Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

At Amazon, some coders say their jobs have begun to resemble warehouse work

nytimes.com

Post reply on HN
Print view View original

701–710 of 897 posts

Re: At Amazon, some coders say their jobs have begun to resemble warehouse work

#701

Post by pitaj »

lightedman wrote: post #404
intended wrote: post #387

Earlier quoted context omitted.

The fact that they used the auto industry as an example is funny, because the Toyota way and six sigma came out of that industry. So you are telling me, your AI code passes a six sigma grade of quality control? I have a bridge to sell you. No, Bridges!

"The fact that they used the auto industry as an example is funny, because the Toyota way and six sigma came out of that industry." It's even funnier when you consider that Toyota has learned how bad of an idea lean manufacturing/6-Sig/5S can be thanks to the pandemic - they're moving away from it in some degrees, now.

JIT is totally independent from 6-sig.

Re: At Amazon, some coders say their jobs have begun to resemble warehouse work

#702

Post by skeeter2020 »

33hsiidhkl wrote: post #683
close04 wrote: post #660

Earlier quoted context omitted.

I don't know what you mean with "the code it creates is deterministic" but the process an LLM uses to generate code based on an input is definitely not entirely deterministic. To put it simply, the chances that an LLM will output the same result every time given the same input is low. The LLM does not operate deterministically, unlike the manufacturing robot who will output the same door panel every single time. Or a…

For any given seed value, the output of an LLM will be identical- it is deterministic. You can try this at home with Llama.cpp by specifying a seed value when you load a LLM, and then seeing that for a given input the output will always be the same. Of course there may be some exceptions (cosmic ray bit flips). Also, if you are only using online models, you can't set the seed value, plus there are multiple models, so…

I feel this is technically correct but intentionally cheating. no one - including the model creators - expects that to be the interface; it undermines they entire value proposition of using an LLM in the first place if I need to engineer the inputs to ensure reproducability. I'd love to hear some real world scenarios that do this where it wouldn't be simpler to NOT use AI.

Re: At Amazon, some coders say their jobs have begun to resemble warehouse work

#703

Post by awinter-py »

ah yes, how the kiva robot lifts a mobile shelf full of Codes and brings them to me for my inspection. I use my left hand to reach for the Code I need and place it in the tote, because my right shoulder is injured from repeating this motion.

The tote goes onto a conveyor where it is shot forwards, backwards and sideways, all around the warehouse until it reaches the compiler station. There, another worker will lint each Code with a lint roller and compile them by selecting the appropriately sized cardboard box.

Next, this executable bundle will be linked with its customer, possibly by means of a plane or 18-wheeler truck, with last-mile linking performed by a smaller vehicle suitable for urban traffic ...

Re: At Amazon, some coders say their jobs have begun to resemble warehouse work

#704

Post by skeeter2020 »

d0gsg0w00f wrote: post #588
lolinder wrote: post #360

Earlier quoted context omitted.

Also, you know who did measure every angle to make sure it was correct? The engineers who put together the initial design. They sure as hell took their time getting every detail of the design right before it ever made it to the assembly line. Who's filling that role in this brave new world?

This is an underrated comment. Who's job is it to do the thinking? I suppose it's still the software engineer, which means the job comes down to "code prompt engineer" and "test prompt engineer".

Wild times where a task that used to be described as "good at using google" now gets the title of "Engineer". It was bonkers enough when software devs co-opted the title.

Re: At Amazon, some coders say their jobs have begun to resemble warehouse work

#705

Post by skeeter2020 »

noosphr wrote: post #573
b112 wrote: post #566

Earlier quoted context omitted.

This also seems, to me, like composer/npm issues. An entire generation of devs, who grew up using unaudited, unverified, unknown license code. And which at a moments notice, can be sold to a threat actor. And I've seen devs try to add packages to the project without even considering the source. Using forks of forks of forks, without considering the root project. Or examing if it's just a private fork, or what is most…

After putting off learning JS for a decade, I finally bit the bullet since I can talk to an LLM about it while going through the slog of getting a mental model up and running. After a month, I can say that the inmates run that whole ecosystem, from the language spec, to the interpreter, to packaging. And worse, the tools for everyone else have to cater to them. I can see why someone who has never had a stable foundat…

since you mention JS specifically, I think it's important to seperate that from the framework ecosystem. I'd suspect that most LLMs don't which is part of the problem. I had a similar experience with Python lately, where the LLM-generated code (once I could get it to run) resulted in code that I would generously evaluate as "Excel VBA Macro quality". It does the task - for now - but I didn't learn much about what production-grade python would look like.

Re: At Amazon, some coders say their jobs have begun to resemble warehouse work

#706

Post by skeeter2020 »

XenophileJKO wrote: post #600
lukan wrote: post #510

Earlier quoted context omitted.

In a world of vibe coders, those who can still debug on their own, will have quite a valuable skill.

Are least for a few more generations of model. I just finished creating a multiplayer online party game using only Claude Code. I didn't edit a single line. However, there is no way someone who doesn't know how to code could get where I am with it. You have to have an intuition on the sources of a problem. You need to be able to at least glance at the correct and understand when and where the AI is flailing, so you k…

so if you're an experienced, trained developer you can now add AI as a tool to your skill set? This seems reasonable, but is also a fundamentally different statement that what every. single. executive. is parroting to the echochamber.

Re: At Amazon, some coders say their jobs have begun to resemble warehouse work

#707

Post by icedogma »

I worked at Amazon Robotics low level software team, as a contractor, for a few months. This article sounds about right. Crazy pace of work with no room for reflection or — especially — groundbreaking invention, is given. And that was a year ago, when AI was totally optional.

The engineers are smart and fast folks, which helps a great deal. But don’t seek true creativity and deep expertise at Amazon, unless we’re talking about legacy L6+ (or likely even L7+) level engineers who get to have the leeway needed. The picker robots I had been working with were created, originally, by a Netherlands company, not Amazon. Amazon’s job was scaling it all up for their massive warehouse operation. It was a bunch of hacks on top of hacks. Heck, not even Harel state machines (which is kind of what one would expect to have on a properly designed event-driven robot that’s not overly tied to its PLC implementation, where things can be modularized, in and out, as needed). Just a bunch of switch statements and a menagerie of globals, with only two key people intimately familiar with most of its key code structure, one of whom decided to leave the group in frustration.

In fairness, the mess didn’t originate just at Amazon, but there never was a technical push at Amazon to have the initial work done in the right way. Not the software part anyway.

I can see that the craziness not only didn’t abate but has gotten worse. The money was good but it’s not worth the stress and other frustrations IMO, unless you’re a noncontractor and at least a technical program manager or above, where you won’t be coding as much as reviewing others’ code anyway.

Otherwise, I can totally understand why people just work there for a couple years, to get their resume looking good and then leave

Re: At Amazon, some coders say their jobs have begun to resemble warehouse work

#708

Post by close04 »

33hsiidhkl wrote: post #683
close04 wrote: post #660

Earlier quoted context omitted.

I don't know what you mean with "the code it creates is deterministic" but the process an LLM uses to generate code based on an input is definitely not entirely deterministic. To put it simply, the chances that an LLM will output the same result every time given the same input is low. The LLM does not operate deterministically, unlike the manufacturing robot who will output the same door panel every single time. Or a…

For any given seed value, the output of an LLM will be identical- it is deterministic. You can try this at home with Llama.cpp by specifying a seed value when you load a LLM, and then seeing that for a given input the output will always be the same. Of course there may be some exceptions (cosmic ray bit flips). Also, if you are only using online models, you can't set the seed value, plus there are multiple models, so…

> the process an LLM uses to generate code based on an input is definitely not entirely deterministic

Technically correct is the least useful kind of correct when it's wrong in practice. And in practice the process AI coding tools use to generate code is not deterministic which is what matters. To make matters worse in the comparison with a manufacturing robot, even the input is never the same. While a robot get the exact command for a specific motion and the exact same piece of sheet metal, in the same position, a coding AI is asked to work with varied inputs and on varied pieces of code.

Even stamping metal could be called "non-deterministic" since there are guaranteed variations, just within determined tolerances. Does anyone define tolerances for generated code?

That's why the comparison shows a lack of understanding of either of the systems.

Re: At Amazon, some coders say their jobs have begun to resemble warehouse work

#709

Post by dbolgheroni »

rwmj wrote: post #684
JKCalhoun wrote: post #668

Earlier quoted context omitted.

My own tangential gripe (a bit related to yours though): the factory work began when Agile crept into the workplace. Additionally, lint, unit tests, code reviews... all this crap just piled on making programming worse still. It stopped being fun to code around that point. Too many i's to dot to make management happy.

Agile yes, it's micromanagement at scale. But writing tests and doing code reviews is good practice.

Code review is another sacred process that seems too good not to have, but many teams use it as a "we care about quality" stamp when in fact they do not. Used for just nitpicking code style (important but not the whole reason to have CR, and there are tools for this), issue comments like "LGTM" and approve whatever arrives at the pull request anyway.

Re: At Amazon, some coders say their jobs have begun to resemble warehouse work

#710

Post by guax »

mrheosuper wrote: post #548
stock_toaster wrote: post #218

Earlier quoted context omitted.

I wonder about codebase maintainability over time. I hypothesize that it takes some period of time for vibe-coding to slowly "bit rot" a complex codebase with abstractions and subtle bugs, slowly making it less robust and more difficult to maintain, and more difficult to add new features/functionality. So while companies may be seeing what appears to be increases in output _now_, they may be missing the increased dra…

I wonder whether we have the same talk when the C compiler first came out. People may worry that the "ASM" codebase will be bit-rot and no one can understand the compiler output or add new feature to the ASM codebase.

My guess is that the discussion trended around performance and not correctness since compilers are pretty well understood. Why a LLM output what they do are not understood by anyone to the same degree.
Post reply on HN