking about how to avoid errors and maintain larger codebases.
In the world I work in, it's common to see functions with 10,000 lines. Many people don't consider coupling or cohesion. So these days, instead of focusing on programming syntax, I'm studying programming theory and thinking about how to handle code when it becomes massive and turns into a black box through vibe coding. And I think this approach is right, because I believe I need to get used to using AI, so I keep coding with it.
But due to my cognitive limits, I've restricted myself to C# and TypeScript, which I'm comfortable with. C++ has too much to memorize and is hard to keep up with. In my region, there are very few C++ jobs, and those that exist are either extremely high-paying or garbage-tier jobs. There's nothing in between. So I stick with C# and TypeScript.
In practice, when building large programs, I often just set external configuration values I don't fully understand and code based on heuristics. I don't know the internals of Kafka, RabbitMQ, or PostgreSQL. I just know how to use them. And yet they work fine.
I feel the same way about AI code. Even if AI code is messy, if it runs, I use it. When bugs appear or performance is off, I just plug in debuggers or print statements and fix the necessary parts, like working with legacy code. Programming is so complex that if you try to understand everything, you can only design very small parts. Do the people who wrote Linux understand the entire codebase? They trust people they can rely on.
I've also reached an internal agreement to trust AI code. To support that, I'm spending time on creating rules for how to get good code from AI. Things like adding gates or CI, and seeing if that improves the code.
The problem is, I know this means no one will want to use other people's work or collaborate. The middle layer will disappear. There will be only highly admired projects or personal projects. In the past, even mid-sized projects had humans helping each other. But now mid-sized projects barely need human help. So I think projects will become increasingly polarized and become a zero-sum game.
Brooks divided complexity into two types in The Mythical Man-Month: Essential Complexity and Accidental Complexity. Personally, I think AI has greatly reduced Accidental Complexity. However, the essential difficulty, the problem of modeling, still needs to be done by humans. Because AI has no physical embodiment, it's inherently hard for it to understand domains the way humans do. Learning about something is different from experiencing it.
So I've decided to believe that vibe coding is also a valid approach. Supporters talk about compilers being deterministic, but LLMs are not deterministic. Critics say AI only produces garbage code, but I've seen that with high-quality prompts, the output becomes much better. Math PhDs say AI is good at things like theorem proving, and most of coding is similar to theorem proving.
It's not about good or bad. I've decided to believe it's just another approach. Yes, this is just a religion. My religion.
No matter how much people say vibe coding is bad, those who use it well do use it well. And there's no reason to criticize those who don't use it. I've just decided to treat this programming approach as a religion. Arguing about what's right or wrong is pointless anyway. Everyone has different values based on their environment, and convincing others is a waste of time.
People in open source communities might feel like AI code is destroying their communities. The code they used to communicate with, and the time they spent on it.
But for someone like me, who's been in delivery and on-site work, it feels like an escape hatch. It freed me from the hell of dealing with difficult people. So I've decided to rationalize it to myself: AI coding is just one way of doing things
I just think new methodologies will emerge. Instead of dividing code by functions or methods, people will think about how to divide things at a larger scale.
I'm just living to adapt to this era. I have nothing to lose anyway. I'm just waiting for the new era.
At some point the path from human language to machine code is not going to require an interstitial step writing to higher level languages as we know them.
Where the "tower" was once a company (or team?) of human devs, it can now be a single dev and their agents.
The right engineer can likely replace non-technical co-founders with a couple LLMs. Geez, I can't wait to write that article...