All the devs saying that Anthropic’s code quality is “normal” are telling on themselves and everybody they’ve worked with
-
All the devs saying that Anthropic’s code quality is “normal” are telling on themselves and everybody they’ve worked with
(Also supports what many have been saying about software quality being a crisis that precedes LLMs, but that’s another story)
@baldur Remember when Elon declared he was going to use lines of code produced on printouts as something he wanted to judge Twitter employees on when deciding who should and shouldn't be fired? It's that mindset made into a machine.
-
@baldur the software crisis is definitely what enabled vibe coding, I feel.
When making software was an artisan process, it was hard for corporations to scale it and treat programmers like cogs in a machine.
So they've been trying more and more to fit software development into a neat mold, essentially dumbing down the process, to the point where making an app has become "just slap together some libraries and hey presto".
That has 100% laid the foundation for LLM-generated code, I feel.
@Tijn Yup. I agree
-
All the devs saying that Anthropic’s code quality is “normal” are telling on themselves and everybody they’ve worked with
(Also supports what many have been saying about software quality being a crisis that precedes LLMs, but that’s another story)
@baldur I ended up in a couple of places like that and never lasted long, because I was told, to my face and quite seriously, that a) my standards were too high and b) that I care too much about my values. They meant it as a negative, which I will never, ever understand. (all of them advertised themselves as "values-based" organizations. fuck startups.
) -
@baldur the software crisis is definitely what enabled vibe coding, I feel.
When making software was an artisan process, it was hard for corporations to scale it and treat programmers like cogs in a machine.
So they've been trying more and more to fit software development into a neat mold, essentially dumbing down the process, to the point where making an app has become "just slap together some libraries and hey presto".
That has 100% laid the foundation for LLM-generated code, I feel.
-
R relay@relay.infosec.exchange shared this topicR relay@relay.publicsquare.global shared this topic
-
@raffaella @Tijn @baldur anyone *can* code, given the right training. But not everyone *wants* to code. And the people who do it, only because they want a paycheck, and not because they want to code, were the start of this decline.
-
@raffaella @Tijn @baldur anyone *can* code, given the right training. But not everyone *wants* to code. And the people who do it, only because they want a paycheck, and not because they want to code, were the start of this decline.
-
@baldur the software crisis is definitely what enabled vibe coding, I feel.
When making software was an artisan process, it was hard for corporations to scale it and treat programmers like cogs in a machine.
So they've been trying more and more to fit software development into a neat mold, essentially dumbing down the process, to the point where making an app has become "just slap together some libraries and hey presto".
That has 100% laid the foundation for LLM-generated code, I feel.
-
All the devs saying that Anthropic’s code quality is “normal” are telling on themselves and everybody they’ve worked with
(Also supports what many have been saying about software quality being a crisis that precedes LLMs, but that’s another story)
@baldur Like I always said, closed-source code is closed because they have something to hide. Mostly, how atrociously bad it is.
-
All the devs saying that Anthropic’s code quality is “normal” are telling on themselves and everybody they’ve worked with
(Also supports what many have been saying about software quality being a crisis that precedes LLMs, but that’s another story)
@baldur Is it possible your opinion about their internal code quality just isn't the constraint that you think it is? After all it's only the scaffolding around the LLM that does the actual work.
Interesting article on them from a couple of weeks ago telling their basing on $21bn of Google's TPUs (which run Gemini for Google) going forward rather than GPUs.
-
@dalias @baldur Yeah this is fair imho!
What I was mainly trying to convey was a sense of caring about what is you're building as a developer.
That's the thing that's lacking imho from corporate coding practices. The workers don't have to care about the software, they don't have to understand the whole thing, they just need to solve the ticket and move on.
This mindset fundamentally undermines the quality of any software project. And also perfectly lays the foundation for LLM-generated code.
-
@baldur Is it possible your opinion about their internal code quality just isn't the constraint that you think it is? After all it's only the scaffolding around the LLM that does the actual work.
Interesting article on them from a couple of weeks ago telling their basing on $21bn of Google's TPUs (which run Gemini for Google) going forward rather than GPUs.
-
-
-
-
@baldur I ended up in a couple of places like that and never lasted long, because I was told, to my face and quite seriously, that a) my standards were too high and b) that I care too much about my values. They meant it as a negative, which I will never, ever understand. (all of them advertised themselves as "values-based" organizations. fuck startups.
) -
@dalias @baldur Yeah this is fair imho!
What I was mainly trying to convey was a sense of caring about what is you're building as a developer.
That's the thing that's lacking imho from corporate coding practices. The workers don't have to care about the software, they don't have to understand the whole thing, they just need to solve the ticket and move on.
This mindset fundamentally undermines the quality of any software project. And also perfectly lays the foundation for LLM-generated code.
-
All the devs saying that Anthropic’s code quality is “normal” are telling on themselves and everybody they’ve worked with
(Also supports what many have been saying about software quality being a crisis that precedes LLMs, but that’s another story)
I mean, we've seen similar issues with code quality from certain types of outsourcing setups where companies send out under-thought and over-written tickets to an outsourcing mill and gets back knee-jerk code that appears as if it's been written at haste while wearing blinkers in a snow-storm, and without any traces of "second-layer thought." (_Not_ the fault of the individual programmers who are often very bright, mind. It's a system failure.)
-
All the devs saying that Anthropic’s code quality is “normal” are telling on themselves and everybody they’ve worked with
(Also supports what many have been saying about software quality being a crisis that precedes LLMs, but that’s another story)
@baldur@toot.cafe
I'm just glad the leak also showed LLMs are just nested regex. -
All the devs saying that Anthropic’s code quality is “normal” are telling on themselves and everybody they’ve worked with
(Also supports what many have been saying about software quality being a crisis that precedes LLMs, but that’s another story)
@baldur
Sometimes I wonder if the practice of shipping every GUI-era computer with a development environment (getting the "rubes" invested in programming) would've prevented the current situation. -
