Remix.run Logo
StilesCrisis a day ago

The babysitting work would still be impossible if you didn't actually know how to code.

irjustin a day ago | parent | next [-]

For now, I can imagine a not too distant future where this is largely untrue.

LLMs are an abstraction just like machine code -> assembly -> C/JVM -> some lang -> LLMs?

At some point you stopped needing to understand the layer down because the layer you were on became so good. Yes there are always corner cases, but for the vast majority of developers/engineers out there, staying at your layer was enough to make a career out of it once your layer hit a certain maturity.

palmotea a day ago | parent | next [-]

> LLMs are an abstraction just like machine code -> assembly -> C/JVM -> some lang -> LLMs?

People keep trying to make that analogy, but it doesn't really work because LLMs aren't deterministic like compilers and assemblers.

maccard a day ago | parent | next [-]

They don’t need to be dererministic, only reproducible.

They’re not reproducible, nor are they even reliable right now.

bloppe 4 hours ago | parent [-]

I don't think there's a meaningful difference between "deterministic" and "reproducible" the way you just used them. How can something be reproducible if you can't accurately predict what it will do?

Regardless, you can make them deterministic by turning the temperature down to zero. Just nobody likes doing that for whatever reason. I guess it ruins some sort of illusion people seem to like.

Izkata 14 hours ago | parent | prev | next [-]

Also to be a compiler-type next-level-up abstraction we'd have to be at the point where we commit the prompts and throw away the code.

(Which is pretty much what determinism would get us, but in these conversations way too many people seem not to understand what determinism is, so describing it in terms of actions the developer takes might work better?)

jonfw 20 hours ago | parent | prev [-]

It is not obvious to me that determinism is a requirement for an abstraction

Toutouxc 19 hours ago | parent | next [-]

IMO delegation isn't abstraction.

I work at a certain level, like Ruby code. That's what I write, debug and maintain. I don't really care about the internals of the interpreter or about the source code of Linux, because these layers are taken care of, they're reliable and they're being developed by competent people. I [think I] know what Ruby code should look like in order to remain [reasonably] fast, maintainable and reliable, and it's my job to build a product out of that kind of code. If I keep writing code like that, I know for sure that I'll be able to keep building the product, because the layers underneath are deterministic. It's like the certificate chain of trust, but with "surely these people are not idiots". And that's simply not the case with LLMs.

Not only can the LLM be a massive idiot, but also an unpredictable one. I can try to warn it, steer it, police it or review as much as you ask me to, but ultimately you're asking me to delegate my job and my responsibilities to an intermediate whose reasoning I don't understand, who has no loyalty, no sense of pride, no sense of ethics, can't be taught and can't be fired.

palmotea 19 hours ago | parent | prev [-]

> It is not obvious to me that determinism is a requirement for an abstraction

But it is a requirement for them to be an "abstraction just like" assemblers or compilers, which is what was being claimed.

bloppe a day ago | parent | prev | next [-]

If this is true, then at some point we will stop committing source code to repositories, and instead commit prompts. The LLM would be a "compiler" that can reliably turn a vast collection of natural-language prompts into a complete system.

I just don't see that, not least of all because natural language is inherently ambiguous, whereas all the other rungs in your latter ("machine code -> assembly -> C/JVM -> some lang") are completely unambiguous by design. Consider "I saw the man with the binoculars". Does that mean "I used binoculars to look at the man", or "The man I looked at was holding binoculars"? This is the kind of inherent ambiguity that Lojban was invented to mitigate. Maybe some day we'll write "natural" language prompts in Lojban that can be unambiguously translated by an LLM, but that sounds a lot like just using a "some lang".

weakfish 21 hours ago | parent | prev | next [-]

IMO English as a programming language is a fools errand. Part of the reason why we _have_ programming languages is because it turns out human language is too ambiguous.

alightsoul a day ago | parent | prev | next [-]

There's an increasing amount of jobs where the job role is analyst but you're just feeding the ai with whatever the task assigned to you was. These are not software jobs but are business jobs. Like User acceptance testing quality assurance aka bureaucracy

kbelder 9 hours ago | parent | prev | next [-]

But at every point, your product was better if you still understood the lower level. Your Python is better if you understand Assembly.

skydhash a day ago | parent | prev | next [-]

> LLMs are an abstraction just like machine code -> assembly -> C/JVM -> some lang -> LLMs

The what is the semantic mapping between <some lang> and LLMs?

I know the semantic mapping between maching code and assembly (some light weight syntax manipulation and macros). I know the one between assembly and C (the C abstract machine, which is mostly about the stack and whatever call/ret instructions pair). I know the one between C and something like python (not so much different than the one between C and assembly in mechanism).

Please talk about how you go from A LLM prompt to a piece of code in Python and guarantee the intent remains unchanged.

tibbar a day ago | parent [-]

The non-determinism is one of the relevant features of this layer of abstraction! And one can learn to validate that the translation is being done properly. Some of the tools you have include writing extremely detailed specs, generating visualizations of the internals of the tool, or (perhaps) reading the code, though that becomes less feasible with volume.

Basically it turns out that code is full of incidental details and what you really want is to verify the important parts, while receiving a guarantee that the vast tail of incidentals is handled "reasonably."

lenkite a day ago | parent | next [-]

Writing Specs that are thousands of lines wrong without formalism, type and correctness checking is a major PITA. Wish we had a..compiler that can check our code..err I mean spec.

duskdozer a day ago | parent [-]

Oh, are you still handwriting your specs? How Q2 2026. Handcrafted artisan specs are obsolete and you will be left behind in the dust if you still insist on them.

weakfish 20 hours ago | parent [-]

Poe’s law displayed beautifully here

Izkata 11 hours ago | parent [-]

In which direction? It wasn't long ago having the agent generate a plan before acting on it was a "best practice".

bcrosby95 a day ago | parent | prev | next [-]

Writing specs validates the translation is being done properly? Was this written by an LLM?

tibbar 3 hours ago | parent [-]

Writing detailed specs is an important tool to help the LLM translate the idea in your head to code. The spec is an intermediate artifact.

bitwize a day ago | parent | prev [-]

LLMs will still blithely ignore the specs and steering documents, apologize profusely for doing so after the fact, and tell you "I'll do better next time" which they might do once or twice. But after the context is cleared or a new session opened, the Dixie Flatline gets reset and it doesn't remember it screwed up, or that you told it not to.

This happened to a coworker of mine. Generally the response from one-shotted devs is a shrug of the shoulders and "wellp, them's the breaks! As long as it looks sensible from 10,000 feet up it's still a huge productivity win." But the devil, as they say, is in the details.

glouwbug a day ago | parent | prev | next [-]

They aren't because we're not committing prompts. The analogous would be compiling to and then committing unmaintainable assembly when the higher level language, which is a deterministic compiler, already exists

chii a day ago | parent [-]

> not committing prompts.

nothing prevents people from committing their prompts. I've started seeing prompts being committed into repos, or at least as part of the commit message.

In any case, if in the future there's a prompt specific language, it would be committed. I dont think we've reached there yet, but i dont doubt this is on the path to the future.

Toutouxc 16 hours ago | parent | next [-]

> if in the future there's a prompt specific language, it would be committed. I dont think we've reached there yet, but i dont doubt this is on the path to the future

You mean, like a ... programming language? Honestly I can't tell these days what is satire and what isn't.

alightsoul a day ago | parent | prev [-]

I posted a while ago about connecting an LLM directly to an assembler and people thought it was a bad idea. Now we're going in exactly that direction

archagon 14 hours ago | parent | prev | next [-]

LLMs are as much an "abstraction" as outsourcing your app to a bunch of contractors: a silly dilution of the word to the point of meaninglessness.

woodpanel a day ago | parent | prev [-]

There is a always going to be the gap of what you can do with LLMs if you know how to code vs if you don’t

lowbloodsugar a day ago | parent [-]

This is only true if it plateaus. What you are saying is “I don’t think LLMs will achieve superintelligence.” Which is a fine opinion to have, but it’s an opinion.

woodpanel a day ago | parent [-]

once, if ever, the plateauing happens. Until then there is going to be this gap.

In other words, superintelligence often referred to as AGI might either be months away or just VC-Money induced cult-speak many fall victim to.

It doesn’t matter because the only certainty is that it’s not here now, and neither tomorrow etc.

sfn42 a day ago | parent | next [-]

To me it looks like they've been pretty plateaued for a good while already? Sure they do marginally better on benchmarks or whatever, but to me as a user there's really not much difference between chatgpt today and chatgpt a couple years ago. Main difference is just added capabilities like web search, image recognition etc.

For coding you still need to be in control if you want a good result.

simg a day ago | parent | prev [-]

AGI is human level intelligence. superintelligence is ASI

DrewADesign a day ago | parent | prev | next [-]

Right, but if things progress as they promised, it will require far fewer people to do the same work, which means the industry already has all the people it so need for at least a couple decades. That’s what happened with tool and die guys when offshoring kicked off, and now they’re scrambling to get apprentices because the last of the OG ones are retiring. There were decades that pretty much nobody (relative to prior eras) got trained for that work.

drTobiasFunke a day ago | parent | next [-]

You are assuming that the amount of work is finite, so more productive people implies fewer people are required. Is that the case though? Has it ever been?

DrewADesign 16 hours ago | parent | next [-]

The amount of work in any given market at any given time is finite. Beyond that, developers likely won’t even be the bottleneck if everyone is a 10x ultradev 3000.

> Has it ever been?

Well… yes? So very many industries shrank, even disappeared in practical terms, because efficiency, automation and technological improvements. Industrial revolution? Calculators? Computerized accounting? I mean the list is giant.

david-gpu a day ago | parent | prev | next [-]

> You are assuming that the amount of work is finite, so more productive people implies fewer people are required. Is that the case though? Has it ever been?

How many horse farriers have you met? How many coopers, blacksmiths, or shoemakers?

drTobiasFunke a day ago | parent | next [-]

Not a fair comparison. Unlike shoemaker or coopers who do ONE thing, a SWE doesnt do just one thing.. atleast in general. I churned through basic, logo, c, c++, java, c#, python, go and now agents. The executable building on single machine died when distributed systems came in. It again evolved when cloud took over. We have been reinventing our work every few years. Every job change requires us to learn new skills.

How much did a horse farrier have to learn if they switched their employers?

DrewADesign a day ago | parent [-]

No profession uses only one skill. I’ve had a few, the longest-running as a developer, and there is nothing uniquely varied about working in software. One thing that is pretty unique to software is how superior they feel to most other professionals, when in reality, it’s a pretty mediocre discipline compared to electrical or mechanical engineering, medicine, chemistry…

Cobblers design, make and repair shoes of various kinds, boots for various purposes, slippers and moccasins with leather, cloth, rubber, and many kinds of threads using punches, knives, various machines, glues…

How much does a developer need to learn about their core competency when switching employers? Not even close to as much as a machinist. It’s not a useful comparison.

alightsoul a day ago | parent | prev [-]

They absolutely still exist just not in the US. Bad comparison IHMO. What they make is just imported into the US and they often work in factories in low wage countries like Bangladesh, Pakistan, Cambodia. They were just offshored

david-gpu a day ago | parent [-]

> They absolutely still exist

That is precisely the point. They still exist, but it is a far less common occupation than it once was.

IshKebab a day ago | parent | prev [-]

The amount of work is not a fixed number, it's more of a supply/demand curve. Do we need 100 programs? Yes definitely! 10k? Yes. 1m? Probably. 100m? Uhm... 1bn? Probably not.

So there will always be a point where people aren't willing to hire more software developers because there are enough already.

DrewADesign 7 hours ago | parent [-]

Considering the dev job market right now, lack of developers doesn’t seem to be a bottleneck for the industry.

patch_dev a day ago | parent | prev [-]

You're making an assumption on what AI lab have promised? The same AI labs that have made a strategy out of lying for hype? Thats crazy man, regardless of whether anything else you're saying is valid.

DrewADesign a day ago | parent [-]

That’s why I said “if” and not “when.”

mellosouls a day ago | parent | prev | next [-]

Consider the difference between capabilities when gpt3 was released and now - the "increasingly babysitting" is exactly right.

Knowing how to code (and more generally software engineering and other roles in software teams) is definitely still extremely useful, but is rapidly becoming less vital as a human-provided skill as models and harnesses greedily hoover up the knowledge margin.

left-struck a day ago | parent | prev | next [-]

I think there’s potentially a future where software engineers learn how to “babysit” models instead of the details of programming. Kind of how software engineering students for the last decade at least haven’t learnt a great deal of assembly or cpu architecture. Maybe you had a unit on CPUs but it’s not central to the course.

I’m not saying I like that future, but I can imagine it.

bloppe a day ago | parent [-]

Imagine managing a team of people with no knowledge of their craft, like a team of doctors when you know nothing about being a doctor. You might be pretty confident that you can do it really well. The doctors would probably think that your "management" is unnecessary and potentially counter-productive.

JumpCrisscross a day ago | parent | prev | next [-]

> babysitting work would still be impossible if you didn't actually know how to code

It will be fewer and fewer people with, probably, deeper and deeper knowledge (and job security and compensation to boot).

Poet is a bad comparison. But something like low-level semiconductor physics or assembly is closer to the mark.

stronglikedan a day ago | parent | prev | next [-]

I fear the capability of the models will quickly outpace the need for a human to validate their output. They won't be juniors for much longer.

danielbln 14 hours ago | parent [-]

Agreed, you can feel this change strongly with Opus vs Fable. Fable doesn't feel like a junior developer anymore. And it requires a lot less ceremony to boot.

deeprack4sure a day ago | parent | prev [-]

And thats why noone should learn to code.