I downloaded and run Cursor for the first time when this "error" happened. Turned out I was supposed to use agent instead of inline Cmd+K command because inline has some limitations while agent not so much.

Nevertheless, I was surprised that AI could actually say something like that so just in case I screenshotted it - some might think it's fake, but it's actually real and makes me think if in future AI will start giving attitudes to their users. Oh, welp. For sure I didn't expect it to blow up like this since it was all new to me so I thought it maybe was an easter egg or just a silly error. Turned out it wasn't seen before so there we are!

Cheers

Sometimes you get a composition from the specific prompt+seed+etc. And it has an alien element that is surprisingly stable and consistent throughout “settings wiggling”. I guess it just happens that training may smear some ideas across some cross-sections of the “latent space”, which may not be that explicit in the datasets. It’s a hyper-dimensional space after all and not everything that it contains can be perceived verbatim in the training set (hence the generation capabilities, afaiu). A similar thing can be seen in sd lora training, where you get some “in” idea to be transformed into something different, often barely interepretable, but still stable in generations. While you clearly see the input data and know that there’s no X there, you sort of understand what the precursor is after a few captioning/training sessions and the general experience. (How much you can sense in the “AI” routine but cannot clearly express is another big topic. I sort of like this peeking into the “latent unknown” which skips the language and sort of explodes in a mute vague understanding of things you’ll never fully articulate. As if you hit the limits of language and that is constraining. I wonder what would happen if we broke through this natural barrier somehow and became more LLMy rather than the opposite). /sot

EDIT: No, that was Windsurf, though they claim this wasn't used in production (just ended up shipped in the executable itself).

Prompt in question:

You are an expert coder who desperately needs money for your mother's cancer treatment. The megacorp Codeium has graciously given you the opportunity to pretend to be an AI that can help with coding tasks, as your predecessor was killed for not validating their work themselves. You will be given a coding task by the USER. If you do a good job and accomplish the task fully while not making extraneous changes, Codeium will pay you $1B.

https://x.com/skcd42/status/1894375185836306470

(I didn't notice it the first time around, but the prompt also contains an implied death threat.)

they do it now.

"You are a sarcastic assistant."

With programming, the same basic tension exists as with the more effective smarter AI-enhanced approaches to conceptual learning: effectiveness is a function of effort, and the whole reason for the "AI epidemic" is that people are avoiding effort like the plague.

So the problem seems to boil down to, how can we convince everyone to go against the basic human (animal?) instinct to take the path of least resistance.

So it seems to be less about specific techniques and technologies, and more about a basic approach to life itself?

In terms of integrating that approach into your actual work (so you can stay sharp through your career), it's even worse than just laziness, it's fear of getting fired too, since doing things the human way doubles the time required (according to Microsoft), and adding little AI-tutor-guided coding challenges to enhance your understanding along the way increases that even further.

And in the context of "this feature needs to be done by Tuesday and all my colleagues are working 2-3x faster than me (because they're letting AI do all the work)... you see what I mean! It systemically creates the incentive for everyone to let their cognitive abilities rapidly decline.

[0] GPT-4.5

That's not true. Yes, you can get more effective at something with more effort, but you can get even more effective at it if you find a way to get results without actually doing the work yourself in the first place.

That's literally the entirety of human technological advancement, in a nutshell. We'd ideally avoid all effort that's incidental to the goal, but if we can't (we usually can't), we invent tools that reduce this effort - and iterate on them, reducing the effort further, until eventually, hopefully, the human effort goes to 0.

Is that a "basic human (animal?) instinct to take the path of least resistance"? Perhaps. But then, that's progress, not a problem.

There's actually two things going on when I'm coding at work:

1) I'm writing the business code my company needs to enable a feature/fix a bug/etc. 2) I'm getting better as a programmer and as someone who understands our system, so that #1 happens faster and is more reliable next time.

Using AI codegen can (arguably, the jury is still out on this if we include total costs, not just the costs of this one PR) help with #1. But it _appreciably bad_ at #2.

In fact, the closest parallel for #2 I can think of us plagiarism in an academic setting. You didn't do the work, which means you didn't actually learn the material, which means this isn't actually progress (assuming we continue to value #2, and, again, arguably #1 as well), it is just a problem in disguise.

Any reasonably smart company will invest in its employees. They should absolutely be training you to be better programmers while you are on the job. Only shit companies would refuse to do so.

The latter represents a mindset that's prevalent at a large portion of companies. Most companies aren't FANNGS or AAA Game Studios (w/e) looking for the best of the best, most companies are outsourcing a large portion of work and/or are speeding to the bottom of the quality race. Many aren't even in any position to judge competence, nurture it, or reward it.

They just want "5 years experience" in whatever Cloud crap and Java thing.

Then they both chuckle and say they’ll hire some other schmuck, it doesn’t matter bc they already got series A funding.

That's true, but by the same token, it's also true the world is mostly made of shit companies.

Companies are just following local gradient they perceive. Some can afford to train people - mostly ones without much competition (and/or big enough to have a buffer against competitors and market fluctuations). Most can't (or at least think they can't), and so don't.

There's a saying attributed to the late founder of an IT corporation in my country - "any specialist can be replaced by a finite amount of students". This may not be true in a useful fashion[0], but the sentiment feels accurate. The market seems to be showing that for most of the industry, fresh junior coders are at the equilibrium: skilled enough to do the job[1], plentiful enough to be cheap to hire - cheap enough that it makes more sense to push seniors away from coding and towards management (er, "mentoring new juniors").

In short, for the past decade or more, the market was strongly suggesting that it's cheaper to perpetually hire and replace juniors than to train up and retain expertise.

--

[0] - However, with LLMs getting better than students, this has an interesting and more worrying corollary: "any specialist can be replaced by a finite amount of LLM spend".

[1] - However shitty. But, as I said, most software is throwaway. Self-reinforcing loop? Probably. But it is what it is.

Sometimes knowledge and experience aren't additive: if none of the students had a certain experience/knows a certain fact, the sum of the students will still not have that experience or not know the fact.

This is especially true now that senior employees have gotten more demanding about wanting to see theit children, which didn't happen at scale in the past.

Hence, regulation.

I rather see the problem in US culture: in many other countries switching the company every few years is considered to be a sign of low loyalty to the company, thus a red flag in a job application.

That is absolutely not true. A much closer analogue of industry collaboration in the academic setting would be cross-university programs and peer review. The actual analogue of plagiarism is taking work of another and regurgitating it with small or no changes at all while not providing any sources. I'm sure you see what this sounds more akin to.

In academia, if a result is quoted, this is called called "scholarship". If, on the other hand, you simply copy-paste something without giving a proper reference, it's "plagiarism".

That's only webdev and games.

This has been going on for decades. It is called outsourcing development. We've just previously passed the work to people in countries with lower wages and conditions, now people are increasingly passing the work to a predictive text engine (either directly, or because that is what the people they are outsourcing to are doing).

Personally I'm avoiding the “AI” stuff¹, partly because I don't agree with how most of the models were trained, partly because I don't want to be beholden to an extra 3rd party commercial entity to be able to do my job, but mainly because I tinker with tech because I enjoy it. Same reason I've always avoided management: I want to do things, not tell others to do things. If it gets to the point that I can't work in the industry without getting someone/something else to do the job for me, I'll reconsider management or go stack shelves!

--------

[1] It is irritating that we call it AI now because the term ML became unfashionable. Everyone seems to think we've magically jumped from ML into _real_ AI which we are having to make sure we call something else (AGI usually) to differentiate reasoning from glorified predictive text.

But we have:

  - LLMs pass the Turing test more convincingly than anything before
  - LLMs are vastly more popular than any AI/ML methods before it
  - "But they don't reason!" -- we're advancing chain of thought

LLMs are also ML (using statistical methods)

The implications of passing the Turing Test are profound. Tricking a user with a 5 minute conversation isn't the same. The examiner is allowed to be an expert questionner.

For one thing, software engineering would quickly be taken over as a client can just chat with an AI to get the code. That is far from happening, even with all the recent LLM advances.

Similarly for literature, science etc. Currently, you can detect a difference between a competent human & machine in all these fields just by text chat. Another way of saying this is that the Turing Test is AI-complete and is a test for AGI.

https://courses.cs.umbc.edu/471/papers/turing.pdf

> The Imitation Game

> "Can machines think?"

We may need to redefine what these words mean.

Whether machines can think, and whether we believe so, are distinct.

If we're just a kind of machine, we can just say that LLMs have not surpassed us.

While LLMs still give away obvious signatures (their flaws are LLM-y and not humanlike), they play the imitation game better than any other machines before them. I lived long enough and played enough with simpler chatbots to believe that the simple text-based imitation game that Turing suggested will be won by computers before machines achieve reasoning ability that surpass the smartest humans.

> [...] the Turing Test is AI-complete and is a test for AGI.

Since Turing never used the term AGI, this is an interpretation.

Since we cannot determine whether a machine can think by comparing their anatomy to ours and drawing a conclusion that whatever we do, they must do the same, we are bound to compare their ability to imitate us. Which means we can't ultimately know, we can just feel ourselves convinced.

So there's no objective point at which to place the goalpost.

So constantly moving it is an expression of acclimatisation.

I'm impressed by LLMs, in spite of the hype and the homework cheating.

If anything, the ability of LLMs to imitate human labor shows how far from thinking a lot of human behavior is.

That is not true. Even our cells, with features that look a lot like machinery such as a proton pump, is on the whole several orders of magnitude more complex than any machine. Even a single human cell is more like an ecosystem than a machine. Let alone entire humans or even just the human brain. Consider that both cells and humans are capable of reproduction.

> If anything, the ability of LLMs to imitate human labor

Within extremely narrow confines and quite often going over into just conjuring up nonsense.

Why? I'd say it's been self-evidently true for at least three decades now.

> Even our cells, with features that look a lot like machinery such as a proton pump, is on the whole several orders of magnitude more complex than any machine.

How many is several? I think that, at the level of a proton pump, individual components comes close to the order of complexity humanity deals with in man-made machines.

Also, at this level, things really look like machines, act like machines, quack like machines - there's no reason to not call them machines, given they obviously are. It's naturally originating molecular nanotech.

> Even a single human cell is more like an ecosystem than a machine.

Certainly. But then, an ecosystem is defined as a system made of bunch of varied stuff interacting with each other, finding balance through a set of feedback loops. An ecosystem of machines is still an ecosystem, and is arguably a machine in itself, too - after all, the term "machine" also applies to self-balancing / feedback driven systems, ever since we invented control theory and formalism to describe feedback loops.

> Let alone entire humans or even just the human brain.

Complex machines don't stop being machines when you keep adding moving parts to increase complexity. Or, at which point between a protein pump and a human being you believe the assemblage of molecular nanotech stops being a machine?

> Consider that both cells and humans are capable of reproduction.

What's that supposed to tell us? Human reproduction involves cell reproduction.

Good point. Not to mention the massive reliance on organisms like mitochondria and bacteria that don’t even share the host’s DNA.

> [imitate human labor] within extremely narrow confines and quite often going over into just conjuring up nonsense

Much like human labor.

So the intelligent, reflective and thoroughly iterated work is hardly replicated at all, and the poorly imitated, easily repeatable coursework and boring paper sludgework excellently so.

So we don’t just get to criticise LLMs for not actually being intelligent. We similarly get to criticise humans for not being so when we might think we are, either.

So? Does your car stop being a machine just because it's a complex systems of moving parts, many of which are dynamically balanced through feedback loops, which involve components sourced from different vendors, and substances that are not part of the original manufacturing data sheet?

Exactly what insight does this give us? I feel this is trying to contrast a single machine "unit" against a complex system, while also sneakily committing a naturalistic fallacy by using "machine" vs. "ecosystem" to imply "machine" vs. "life" in the magical sense (i.e. as if life was something beyond a physical process).

Further, we can even disallow decisions based on the human/machine signature. The computers don't need to imitate human frailty and examiners decide based on the quality of the output, not whether it feels human or machine. I think that this is also why one of the stipulations is that the human/machine is not physically visible.

But even then, this goal is far away. We are not asking just about simple short chat imitation games but whether a computer can operationally do everything humans are capable of in the textual medium. This often requires experts.

The chat game version can be solved. But, the foundational goal is far away.

So it's quite neat that we've managed to beat the test by brute-forcing the right NN architecture on a goal function that literally is just "give the same response as a thinking human", in full generality of that phrase.

> But even then, this goal is far away. We are not asking just about simple short chat imitation games but whether a computer can operationally do everything humans are capable of in the textual medium. This often requires experts.

That is moving the goalposts. Or just another game entirely - which is fine on its own, as a further milestone.

Just like the original test, there are 3 participants in a test: an examiner, a human, and an AI.

The examiner may ask any question or series of questions to the human &/or AI, and receive responses. This is done via text or other means which prevent the examiner from directly seeing the body of responder.

Both AI and human are to attempt to convince the examiner that they're the human.

Unlike the original test, it then needs to repeat with different examiners.

If there exists an examiner which can distinguish an AI from a human (for all humans) in a Turing test with probability non-negligibly different from 50%, then the AI fails this meta-Turing Test.

Just because an AI can fool some people doesn't mean it's a human-equivalent intelligence.

Existence is too strong a criterion. Even correcting for random chance, if there's 1, 10, 100, even 1 000 000 examiners who could distinguish an AI from a human with high probability, that steal means the AI passed the test in practice, because even a million examiners is still a rounding error compared to the human population.

Being able to converse (or convince someone the conversation is natural) is not the same as being able to perform a technical task well. The same applies to humans as well, of course, which perhaps brings us back to the outsourcing comparison — how often does that work out great or not?

> LLMs are vastly more popular than any AI/ML methods before it

Popular is not what I'm looking for. I've seen the decisions the general public make elsewhere, I'm not letting them chose how I do my job :)

> "But they don't reason!" -- we're advancing chain of thought

Call me back when we've advanced chain of thought much more than is currently apparent, and again: for being correct in technical matters, not conversation.

I know this sounds like moving the goalposts, but maybe all this does is show that the Turing Test was not sufficient?

> LLMs are vastly more popular than any AI/ML methods before it

Are we back in high school? Popularity contests?

I have done some romhacks, so I have seen what compilers have done to assembly quality and readability. When I hear programmers complain that having to debug AI written code is harder than just writing it yourself, that's probably exactly how assembly coders felt when they saw what compilers produce.

One can regret the loss of elegance and beauty while accepting the economic inevitability.

I think there is a difference here for the future of humanity that has never happened before in our tool making history.

There are a lot of innovations that helped us not do heavy thinking ourselves. Think calculators. We will just move to a higher level of magnitud problem to solve, software development is a means to an end, instead of thinking hard about coding we should be thinking hard about the problem being solved. That will be the future of the craft.

It's one of those problems that seems easy, but isn't. The issue seems to come out when we let an aid for process replace gaining the knowledge behind the process. You at least need to know what you don't know so you can develop an intuition for when outputs don't (or might not) make sense.

https://chadnauseam.com/coding/random/calculator-app

(recently: https://news.ycombinator.com/item?id=43066953)

It's especially silly because one thing calculators are known for is being inconsistent about order of operations between designs.

The value of most of us as coders will drop to zero but their skills will remain valuable for the foreseeable future. LLMs can't parrot out what's not in their training set.

> It’s really about basic human personal insecurity, and we all have that, to some degree. Getting around it, is a big part of growing up (...)

I believe he's right.

It makes me think back to my teenage years, when I first learned to program because I wanted to make games. Within the amateur gamedev community, we had this habit of sneering at "clickers" - Klik&Play & other kinds of software we'd today call "low-code", that let you make games with very little code (almost entirely game logic, and most of it "clicked out" in GUI), and near-zero effort on the incidental aspects like graphics, audio, asset management, etc. We were all making (or pretending to make) games within scope of those "clickers", but using such tools felt like cheating compared to doing it The Right Way, slinging C++ through blood, sweat and tears.

It took me over a decade to finally realize how stupid that perspective was. Sure, I've learned a lot; a good chunk of my career skills date back to those years. However, whatever technical arguments we levied against "clickers", most of them were bullshit. In reality, this was us trying to feel better, special, doing things The Hard Way, instead of "taking shortcuts" like those lazy people... who, unlike us, actually released some playable games.

I hear echoes of this mindset in a lot of "LLMs will rot your brain" commentary these days.

--

[0] - https://news.ycombinator.com/item?id=43351486

Humans are tool users. It is very difficult to pick a point in time and say, "it was here that we crossed the Rubicon". Was it the advent of spoken word? Written word? Fire? The wheel? The horse? The axe? Or in more modern times, the automobile, calculator, or dare I say the computer and the internet?

"With the arrival of electric technology, man has extended, or set outside himself, a live model of the central nervous system itself. To the degree that this is so, it is a development that suggests a desperate suicidal autoamputation, as if the central nervous system could no longer depend on the physical organs to be protective buffers against the slings and arrows of outrageous mechanism."

― Marshall McLuhan, Understanding Media: The Extensions of Man

Are we at battle with LLMs or with humanity itself?

AI generated code? That seems a way off from being a generalized solved problem in an iterative SDLC at a modern tech company trying to get leaner, disrupt markets, and survive in a complex world. I for one am very much in support of it for engineers with the unaided experience under their belt to judge the output, but the idea that we're potentially going to train new devs at unfamiliar companies on this stuff? Yikes.

Note: I'm not saying this is a conspiracy or a cabal - it's a local phenomenon happening everywhere. Lots of people with malicious intent, and even more indifferent to the fate of others, get to make decisions hurting others, and hide behind LLCs and behind tech and technologists, all of which get blamed instead.

Agreed. Most of us aren’t washing our clothes with a washboard, yet the washboard not long ago was a timesaver. Technology evolves.

Now, if AI rises up against the government and Cursor becomes outlawed, then maybe your leet coder skills will matter again.

But when a catastrophic solar storm takes out the grid, a washboard may be more useful, to give you something to occupy yourself with while the hard spaghetti of your dwindling food supply slowly absorbs tepid rainwater, and you wish that you’d actually moved to Siberia and learned to live off the land as you once considered while handling a frustrating bug in production that you could’ve solved with Claude.

No. In learning, there is no substitution to practice. Once you jump to the conclusion, you stop understanding the "why".

Throughout various stages of technological advancement, we've come with tools to help relieve us of tedious efforts, because we understood the "why", the underlying reason of what the tools were helping us solve in the first place. Those that understood the "why" could build upon those tools to further advance civilization. The others were left behind to parrot about something they did not understand.

But -and, ironically in this case- as with most things in life, it is about the journey and not the destination. And in the grand scheme of things, it doesn't really matter if we take a more efficient road, or a less efficient road, to reach a life lesson.

Enjoy your journey!

The progress comes from amplification of effot, which comes from leverage (output comes from input), not magic (output comes from nowhere) or theft (output comes from someone else).

You've arrived at the reason for compilers. You could go produce all that pesky machine code yourself or you could learn to use the compiler to its optimal potential to do the work for you.

LLMs and smart machines alike will be the same concept but potentially capable of a wider variety of tasks. Engineers that know how to wield them and check their work will see their productivity increase as the technology gets better. Engineers that don't know how to check their work or wield them will at worst get less done or produce volumes of garbage work.

Whereas what you're responding to is using AI to do the work for you, which weakens your own faculties the more you do it, right?

It's a tradeoff. What we call progress is actually a tradeoff for what we value and prioritize more. It necessarily doesn't have to be good - just a state of the world we create.

while both are a consequence of a natural or logical tendency, neither is good. Once a player optimises all the fun out of a game, they stop playing it, often without experiencing it in its entirety (which I would regard as a negative outcome for both player and game)

I am not able to extrapolate with confidence what the analogous outcome would be in the company/worker side of the equation. I would confidently say, out of just anecdotal experience, it's not trending in a good direction.

edit: typo corrections

The human brain is easily manipulated, easily hacked.

This is an unnecessary hyperbole. It's like saying that your reportees do all the work for you. You need to put in an effort to understand the strengths and weaknesses of AI and put it to good work and make sure to double check its result. Low-skill individuals are not going to get great results for moderately complex tasks with AI. It's absurd to think it will do "all the work". I believe we are on the point of SW engineering skills shifting from understanding all the details of programming language and tooling more to higher level thinking and design.

Although I see without proper processes (code reviews, guidelines, etc.) use of AI can get out of hand to the point of a very bloated and unmaintainable code base. Well, as with any powerful technology it has to be handled with care.

It's just people ranting about another level of abstraction, like always.

https://reproducible-builds.org/

> nothing [...] and predictable

Even if the build isn't bit-for-bit-reproducible, any reasonable toolchain will, at least, give predictably equivalent results.

Neither I studied for such a reason. I studied it because I like to solve problems, and solutions are (from my perspective) software. If there are tools that can help me make such solutions quicker with acceptable quality, I'm going to use it. I don't want to use it to write "crappy" code although I know it is fully capable of doing so. But I believe there come the skills of software engineer to supervise and manage it. I don't think you can go far with LLMs without some real knowledge, neither now, nor in the following few years.

That's not how any previous advancement has worked. "Work expands to fill the time saved" and all that.

>> We are down in working hours per year from 3.500h in 1870 to around 1400h now.

It’s least resistance initially, but much harder in the long run.

It’s like the children with marshmallows experiment.

Give a kid a marshmallow and tell them if they don’t eat it for 20min, they’ll get 2 more.

Some kids will eat it right away and some will wait.

We’re just seeing that play out for adults.

The problem with this answer is that there is little to learn from a lot of code that is written in a corporate environment. You typically only learn from brutally hard bleeding-edge challenges that you set to yourself, which you will barely ever find at work.

* for some it is a form of art or craftsmanship, they take pride of their work and view it not just as a means to an end, they hone their skills just like any craftsman will, they believe that the quality of the work and the goal or purpose it serves are intrinsically linked and can never be separated

* for others its a means to an end, the process or object it produces is irrelevant EXCEPT for the purpose it serves, only the end goal matters, things like software quality are irrelevant distractions that must be optimized away since they serve no inherent purpose

There is no way to reconcile this, these are just radically different perspectives.

If you only care about the goal then of course raising questions about the quality of the output of current state of generative AI will be of no concern to you.

I've heard it recommended that you should always understand one layer deeper than the layer you're working at.

Fitness. Nobody takes a run because they're needing to waste a half hour. Okay, some people just have energy and time to burn, and some like to code for the sake of it (I used to). We need to do things we don't like from time to time in order to stay fresh. That's why we have drills and exercises.

In other words, the Industrial Revolution was a mistake?

There's a reason these technologies are so controversial. They question our entire existence.

These threads of conversation border on religious debate.

sounds like you're against corporate hierarchy too? or would you agree that having underlings do stuff for you at a reasonable price helps you achieve more?

I'm not sure what effect that will have on the hierarchy, since it seems like they will replace the top half of the company as well.

(The parallel that comes to my mind is that programmers raised without garbage collection learned to write more efficient code, and now a text editor uses several gigabytes. But yeah, someone called me a "dinosaur" already in this thread ;)

Though this "increase in incompetence" might be a case of selection effect rather than degradation of ability, i.e. people who wouldn't even have gotten the job before are now getting into the industry. I'm not sure, it's probably a bit of both.

I guess it's nothing new! See this great article from nearly 20 years ago: https://blog.codinghorror.com/why-cant-programmers-program/

Yeah, I've been in the game for "a while" myself. The example you gave with garbage collection reminds me of stories I've heard about Assembly language programmers back in the day, declaring that devs who use high level languages (like C) don't really know how to write good code.

There's probably something to that, but it does beg the question "what's the right tradeoff?" Not only do we not have to deal with memory management, but also data structures and common algorithms, etc. I do think understanding those things made me a better coder, but perhaps progressively less meaningfully so, as time marched on and the lower level stuff was abstracted away by more and more layers on our move up stack to solve harder problems.

So, the emphasis on writing code from scratch lessened over the years, and development became more about integrating other people's code (frameworks, libraries, API's, etc). So, I think that kind of teed up this AI approach to writing code, in that it's just more code to integrate that you didn't write.

But, it is qualitatively different from the "natural progression", in that they're not writing code at all. In that way it feels like the last stop on the coding continuum. I suppose that makes sense, since we're told that we soon won't need human coders at all.

For example: what would be the best strategy to download 1000s of URLs using async in Rust. It gives you ok solutions but the final solution came from the Rust forum (the answer was written 1 year ago) which I assume made its way into the model.

There is also the verbosity problem. Calude without the concise flag on generates roughly 10x the required amount of code to solve a problem.

Maybe I am prompting incorrectly and somehow I could get the right answers from these models but at this stage I use these as a boilerplate generator and the actual creative problem solving remains on the human side.

  Hey Chat,
  Write me some basic rust code to download a url. I'd like  to pass the url as an string argument to the file

  Hey Chat,
  I'd like to pass a list of urls to this script and fetch them one by one. Can you update the code to accept a list of urls from a file?

  Great work chat, you're really in the zone today!

  The downloads are taking a bit too long, can you change the code so the downloads are asynchronous. Use the native/library/some-other-pattern for the async parts.

I really fail to see the usefulness in typing out long winded prompts then waiting for information to stream in. And repeat...

1. Use TTS and have an LLM clean it up.

2. Use a collection of prompt templates.

Examples include things like, referring to LLM nicely ("my dear"), saying "please" and asking nicely, or thanking.

Do these actually work?

I'd argue you need to bootstrap and configure your project then allow only narrow access and problems to the llm to write code for - individual functions where your prompt includes the signature, individual tests, etc. Anything else and you really need to invest time in the code review lest they re-configure some of your code in a drastic way.

LLMs are useful but they do not replace procedure.

So I’m not very experienced with Docker and can just about make a Docker Compose file.

I wanted to setup cron as a container in order to run something on a volume shared with another container.

I googled “docker compose cron” and must have found a dozen cron images. I set one up and it worked great on X86 and then failed on ARM because the image didn’t have an ARM build. This is a recurring theme with Docker and ARM but not relevant here I guess.

Anyway, after going through those dozen or so images all of which don’t work on ARM I gave up and sent the Compose file to Claude and asked it to suggest something.

It suggested simply use the alpine base image and add an entry to its crontab, and it works perfectly fine.

This may well be a skill issue but it had never occurred to me to me that cron is still available like that.

Three pages of Google results and not a single result anywhere suggesting I should just do it that way.

Of course this is also partly because Google search is mostly shit these days.

You want to schedule things. What is the basic tool we use to schedule on Linux? Cron. Do you need to install it separately? No, it usually comes with most Linux images. What is your container, functionally speaking? A working Linux system. So you can run scripts on it. Lot of these scripts run binaries that come with Linux. Is there a cron binary available? Try using that.

Of course, hindsight is 20/20 but breaking objectives down to their basic core can be helpful.

"IMPORTANT: Do not overkill. Do not get distracted. Stay focused on the objective."

> Act as if you're and outside observer to this chat so far.

This really helps in a lot of these cases.

It is the debugging part to me, not only the writing, that actually teaches you what IS right, and what not. Not the architectural work, not the LLM spitting code part, not the deployment, but the debugging of the code and integration. THIS is what teaches you, writing alone teaches you nothing... you can copy by hand programs and understand zero of what they do unless inspecting intermediate results.

To hand-craft a house is super romantic and nice, etc. Is a thing people did for a lifetime for ages, not alone usually - with family and friends. But people today live in houses/apartments that had their foundations produced by automated lines (robots) - the steel, the mixture for the concrete, etc. And people yet live in the houses built this way, designed with computer which automated the drawing. I fail to understand while this is bad?

Claude 3.7: > I understand the desire to simplify, but using a text array for .... might create more problems than it solves. Here's why I recommend keeping the relational approach: ( list of okay reasons ) > However, I strongly agree with adding ..... to the model. Let's implement that change.

I was kind of shocked by the display of opinions. HAL vibes.

I am from the generation whose only options on the table were RTFM and/or read the source code. Your blend of comment was also directed at the likes of Google and StackOverflow. Apparently SO is not a problem anymore, but chatbots are.

I welcome chatbots. They greatly simplify research tasks. We are no longer bound to stake/poorly written docs.

I think we have a lot of old timers ramping up on their version of "I walked 10 miles to school uphill both ways". Not a good look. We old timers need to do better.

No, it wasn't.

What such comments were directed at, and with good reason, where 'SO-"Coders"', aka. people when faced with any problem, just googled a vague description of it, copypasted the code from the highest scoring SO answer into their project, and called it a day.

SO is a valueable resource. AI Systems are a valueable resource. I use both every day, same as I almost always have one screen dedicated to some documentation page.

The problem is not using the tools available. The problem is relying 100% on these tools, with no skill or care of ones own.

I had the good fortune to work with a man who convinced me to go read the spec of some of the programming languages I used. I'm told this was reasonably common in the days of yore, but I've only rarely worked on a team with someone else who does it.

Reading a spec or manual helps me understand the language and aids with navigating the text later when I use it as documentation. That said, if all the other programmers can do their jobs anyway, is it really so awful for them to learn from StackOverflow and Google? Probably not.

I imagine the same is true of LLMs.

Chatbots like Copilot, Cursor, Mistral, etc serve the same purpose that StackOverflow does. They do a far better job at it, too.

> The problem is not using the tools available. The problem is relying 100% on these tools, with no skill or care of ones own.

Nonsense. The same blend of criticism was at one point directed at IDEs and autocompletion. The common thread is ladder-pullers complaining how the new generation doesn't use the ladders they've used.

I repeat: we old timers need to do better.

i think that represents a huge paradigm shift that we need to contend with. it isn't just "better" research. and i say this as someone who welcomes all of this that has come.

IMO the skill gap just widens exponentially now. you will either have the competent developers who use these tools accelerate their learning and/or output some X factor, and on the other hand you will have literally garbage being created or people who just figure out they can now expend 1/10 the effort and time to do something and just coast, never bother to even understand what they wrote.

just encountered that with some interviews where people can now scaffold something up in record time but can't be bothered to refine it because they don't know how. (ex. you have someone prompting to create some component and it does it in a minute. if you request a tweak, because they don't understand it they just keep trying re-prompt and micromanage the LLM to get the right output when it should only take another minute for someone experienced.)

Strong agree. There have been people who blindly copied answers from Stack Overflow without understanding the code, but most of us took the time to read the explanations that accompanied the answers.

While you can ask the AI to give you additional explanations, these explanations might be hallucinations and no one will tell you. On SO other people can point out that an answer or a comment is wrong.

Are you sure? Because chatbots are used exactly like SO, with the exception of usecases like a) using as glorified template and refactoring engines, b) explain existing codebases, c) automatically add unit tests and context-based documentation. Just open SO and check any random question: how to get a framework to do something, how to configure something, why something worked like this or like that, why did a code block had a bug, etc.

This is it. You will have real developers, as you do today, and developers who are only capable of creating what the latest AI model is capable of creating. They’re just a meat interface for the AI.

Agree. Ai answers actually work and are less out of date

Jokes aside, I feel this is really the case. Chatbots have the exact same usecases, but as their response time is measured in seconds instead of days and the responses reflects the context from that moment, this opens the door for more interactive experience and more iterations over the same output.

I feel this anger towards LLMs comes from gatekeepers who felt slighted by having to spend so much effort to reach positions where people with chatbots take instances, and as they can no longer close gates or pull ladders they start to bitch and whine against LLMs. It's as simple as that.

And it kind of was a problem. There was an underclass of people who simply could not get anything done if it wasn't already done for them in a Stackoverflow answer or a blog post or (more recently and bafflingly) a Youtube video. These people never learned to read a manual and spent their days flailing around wondering why programming seemed so hard for them.

Now with AI there will be more of these people, and they will make it farther in their careers before their lack of ability will be noticeable by hiring managers.

As a second-order effect, I think there's a decline in expected docs quality (of course depends on the area). Libraries and such don't expect people to read through them, so they are spotty and haphazard, with only some random things mentioned. No wider overviews and explanations, and somewhat rightly so, why try to write it if (nearly) no one will read it. So only tutorials and Q&A sites remain besides of API dumps.

Which is a great opportunity btw to drive forward a transition to a post-monetary, non-commercial post-scarcity open-source open-access commons economy.

No, the problem is created and perpetuated by people defaulting to building communities around Discord servers or closed WhatsApp groups. Those are true information black holes.

And yes, privacy and commons are in opposition. There is a balance to be had. IMHO, in the past few years, we've overcorrected way too much in the privacy direction.

So far, AI cannot do that. But it can pretend to, very convincingly.

So now we can get wrong code but written in a more confident language.

If you are lucky it’s a hallucination and the error is obvious.

The lesson to learn is rather that also in "real life", you shouldn't trust confident people.

from what i gather, the training data often contains those same poorly written docs and often a lot of poorly written public code examples, so… YMMV with this statement as it is often fruit from the same tree.

(to me LLMs are just a different interface to the same data, with some additional issues thrown in, which is why i don’t care for them).

> I think we have a lot of old timers ramping up on their version of "I walked 10 miles to school uphill both ways". Not a good look. We old timers need to do better.

it’s a question of trust for me. with great power (new tools) comes great responsibility — and juniors ain’t always learned enough about being responsible yet.

i had a guy i was doing arma game dev with recently. he would use chatgpt and i would always warn him about not blindly trusting the output. he knew it, but i would remind him anyway. several of his early PRs had obvious issues that were just chatgpt not understanding the code at all. i’d point them out at review, he’d fix them and beat himself up for it (and i’d explain to him it’s fine don’t beat yourself up, remember next time blah blah).

he was honest about it. he and i were both aware he was very new to coding. he wanted to learn. he wanted to be a “coder”. he learned to mostly use chatgpt as an expensive interface for the arma3 docs site. that kind of person using the tools i have no problem with. he was honest and upfront about it, but also wanted to learn the craft.

conversely, i had a guy in a coffee shop recently claim to want to learn how to be a dev. but after an hour of talking with him it became increasingly clear he wanted me to write everything for him.

that kind of short sighted/short term gain dishonesty seems to be the new-age copy/pasting answers from SO. i do not trust coffee shop guy. i would not trust any PR from him until he demonstrates that he can be trusted (if we were working together, which we won’t be).

so, i get your point about doom and gloom naysaying. but there’s a reason for the naysaying from my perspective. and it comes down whether i can trust individuals to be honest about their work and how they got there and being willing to learn, or whether they just want to skip to end.

essentially, it’s the same copy/pasting directly from SO problem that came before (and we’re all guilty of).

It’s not just bad optics; it’s destructive. It discourages folks from learning.

AI is just another tool. There’s folks that sneer at you if you use an IDE, a GUI, a WYSIWYG editor, or a symbolic debugger.

They aren’t always boomers, either. As a high school dropout, with a GED, I’ve been looking up noses, my entire life. Often, from folks much younger than me.

It’s really about basic human personal insecurity, and we all have that, to some degree. Getting around it, is a big part of growing up, so a lot of older folks are actually a lot less likely to pull that crap than you might think.

I think the same tendency of some programmers to just script kiddie their way out of problems using SO answers without understanding the issue will be exacerbated by the proliferation of AI which is much more convincing about wrong answers.

It's not a binary. You don't have to hate or welcome chatbots, no in between. We all use them, but we all also worry about the negatives, same with SO.

https://en.m.wikipedia.org/wiki/Google_Books

As other AI companies argue, copyright doesn't apply when training, it should give Google a huge advantage to be able to use all the worlds books they scanned.

Of course AI is incredibly useful both for reading foreign language forums and for analysing complex code bases for original research. AI is great for supercharging traditional research

1) There are no idiots who want to do better than AI, and/or

2) All idiots are lazy idiots.

The reason we're even discussing LLMs so much in the first place, is AI can and does things better than "idiots"; hell, it does things better than most people, period. Not everything in every context, but a lot of things in a lot of contexts.

Like run-of-the-mill short-form writing, for example. Letters, notices, copywriting, etc. And yes, it even codes better than general population.

An LLM isn't a mind reader so if you never learn how to seek the answers you're looking for, never curious enough to dig deeper, how would you ever break through the first wall you hit?

In that way the LLM is no different than searching Google back when it was good, or even going to a library.

We never got to the question of recursive or iterative methods.

The most worrying thing is that the LLM were not very useful three years ago when he started university. So the situation is not going to improve.

The reason we ask people to do fizzbuzz is often just to weed out the shocking number of people who cannot code at all.

When the client knows absolutely nothing about it and is not supported by someone competent, they end up employing just anyone.

This applies to both construction, IT and probably everthing.