2026年8月7日 · 星期五
● 每日更新·改变自己
Eurekar·TOP
捕捉真实世界的英语信号
25信息来源
10,499精选文章
168单词卡片
18照片图片
全部7,114口语2,068免费777帖子3,025新闻1,599hackernews797tmz616techmeme532slashdot379随笔323techcrunch295外刊291arstechnica288Cards168simonwillison100动态69bloomberg59sethgodin50图片18youtube7

/dpt/ - Daily Programming Thread

/dpt/ - 每日编程串

4chan /g/ · Anonymous · 2026-08-02 09:51 · 311 帖 · 原文 ↗

← 上一篇返回列表下一篇 →
#1 /dpt/ - Daily Programming Thread / /dpt/ - 每日编程串
Anonymous · 2026-08-02 09:51

prev: >>109414529

上一条:>>109414529

图片: https://i.4cdn.org/g/1785664313070962.png

#2 No.109435690
Anonymous · 2026-08-02 10:27

>not asking us what we're working on

>不问问我们在搞什么 ┋ 搞不成了

not gonna make it

成不了啊

#3 No.109435711
Anonymous · 2026-08-02 10:31

>>109435690

we already know the answer lil bro

咱都知道答案了老弟

#4 No.109435719
Anonymous · 2026-08-02 10:33

Tell me then, autist.

那你说啊,自闭儿。

What am I actually working on?

我到底在搞啥?

>no, it's not nothing

>不,不是什么都没有

#5 No.109435764
Anonymous · 2026-08-02 10:41

>>109435719

show us your portfolio and we will know what you're "working" on lil bro

把你的作品集亮出来,咱就知道你“搞”的是啥了老弟

#6 No.109435809
Anonymous · 2026-08-02 10:50

>>109435690

me dumb ogre.

我是傻大个食人魔。

me eat my poos.

我吃我自己的屎。

nom nom.

嚼嚼嚼。

me try read technical document and make giga softwares.

我想读技术文档然后做超级软件。

#7 No.109435851
Anonymous · 2026-08-02 11:00

what are you working on, /dpt/?

你在搞什么,/dpt/?

#8 No.109435855
Anonymous · 2026-08-02 11:01

>it needs information

>它需要信息

>so that it can repeat said information

>这样才能重复那些信息

I swear to God this is exactly why humanity would've been better off literally slaughtering autistic scum like Newton. We already would be at Mars if we had tortured his entire fucking ilk four hundred years ago.

我发誓这他妈就是为什么人类当初还不如把牛顿那种自闭垃圾全宰了。四百年前要是折磨死他那整个同类,我们早就上火星了。

#9 No.109436028
Anonymous · 2026-08-02 11:35

>>109435518

Right in this moment I am not coding.

此刻我没在写代码。

The weekend if for relaxation. I think I almost stop sweating now, so soon (tm) I can go shower.

周末是用来放松的。我觉得我几乎不出汗了,所以很快(tm)就能去洗澡了。

So I can't tell you right now what I will be coding today, I am sorry

所以我这会儿没法告诉你我今天要写什么代码,抱歉。

图片: https://i.4cdn.org/g/1785670515359721.png

#10 No.109436094
Anonymous · 2026-08-02 11:46

>and yet they claim Germans are not autistic by definition

>然而他们却声称德国人按定义不是自闭的

#11 No.109436129
Anonymous · 2026-08-02 11:52

>>109435855

You could go to mars today if your portfolio wasn't just your last month's neetbux and $50 from mommy.

你要真能去火星,那你的作品集就不会只是上个月的福利金加妈妈给的50美元了。

#12 No.109436140
Anonymous · 2026-08-02 11:53

>>109436094

Ok no problem for me. I just finished the shower, had a snack and now I have a Club Mate (TM). Life's Good ;-) (LG) (TM)

行,我无所谓。我刚洗完澡,吃了点零食,现在来了一瓶Club Mate(TM)。生活真好 ;-)(LG)(TM)

#13 No.109436145
Anonymous · 2026-08-02 11:54

>>You could go to mars today

>>你能去火星

>this is what autists actually, factually believe

>这就是自闭症患者实际认为的事实

Holy.

#14 No.109436150
Anonymous · 2026-08-02 11:54

And don't hurry coming back from mars, lil bro, stay there, fizzbuzzing your registry dumbing nocodeshittery there until you die of cancer.

别急着从火星回来,老弟,就待那儿,在那儿fizzbuzz你的注册表、搞你的无代码屎山,直到你得癌症死掉。

#15 No.109436186
Anonymous · 2026-08-02 12:00

>existing piece of software gets vibeshitted

>现有软件被vibeshit化

>quantity of retarded features explodes

>智障功能数量爆炸

>quality of software divebombs

>软件质量俯冲坠毁

we need a Mythical Man-Month for sloptards

我们需要一本给这帮slop废物看的《人月神话》

#16 No.109436189
Anonymous · 2026-08-02 12:01

>autists autisming about muh fizzbuzzing registry dumbing nocodeshittery

>自闭症患者对着“muh fizzbuzz注册表弱智无代码屎山”发癫

Holy.

Anyone who argues your ilk has any right to exist would be immediately branded as an enemy to humanity in any serious society.

任何敢说你这类货色有存在权利的人,在任何正经社会里都会被立刻打成人类公敌。

#17 No.109436210
Anonymous · 2026-08-02 12:03

>>109435851

making schizos explode by ignoring them

无视他们就能让精神病爆掉

#18 No.109436225
Anonymous · 2026-08-02 12:06

>>by ignoring them

>>无视他们

>this is what autists actually, factually believe

>这就是自闭症患者实际认为的事实

#19 No.109436243
Anonymous · 2026-08-02 12:08

>Start the Embedded Linux book and learn how to build a proper SLICK distro for my polarfire SoC

>开始读嵌入式Linux的书,学怎么给我的polarfire SoC搞一个真正顺滑的发行版

>Do actual fun stuff, like image manipulation on the polarfire SoC and the FPGA

>干点真有意思的事,比如在polarfire SoC和FPGA上做图像处理

>Code a game in TempleOS

>在TempleOS上写个游戏

>Code a game on the c64

>在c64上写个游戏

What do?

咋整?

图片: https://i.4cdn.org/g/1785672521156542.png

#20 No.109436276
Anonymous · 2026-08-02 12:13

>>109436243

>Code a game in TempleOS

>在TempleOS上写个游戏

Why would you do that, considering no one's gonna be able to run it.

你干嘛要干这个,反正没人能跑得起来。

#21 No.109436301
Anonymous · 2026-08-02 12:16

>>109436276

>Why would you do that,

>你干嘛要干这个,

Why not?

为什么不?

>considering no one's gonna be able to run it.

>反正没人能跑得起来。

TempleOS is public domain

TempleOS是公有领域的

#22 No.109436314
Anonymous · 2026-08-02 12:18

>>109436301

Useless OS btw.

而且这系统屁用没有。

#23 No.109436349
Anonymous · 2026-08-02 12:24

>>109436314

God talks, you retard. Run TempleOS

上帝在说话,你个傻逼。去跑TempleOS

图片: https://i.4cdn.org/g/1785673460022982.png

#24 No.109436350
Anonymous · 2026-08-02 12:24

>>109436314

Most non-autistic post ITT.

本帖最不自闭的回复。

#25 No.109436365
Anonymous · 2026-08-02 12:26

just let /dpt/ die

就让/dpt/死掉算了

#26 No.109436389
Anonymous · 2026-08-02 12:30

>>109436365

Disagreed, but we have to violently purge its autism.

不同意,但我们必须狠狠清洗它的自闭症。

#27 No.109436579
Anonymous · 2026-08-02 13:01

>>109435809

your attempt to hide your scat fetish by trying to casually talking about eating shit like it's nothing was in vain.

你想装得若无其事聊吃屎来掩盖你的屎癖,白费了。

#28 No.109436726
Anonymous · 2026-08-02 13:23

Finishing up a codewars kata in Java. Doing some Goal (K variant in Go) tutorials, and setting up the clanker on OpenSUSE to try some of my prompts on new deepseek+luna. Probably need to flesh out some more projects as well.

快做完一个Java的codewars kata了。正在看一些Goal(Go的K变体)教程,还在OpenSUSE上配置clanker,准备拿我的提示词去试新的deepseek+luna。可能还得再补几个项目。

#29 No.109437048
Anonymous · 2026-08-02 14:11

>>109435851

I'm making an virtual companion.

我在做一个虚拟伴侣。

#30 No.109437625
Anonymous · 2026-08-02 15:38

>>109435851

I feel appreciated

我感觉自己被赏识了

#31 No.109437883
Anonymous · 2026-08-02 16:19

im sure there's a better way

我肯定有更好的办法

图片: https://i.4cdn.org/g/1785687556985301.png

#32 No.109437922
Anonymous · 2026-08-02 16:23

>>109436186

fork the pre-slop version

分叉出pre-slop版本

then post it here

然后发到这里

then compile a list of pre-slop forks, even if they're maintained by retards

然后整理一个pre-slop分叉列表,哪怕那些分叉是傻逼维护的

then shill that list to the people that hate ai (most programmers)

然后把这个列表推销给讨厌AI的人(大多数程序员)

and once those projects start collapsing and embarrassing themselves, you can start begging for money for your shitty fork. but if and only if you and the people like you can prove that you can actually deliver and do maintainer work

等那些项目开始崩盘丢人现眼,你就可以为你那破分叉讨钱了。但前提是你和你的同类能证明你们真能干成事、能扛起维护工作

#33 No.109438235
Anonymous · 2026-08-02 17:01

>>109437976

I'm not a cuck who would delegate my craft and skills to Americans overlords.

我不是那种会把自己的手艺和技能外包给美国佬的龟男。

Nor do I have the money to spend onto overpriced hardware to selfhost Chinese AI.

我也没有钱去买溢价硬件来自己托管中国AI。

#34 No.109438244
Anonymous · 2026-08-02 17:02

>>109437976

I started to believe there is only a single Indian which posts this same image whenever he finds an AI vs real programmer debate.

我开始相信只有一个印度人,他每次看到AI对真人程序员辩论就发这张图。

#35 No.109438980
Anonymous · 2026-08-02 18:43

You would've thought that even nocodeshitters would now be able to write registry dumpers that finish in seven seconds, yet here we are, two years later, and still nothing.

你会以为连无代码屎工现在都能写出七秒搞定的注册表导出器了,可这都两年了,还是一点动静没有。

图片: https://i.4cdn.org/g/1785696216566843.png

#36 No.109439039
Anonymous · 2026-08-02 18:49

>>109438980

brother don't even bother

兄弟别费劲了

5 years from now windows will be a linux distro; the registry won't be a thing anymore

五年后Windows就是Linux发行版了;注册表这东西就不会存在了

#37 No.109439053
Anonymous · 2026-08-02 18:51

>>109439039

Holy.

The registry is baked into the fucking kernel, you fucking autist.

注册表是焊死在狗屁内核里的,你个傻逼自闭症。

#38 No.109439100
Anonymous · 2026-08-02 18:56

What's faster, Julia or Lua?

Julia和Lua哪个更快?

#39 No.109439107
Anonymous · 2026-08-02 18:57

>>109439100

Never looked at Julia, but Lua's got the most niggerlicious memory allocation bullshit.

从没看过Julia,但Lua的内存分配垃圾得一逼。

>source: Factorio

>来源:异星工厂

#40 No.109439113
Anonymous · 2026-08-02 18:57

>>109439100

you probably want https://terralang.org/

#41 No.109439140
Anonymous · 2026-08-02 19:01

>>109439100

for what?

为了什么?

number crunching probably Julia

数字计算大概用Julia

dynamic stuff Lua

动态部分用Lua

#42 No.109439175
Anonymous · 2026-08-02 19:05

>>109439053

>The registry is baked into the fucking kernel

>注册表直接嵌入内核里了,真他妈服了

yeah, that's why it won't be around once windows adopts the linux kernel

是啊,所以一旦Windows采用Linux内核,这玩意儿就没了

#43 No.109439185
Anonymous · 2026-08-02 19:06

>>109439175

>autists STILL confuse interface with implementation

>神经病还是分不清接口和实现

At this point it's pointless trying to explain this to your ilk. It's like trying to argue with pedophiles.

到现在跟你这种人解释这些已经没意义了,就像跟恋童癖讲道理一样

#44 No.109439191
Anonymous · 2026-08-02 19:07

You lost, progrannies. Vibe gods won.

你输了,程序员奶奶们。氛围神赢了。

#45 No.109439199
Anonymous · 2026-08-02 19:08

>>109438980

#46 No.109439233
Anonymous · 2026-08-02 19:12

>>109439191

this doesn't work because it sounds like you're calling us programmer grannies, which is just funny

这说法不行,听起来像在叫我们“程序员奶奶”,太搞笑了

#47 No.109439333
Anonymous · 2026-08-02 19:24

>>109438980

>microslop in 2026

>2026年还用微软垃圾

Why are people still using this system? Microsoft doesn't want you to use it. They want to get you out. So stop using it

为啥还有人用这系统?微软自己都不想让你用。他们想把你赶出去。所以别用了

#48 No.109439350
Anonymous · 2026-08-02 19:25

>>109439333

Don't tell me, tell the paying customers.

别跟我说,跟付钱的客户说去。

And now make a better dumper.

现在做个更好的转储工具吧。

#49 No.109439356
Anonymous · 2026-08-02 19:26

>>109439350

So are masochistic? I can think of better fun with pain

你是有受虐倾向吗?我能想到更好的疼法玩法

#50 No.109439362
Anonymous · 2026-08-02 19:27

>>109439233

That's what I am calling you though. Get your mind out of the Jewish gutter

我就是这么叫你的啊。把脑子从犹太佬的阴沟里拿出来

#51 No.109439488
Anonymous · 2026-08-02 19:42

>>109439356

Stop deflecting.

别转移话题了。

#52 No.109439493
Anonymous · 2026-08-02 19:43

Should I implement if statements or guards?

我该实现if语句还是守卫条件?

#53 No.109439656
Anonymous · 2026-08-02 20:01

man this general was top tier like 10 years ago

老哥,这个版块十年前可是顶级水准

all the fun tripcode users are gone, all the fun silly projects are gone, we don't even get together to do coding challenges anymore. i hate that this place is a shadow of its past self, but more than that i hate that i'm still on this fucking website all these years later

好玩的tripcode用户都走了,有趣的傻逼项目也没了,我们连coding challenge都不一起搞了。我恨这个地方成了过去的影子,但更恨我这么多年还赖在这个破网站上

图片: https://i.4cdn.org/g/1785700869852213.png

#54 No.109439664
Anonymous · 2026-08-02 20:02

I DRANK CAFFEINEEEEEEEEEEEEEE

#55 No.109439692
Anonymous · 2026-08-02 20:05

>vibecoders thinking using ai will secure them a job while all it will do is make the old guard extend their careers for a decade longer than they used to work for.

>vibecoders以为用AI能保住工作,结果只会让老家伙们多干十年

#56 No.109439707
Anonymous · 2026-08-02 20:07

>>109439656

The entire website is like that. Acknowledge the memories and move on.

整个网站都这样。记住这些回忆,然后往前走吧。

#57 No.109439713
Anonymous · 2026-08-02 20:08

>>109439656

Holy.

I was already there 10 years ago. The two things that have changed:

我十年前就在这儿了。现在变了两件事:

- sepples have been exposed for being complete hacks

- C++被揭穿全是糊弄人的玩意儿

- people ask less questions because LLMs hallucinate the answers for them.

- 人们问的问题少了,因为LLM直接给他们幻觉出答案

#58 No.109439735
Anonymous · 2026-08-02 20:11

genuinely what tf people still post here?

说真的,现在到底谁还在这儿发帖?

It’s getting quite sad

越来越可悲了

#59 No.109439753
Anonymous · 2026-08-02 20:13

cliques... good?

小圈子……好?

#60 No.109439756
Anonymous · 2026-08-02 20:13

>>109439735

They can't program, simple as: >>109438980

他们根本不会编程,就这样:>>109438980

All they do is changing topics and making excuses.

他们只会扯开话题找借口。

#61 No.109439791
Anonymous · 2026-08-02 20:18

i have finally found a medication that actually made a dent in my ability to stay asleep and not wake up after 3 hours, and today i was able to resume work on my concurrent queue design which will totally smash moodycamel. and which still seems to be beyond the ken of LLMs to keep up with. maybe with $1000 of tokens claude could get it done

我终于找到一种药真的能让我睡整觉不半夜醒来了,今天我又能继续搞我的并发队列设计了,绝对能碾压moodycamel。而且这玩意儿看来LLM还是跟不上,也许花1000刀的token让claude能干成

#62 No.109439820
Anonymous · 2026-08-02 20:22

>>109436186

https://codeberg.org/ethical-foss/open-slopware

#63 No.109439837
Anonymous · 2026-08-02 20:24

I have a big dick. I have a big cock. My cock is over average in size.

我有大鸡巴。我有大屌。我屌比平均值大。

#64 No.109439850
Anonymous · 2026-08-02 20:26

Coq

#65 No.109439875
Anonymous · 2026-08-02 20:29

See?

>>109439756

#66 No.109439877
Anonymous · 2026-08-02 20:29

>>109439656

I don't even remember how to do tripcodes.

我都不记得怎么搞tripcode了。

BACK IN THE DAY this thread got spammed with rolls when you posted pic related, too

以前这帖子一发图就全是roll点刷屏

图片: https://i.4cdn.org/g/1785702584837709.jpg

#67 No.109439887
Anonymous · 2026-08-02 20:31

>>109439877

>unironically rolled a 77 by posting this picture

>发这图居然真roll出个77

God literally talks. Damn. Okay next I'll implement AI on an FPGA. But not today.

上帝真在说话。操。好吧,下次我在FPGA上实现AI。但不是今天。

#68 No.109439943
Anonymous · 2026-08-02 20:36

why can't programming languages stop being retarded and just let me have mutually recursive imports.

为什么编程语言就不能别这么傻逼,让我互相递归导入就行啊。

i'm so tired of having to complicate basic things just so the compiler won't screech because two modules need eachother.

我太累了,为了不让编译器因为两个模块互相依赖而尖叫,我得把基本的东西搞复杂。

#69 No.109439959
Anonymous · 2026-08-02 20:38

>>109439943

Just go fucking fix your fucking dependencies.

赶紧去把你那该死的依赖修好。

#70 No.109439993
Anonymous · 2026-08-02 20:41

>>109439959

This

>>109439943

If you use recursive imports, then you should go back to school. Maybe a CS education is not that bad afterall

如果你用循环导入,那你真该回炉重造。也许计算机科学教育没那么差。

#71 No.109439997
Anonymous · 2026-08-02 20:42

>>109439993

>If you use recursive imports, then you should go back to school

>如果你用循环导入,那你真该回炉重造

why?

Explain without fallacies, why they're bad.

别用谬论,解释一下它们为什么不好。

#72 No.109440023
Anonymous · 2026-08-02 20:45

>>109439997

They might not be ""bad"", but they are confusing and unnecessary. You can code easily without them

它们可能不算“不好”,但确实让人困惑而且没必要。不用它们也能轻松写代码。

#73 No.109440062
Anonymous · 2026-08-02 20:48

>>109440023

>but they are confusing and unnecessary

>但它们让人困惑而且没必要

In one reply, we went from "you need to go back to school if you use them" to "well they're not "bad" "

一句话的功夫,我们就从“用它们就该回学校”变成了“嗯它们不算‘不好’”。

lmao

>You can code easily without them

>不用它们也能轻松写代码

I don't want to get pigeon-holed into a paradigm i hate. I want to program and design how I want, that's the entire purpose of a programming language.

我可不想被塞进一个我讨厌的编程范式里。我想按自己的方式来编程和设计,这才是编程语言的整个意义。

#74 No.109440083
Anonymous · 2026-08-02 20:51

>>109440062

>that's the entire purpose of a programming language

>这才是编程语言的整个意义

Well, obviously not of every programming language.

嗯,显然不是每种编程语言都这样。

Go is enforcing things for a reason. It might not be comfortable for YOU, but everyone does it the same way, you can read any other code super easy AND the compiler is super fast, BECAUSE they made a simple to parse language. Some decisions were made just to keep the compiler fast.

Go 强制一些东西是有原因的。可能对你不舒服,但每个人都这么做,你读任何其他代码都超级容易,而且编译器超快,正是因为他们做了个容易解析的语言。有些决定只是为了保持编译器快速。

If you dislike that, well then just use javascript and gpt sol 6.7

如果你不喜欢这样,那就去用 JavaScript 和 gpt sol 6.7 吧。

#75 No.109440092
Anonymous · 2026-08-02 20:52

>>109440083

>Gotard

yeah that makes sense, you should have said this sooner so i could stop thinking you had something worthwhile to say.

对,这说得有道理,你早该这么说,省得我一直以为你有点有价值的话要说。

#76 No.109440128
Anonymous · 2026-08-02 20:57

I hate debugging.

我讨厌调试。

#77 No.109440143
Anonymous · 2026-08-02 20:58

i learned how to program by watching tsoding.

我是看 tsoding 学的编程。

#78 No.109440152
Anonymous · 2026-08-02 21:00

>>109440128

I need to learn tracing function calls in Erlang. I'm tired of io:format/2 debugging.

我需要学在 Erlang 里跟踪函数调用。我烦透了用 io:format/2 调试。

>>109440143

lmao

#79 No.109440215
Anonymous · 2026-08-02 21:06

Is everything an interface or is it just a meme?

是不是所有东西都是接口,还是只是个梗?

#80 No.109440224
Anonymous · 2026-08-02 21:07

>>109440215

Of course not. Sometimes things are just poor implementation.

当然不是。有时候只是实现得很烂。

#81 No.109440229
Anonymous · 2026-08-02 21:08

>>109440215

disregard slop languages. c and bash are all you need.

别管那些垃圾语言。C 和 bash 就够了。

#82 No.109440241
Anonymous · 2026-08-02 21:09

>>109440215

meme

#83 No.109440255
Anonymous · 2026-08-02 21:11

>bash

>where checking if a value is true is its own process

>检查一个值是否为真都要单独开个进程

#84 No.109440270
Anonymous · 2026-08-02 21:12

>>109440255

it is impossible to write a correct bash script anyway

反正写不出正确的 bash 脚本。

#85 No.109440308
Anonymous · 2026-08-02 21:16

I made a language that uses continuation passing and it looks ugly. Any constructive comments?

我做了个用 continuation passing 的语言,看起来很难看。有什么建设性意见吗?

@(\a b =

@a b

) (\cont =

(续)

putchr 65

普奇尔 65

@cont

) (\= putchr 48)

(那是 (\= putchr 48)

#86 No.109440324
Anonymous · 2026-08-02 21:17

>>109440229

Containers are needed more than either of those things.

容器比那两样更需要。

#87 No.109440342
Anonymous · 2026-08-02 21:18

>>109440152

Erlang looks neat. Is it purely functional?

Erlang 看起来挺整洁的。它是纯函数式吗?

#88 No.109440348
Anonymous · 2026-08-02 21:19

>>109440308

It is supposed to be ugly. It should be IR for something else.

它本来就该难看。它应该作为别的东西的中间表示。

SMLNJ and Chicken Scheme do it.

SMLNJ 和 Chicken Scheme 能做到。

#89 No.109440350
Anonymous · 2026-08-02 21:19

>>109440092

NTA, as a smart person I like languages made for dumb people.

不评判,作为聪明人,我喜欢为笨人设计的语言。

#90 No.109440391
Anonymous · 2026-08-02 21:25

>as a smart person

>作为聪明人

Lol

Lmao

Any smart person would like languages that let intelligent people do intelligent things.

任何聪明人都会喜欢能让聪明人干聪明事的语言。

#91 No.109440400
Anonymous · 2026-08-02 21:26

That’s a good idea. Using Y combinators everywhere isn’t that fun from a user perspective.

那是个好主意。到处用 Y 组合子从用户角度看可没那么好玩。

#92 No.109440406
Anonymous · 2026-08-02 21:27

>>109440391

Nah, I want something easy that lets me easily do simple things instead of getting in the way.

不,我想要个简单的东西,能轻松做简单的事,而不是碍手碍脚。

#93 No.109440425
Anonymous · 2026-08-02 21:29

So you're definitely not a smart person, got it.

所以你肯定不是聪明人,懂了。

#94 No.109440448
Anonymous · 2026-08-02 21:32

>>109440425

My IQ is just passing genius. Not ridiculously smart, but I am definitely smart.

我的智商刚好超过天才线。不算离谱的聪明,但我确实是聪明人。

#95 No.109440484
Anonymous · 2026-08-02 21:35

Sure, and I'm the Emperor of China.

当然,我还是中国皇帝呢。

#96 No.109440497
Anonymous · 2026-08-02 21:37

>>109440484

No, you're the registry schizo.

不,你是那个注册表神经病。

#97 No.109440508
Anonymous · 2026-08-02 21:38

Holy autism.

太自闭了。

#98 No.109440513
Anonymous · 2026-08-02 21:38

>>109440448

i don't think you are, mate.

我觉得你不是,哥们。

#99 No.109440543
Anonymous · 2026-08-02 21:41

>>109440513

You're wrong.

你错了。

#100 No.109440551
Anonymous · 2026-08-02 21:42

He's right.

他说得对。

#101 No.109440562
Anonymous · 2026-08-02 21:44

>>109440543

i think you're dumber than a trump supporter. do you have any idea how dumb that is?

我觉得你比川普支持者还蠢。你知道那有多蠢吗?

#102 No.109440569
Anonymous · 2026-08-02 21:44

>>109440513

>>109440551

No, I am a genius. Cope and seethe harder.

不,我是天才。忍着点,继续酸吧。

#103 No.109440579
Anonymous · 2026-08-02 21:46

there has never been an actual genius who uttered the words "I am a genius"

从来没有真正的天才会说出“我是天才”这种话。

#104 No.109440584
Anonymous · 2026-08-02 21:46

>seething and coping about not being a genius

>在这里酸和忍,因为自己不是天才

>while autistically projecting his seething and coping onto others

>还自闭症似的把自己的酸和忍投射到别人身上

Holy.

#105 No.109440606
Anonymous · 2026-08-02 21:48

>>109440579

Wrong. I don't give a fuck about false humility anymore. I am a genius.

错了。我早就不在乎那套虚伪的谦虚了。我就是天才。

#106 No.109440615
Anonymous · 2026-08-02 21:49

>>109440584

>kike is seething over goyim being smarter and genius level.

>犹太佬在酸,因为外邦人更聪明、更天才。

#107 No.109440628
Anonymous · 2026-08-02 21:50

>implying the goyim have intellect

>暗示外邦人有智力

>or souls even

>甚至还有灵魂

#108 No.109440634
Anonymous · 2026-08-02 21:51

TIL registryschizo uses winblows and doesn't believe in geniuses.

今天学到,registryschizo在用winblows,而且不信天才这一说。

#109 No.109440646
Anonymous · 2026-08-02 21:52

>>109440628

Apparently you're also a kike.

显然你也是个犹太佬。

#110 No.109440650
Anonymous · 2026-08-02 21:52

Is this registryschizo in the room with us now?

这个registryschizo现在就在我们房间里吗?

#111 No.109440665
Anonymous · 2026-08-02 21:53

Oy vey, I have a soul now.

哎呀,那我也有灵魂了。

#112 No.109440690
Anonymous · 2026-08-02 21:55

I like Python.

我喜欢Python。

#113 No.109440707
Anonymous · 2026-08-02 21:57

>>109440342

no classical loops in sight

完全没有经典循环的踪影

instead, you use recursion everywhere and make sure they are tail optimized by not doing anything at the end of the function other than calling itself or returning the final value

取而代之的是到处用递归,并且确保函数末尾除了调用自身或返回最终值之外不做任何事,来保证尾递归优化

you store state in a function argument, mostly using a linked list, to which you attach values at the head until you're done

你把状态存在函数参数里,大多用链表,在头部附加值直到完成

it also has really extensive pattern matching, really lends itself well to making/understanding protocols, code below is one such example CBOR decoder excerpt decoding a binary into, in this example, CBOR's major data type 0. it defines how you decode and encode unsigned integers.

它还有非常强大的模式匹配,特别适合编写和理解协议,下面这段代码就是CBOR解码器的一个例子,将二进制解码成,比如说,CBOR的主要数据类型0。它定义了如何解码和编码无符号整数。

first three bits = type

前三位是类型

next 5 bits:

接下来五位:

<24 - value itself

<24 - 值本身

24 - next byte after it is the value

24 - 下一个字节就是值

25 - next 2

25 - 接下来2个字节

26 - next 4

26 - 接下来4个字节

27 - next 8 bytes

27 - 接下来8个字节

the way you'd use this is you'd repeatedly call the cbordataitem function from a recursive function collecting the values and passing the returned binary slice back until you're done

你用法是,从递归函数里反复调用cbordataitem函数,收集值并把返回的二进制片段传回去,直到完成

i don't really know my paradigms, but there is enforced immutability and support for higher order functions

我不太懂编程范式,但这里有强制不可变性和高阶函数支持

other big part of erlang is utilizing processes (not the OS kind, the lightweight VM kind) and message passing between them

Erlang的另一大重点是使用进程(不是操作系统那种,是轻量级VM那种)和它们之间的消息传递

cbordataitem(<<0:3/integer, AddInfo:5/integer-unsigned-big, Rest/binary>>) when AddInfo < 24 ->

cbordataitem(<<0:3/integer, AddInfo:5/integer-unsigned-big, Rest/binary>>) 当 AddInfo < 24 ->

{AddInfo, Rest};

{添加信息,休息};

cbordataitem(<<0:3/integer, 24:5/integer, Value:8/integer-unsigned-big, Rest/binary>>) ->

cbordataitem(<<0:3/整数,24:5/整数,值:8/整数无符号大,其余/二进制>>) ->

{Value, Rest};

{价值,休息};

cbordataitem(<<0:3/integer, 25:5/integer, Value:16/integer-unsigned-big, Rest/binary>>) ->

cbordataitem(<<0:3/整数,25:5/整数,值:16/整数无符号大,其余/二进制>>) ->

{Value, Rest};

{价值,休息};

cbordataitem(<<0:3/integer, 26:5/integer, Value:32/integer-unsigned-big, Rest/binary>>) ->

cbordataitem(<<0:3/整数,26:5/整数,值:32/整数无符号大,其余/二进制>>) ->

{Value, Rest};

{价值,休息};

cbordataitem(<<0:3/integer, 27:5/integer, Value:64/integer-unsigned-big, Rest/binary>>) ->

cbordataitem(<<0:3/整数,27:5/整数,值:64/整数无符号大,其余/二进制>>) ->

{Value, Rest};

{价值,休息};

decode(Binary) ->

解码(二进制)->

decode(Binary, []).

解码(二进制,[])。

decode(<<>>, Acc) ->

解码(<<>>, Acc) ->

Acc;

decode(Binary0, Acc) ->

解码(Binary0,Acc)->

{Value, Binary1} = cbordataitem(Binary0),

{值,Binary1} = cbordataitem(Binary0),

decode(Binary1, [Value | Acc]).

解码(Binary1,[值| Acc])。

i'm no expert though, just a hobbyist

我不是专家,只是爱好者

#114 No.109440781
Anonymous · 2026-08-02 22:03

>>109440707

Neat. I personally like Haskell.

不错。我本人喜欢Haskell。

#115 No.109440782
Anonymous · 2026-08-02 22:03

>>109440690

Use Haskell

用Haskell

#116 No.109440833
Anonymous · 2026-08-02 22:07

>>109440782

I use that too.

我也用那个。

#117 No.109440836
Anonymous · 2026-08-02 22:07

>>109439735

I don't know where else to post about the stuff I'm making

我不知道还能在哪儿发这些关于我做的项目的东西

#118 No.109440869
Anonymous · 2026-08-02 22:10

I am creating an interpreter in C to do string calculations, manipulation and linux administration called Serpent.

我正在用C语言创建一个解释器,用于字符串计算、操作和Linux管理,叫Serpent。

It will be very primitive and will just execute one command after another.

它会非常原始,只是逐条执行命令。

I'll probably implement modes though to stop the console spam.

不过我可能会实现一些模式来避免控制台刷屏。

https://codeberg.org/headlesschook/serpent

图片: https://i.4cdn.org/g/1785708601380589.png

#119 No.109440870
Anonymous · 2026-08-02 22:10

i choose chocolate milk :^)

我选巧克力牛奶 :^)

#120 No.109440901
Anonymous · 2026-08-02 22:11

>>109440869

why do you have a reddit command

你为啥有个reddit命令?

#121 No.109440906
Anonymous · 2026-08-02 22:12

>>109440707

Everything that you described, other languages do better, everything except concurrent processing, I have. If your code does not require erlang processes and concurrency you might as well switch to another language.

你描述的所有东西,其他语言做得更好,除了并发处理,这个我有。如果你的代码不需要erlang进程和并发,不如换别的语言算了。

#122 No.109440936
Anonymous · 2026-08-02 22:15

>>109440901

Why do you call it reddit? Randomized case was used to denote sarcasm on the internet loooooooong before reddit existed.

你为啥叫它reddit?随机大小写早在reddit出现之前很久就被用来表示讽刺了,兄弟。

As far as I am concerned reddit is a dead website moderated by pedophiles and owned by jews like Sam Altman.

在我看来reddit是个死网站,版主是恋童癖,老板是像Sam Altman那样的犹太人。

t.late 30s millennial.

快30末的千禧一代。

#123 No.109440944
Anonymous · 2026-08-02 22:16

Why did Lisp lose the AI wars?

为什么Lisp在AI战争中输了?

#124 No.109440959
Anonymous · 2026-08-02 22:17

>>109440936

>reddit is a dead website moderated by pedophiles and owned by jews like Sam Altman

>reddit是死网站,版主是恋童癖,老板是像Sam Altman那样的犹太人

I wish.

我倒是想。

#125 No.109440980
Anonymous · 2026-08-02 22:20

>>109440906

I'd like to hear about them. The way Erlang does pattern matching feels really good to use.

我想听听它们。Erlang的模式匹配用起来真的很爽。

If you're talking purely about speed then I'm not interested. I don't foresee the need to scale for this project in particular.

如果你只谈速度那我没兴趣。我不觉得这个项目需要扩容。

I don't think I'll be switching languages mid project either. Maybe next project.

我也不会中途换语言。可能下个项目再说吧。

#126 No.109440992
Anonymous · 2026-08-02 22:21

>>109440959

It is dead, It's just bots posting the same karma farming bullshit every single day and every single AI model is trained on the shit.

它确实死了,每天就是机器人发一样的刷分垃圾,每个AI模型都用这堆屎训练出来的。

#127 No.109440999
Anonymous · 2026-08-02 22:22

>>109440959

#128 No.109441038
Anonymous · 2026-08-02 22:27

Is Java faster than LuaJIT?

Java比LuaJIT快吗?

#129 No.109441188
Anonymous · 2026-08-02 22:42

>>109440944

because it was a retarded idea from the start, even in symbolic ai. the japanese had the right idea, but it required infinite money that america didn't want to put into their ai industry.

因为这从一开始就是个蠢主意,就算在符号AI里也是。日本人当时思路是对的,但需要无限的钱,美国不想往AI行业里砸。

#130 No.109441370
Anonymous · 2026-08-02 23:07

>racket is vibecoded now

>racket现在也是vibecoded了

give me alternatives; preferably scheme but any non-common lisp is fine

给我点替代品;最好scheme,但任何非common lisp的都行

>>109441038

it depends

看情况

#131 No.109441538
Anonymous · 2026-08-02 23:30

I fixed a really hard bug and now I'm happy.

我修了一个超难的bug,现在很开心。

#132 No.109441560
Anonymous · 2026-08-02 23:33

This program is prettier.

这个程序漂亮多了。

@(\n f cont = @f n f 1 cont) 3 (\n f r cont =

@(\n f 续 = @f n f 1 续) 3 (\n f r 续 =

| n = @f n - 1 f n * r cont

| n = @f n - 1 f n * r 续

| _ = @cont r

) (\c = putchr c + 48)

#133 No.109441588
Anonymous · 2026-08-02 23:35

>>109441560 (Me)

It's also really fast.

而且速度很快。

#134 No.109441624
Anonymous · 2026-08-02 23:40

My Brainfuck compiler is very well optimized.

我的 Brainfuck 编译器优化得非常好。

图片: https://i.4cdn.org/g/1785714013988583.png

#135 No.109441640
Anonymous · 2026-08-02 23:41

>>109441370

Why is it being vibecoded bad?

为什么说用 vibecoded 写代码不好?

#136 No.109441641
Anonymous · 2026-08-02 23:41

>>109441560

What do the slashes do?

那些斜杠是干嘛的?

#137 No.109441657
Anonymous · 2026-08-02 23:43

>>109441641

They start a lambda.

它们开启一个 lambda。

#138 No.109441680
Anonymous · 2026-08-02 23:47

>>109441640

because llms produce trash code that doesnt even work most of the time

因为 LLM 产出的代码大部分时间根本跑不起来,纯垃圾。

and works only approximately for the rest

剩下那部分也只是近似能用。

llms have limitations that are inherent to the technology

LLM 的局限是技术本身就有的。

unless you call complete handholding- vibecoding too

除非你把完全手把手也算 vibecoding——

which is ultra gay and its 100% on you for not inventing a term that doesnt conjure broken, dyslexic code in the mind of your interocutor

那也太基了,完全是你自己的锅,谁让你不发明一个不会让人联想到烂得掉渣、像阅读障碍写出来的代码的词。

#139 No.109441685
Anonymous · 2026-08-02 23:48

What's the most Christian programming language?

最基督教的编程语言是什么?

图片: https://i.4cdn.org/g/1785714505294928.jpg

#140 No.109441697
Anonymous · 2026-08-02 23:50

>>109441680

>because llms produce trash code that doesnt even work most of the time

>因为 LLM 产出的代码大部分时间根本跑不起来,纯垃圾

You need to stop living in 2023. The rest of the world (LLMs included) have moved on. You sound just like the people who think LLMs can't do math better than pic related.

别活在 2023 年了。全世界(包括 LLM)早就进步了。你听起来就像那些觉得 LLM 数学还不如图里那玩意儿的人。

图片: https://i.4cdn.org/g/1785714608757647.jpg

#141 No.109441704
Anonymous · 2026-08-02 23:50

>>109441685

HolyC

#142 No.109441706
Anonymous · 2026-08-02 23:50

i just wrote a webserver for an esp32-wroom-32d. i have have an esp32 on my network with a webpage that tells the visitor to kys

我刚给 esp32-wroom-32d 写了个网页服务器。我网络里有个 esp32 挂着一个网页,内容就是叫访客去死。

#143 No.109441756
Anonymous · 2026-08-02 23:58

I think I get it now.

我好像懂了。

AI is actually super powerful but they need to train it on pure retardation to dilute it like Wheatly and GLaDOS.

AI 其实超强,但他们得拿纯傻逼数据来训练稀释它,就像 Wheatly 和 GLaDOS 那样。

图片: https://i.4cdn.org/g/1785715122994615.jpg

#144 No.109441769
Anonymous · 2026-08-03 00:00

If LLMs were as good as they say we'd all be programming in assembly, period.

如果 LLM 真像吹的那么好,我们早全在写汇编了,就这样。

#145 No.109441802
Anonymous · 2026-08-03 00:03

>>109441769

It should be able to create a direct binary.

它应该能直接生成二进制文件。

The whole point of a programming language's syntax is to make cpu instructions human readable.

编程语言语法存在的全部意义,就是让 CPU 指令对人类可读。

#146 No.109441924
Anonymous · 2026-08-03 00:19

Is continuation passing naturally fast?

续延传递天生就快吗?

#147 No.109441949
Anonymous · 2026-08-03 00:24

>>109441697

nono

i use llms every day

我天天用 LLM。

even with thousands usd worth of tokens you get basically a copy pasted jumble of disparate solution without any coherent architecture

就算烧掉几千美元的 token,你得到的也基本是零散解决方案的拼贴大杂烩,根本没有什么连贯架构。

the only thing that actually improved is that now we have harnesses which automajically reprompt the model by feeding it the resulting errors

真正变好的只有一点:现在我们有了安全带,能自动把报错喂回给模型重新提示。

if you want to compare the llms proper, compare what can be one-shotted now vs what could be one-shotted in 2025

要比较真正的LLM,得看现在能一枪打中的问题 vs 2025年能一枪打中的问题。

and on that point, not much has changed

就这点而言,变化不大。

#148 No.109442009
Anonymous · 2026-08-03 00:37

>>109441949

we can one-shot a lot more unsolved math problems

我们现在能一枪解决更多未解的数学题了。

#149 No.109442012
Anonymous · 2026-08-03 00:37

>>109441924

No, it has to be optimized/designed for.

不,那必须得专门优化/设计才行。

#150 No.109442026
Anonymous · 2026-08-03 00:40

>>109442009

yeah, sure

是啊,当然。

one shot them so hard you need a team of mathematicians to clean up the mess afterwards

一枪解得那么猛,完了还得请一队数学家来收拾残局。

#151 No.109442078
Anonymous · 2026-08-03 00:48

I ran the procode. The codegram is executing.

我跑了procode。代码图正在执行。

图片: https://i.4cdn.org/g/1785718086837530.png

#152 No.109442182
Anonymous · 2026-08-03 01:06

Cniles in a nutshell

一句话总结Cniles。

图片: https://i.4cdn.org/g/1785719199508266.png

#153 No.109442240
Anonymous · 2026-08-03 01:15

>>109441802

this is such a dogshit stupid idea. llm slop code is already difficult enough to debug and fix at a high level

这真是个狗屎烂主意。LLM生成的垃圾代码本来就够难调试和修了,在高层次上更是如此。

#154 No.109442256
Anonymous · 2026-08-03 01:18

>>109442240

>what do you mean, what would I feel like if I hadn't had breakfast?

>你说啥,要是没吃早饭我会有啥感觉?

>>109441769

#155 No.109442386
Anonymous · 2026-08-03 01:39

J is so cute.

J太可爱了。

#156 No.109442472
Anonymous · 2026-08-03 01:52

>>109442240

If ai was truly intelligent it would not need fixing you fucking retarded nigger.

如果AI真聪明,它就不需要修,你这傻逼黑鬼。

#157 No.109442685
Anonymous · 2026-08-03 02:25

J is truly intelligent.

J才是真聪明。

#158 No.109442801
Anonymous · 2026-08-03 02:38

main = print $ sum[1..5]

主要 = 打印 $ sum[1..5]

#159 No.109442917
Anonymous · 2026-08-03 02:53

Thank you for reading my blog. You may now comment.

谢谢阅读我的博客。现在可以留言了。

#160 No.109442923
Anonymous · 2026-08-03 02:54

>>109442801

>ambiguous type a where (Num a, Show a) defaulted to Integer

>模糊类型a,其中(Num a, Show a)默认成Integer

#161 No.109442932
Anonymous · 2026-08-03 02:55

>>109442923

o and Enum

还有Enum。

Fuck Enum btw

顺便去他妈的Enum。

#162 No.109442945
Anonymous · 2026-08-03 02:57

>>109442923

You're doing something wrong.

你肯定做错了什么。

#163 No.109442973
Anonymous · 2026-08-03 02:59

>>109442945

I didn't even need to run

我甚至不用跑

ghc -e "print (sum [1..3])" -Wall

ghc -e“打印(总和[1..3])”-Wall

to know what it would tell me

就知道它会说啥。

#164 No.109443023
Anonymous · 2026-08-03 03:05

>>109442973

You're doing something wrong.

你肯定做错了什么。

#165 No.109443038
Anonymous · 2026-08-03 03:08

>>109443023

enabling warnings?

开警告?

#166 No.109443199
Anonymous · 2026-08-03 03:26

>>109443038

Warnings don't show that.

警告不会显示那个。

#167 No.109444220
Anonymous · 2026-08-03 05:47

>to delete a file on Windows the file has to be first opened (NtCreateFile/NtOpenFile) with permission to delete the file

>要在Windows删文件,得先用删除权限打开该文件(NtCreateFile/NtOpenFile)

>then inform the kernel that the file is to be deleted once the handle is closed (NtSetInformationFile(FileDispositionInformationEx))

>然后告诉内核,等句柄关闭后这个文件要被删(NtSetInformationFile(FileDispositionInformationEx))

>then close the handle (NtClose)

>然后关闭句柄(NtClose)

Who comes up with this shit?

谁想出来的这堆破玩意?

图片: https://i.4cdn.org/g/1785736063500880.jpg

#168 No.109444264
Anonymous · 2026-08-03 05:56

>>109444220

Smart people. Not autists.

聪明人。不是自闭症患者。

#169 No.109444299
Anonymous · 2026-08-03 06:01

Oh, this was *definitely* done by autists, because these interfaces do not support batches. It's just a question of if these autists were even more retarded than already extraordinarily retarded autists or not.

哦,这*绝对*是自闭症患者搞的,因为这些接口不支持批量操作。问题只在于这帮自闭症患者是不是比已经极度自闭的那批还要更蠢。

图片: https://i.4cdn.org/g/1785736903810782.jpg

#170 No.109444744
Anonymous · 2026-08-03 07:10

>>109439877

roll, babyyyyy

滚起来,宝贝。

#171 No.109444782
Anonymous · 2026-08-03 07:19

>>109444264

win32 is the antithesis of smart.

win32就是聪明的反面。

It even over-complicates the most basic thing.

连最基本的事都能搞复杂。

But conversely, can't tell you specifics for anything meaningful, like when you pass the wrong size to a struct field that is passed to a function.

但反过来,对任何有意义的具体细节都说不出个所以然,比如你给传给函数的结构体字段传错大小时。 ﹚ 这可不是Win32。这是NTAPI。

图片: https://i.4cdn.org/g/1785741573269606.png

#172 No.109444793
Anonymous · 2026-08-03 07:20

>>109444782

This isn't Win32 though. This is the NTAPI.

这可不是Win32,这是NTAPI。

#173 No.109444798
Anonymous · 2026-08-03 07:21

>>109444793

NT is just lower level win32 (windows), it's all M$ garbage

NT就是更低级的win32(Windows),全是微软的垃圾。

#174 No.109444802
Anonymous · 2026-08-03 07:22

>>109443199

-fwarn-type-defaults

-fwarn-类型默认值

#175 No.109444819
Anonymous · 2026-08-03 07:25

>>109444798

No it's not. Win32 has the DeleteFileW call that takes a simple file name to delete it, but internally it does the aforementioned nonsense which you cannot avoid, plus a bunch of string copies and memory allocations of course, which you can avoid as long as you use the NTAPI directly.

不对。Win32有DeleteFileW这个调用,只需要简单文件名就能删,但内部还是得走前面说的那堆破流程,你躲不掉,外加一堆字符串拷贝和内存分配,这些只要你直接用NTAPI就能避免。

#176 No.109445825
Anonymous · 2026-08-03 10:11

does /dpt/ do AI applications?

/dpt/ 做AI应用吗?

I'm working on a system that helps you find MTG cards with specific effect(s).

我在搞一个系统,帮你找有特定效果的MTG牌。

There's an existing database where users assign to every card one or multiple tags out of ~2500 available ones. So, the problem I'm actually solving is finding the related tag(s) from the user's prompt.

有个现成数据库,用户给每张牌标一或多个标签,总共约2500个可选。所以,我实际要解决的问题是从用户提示里找到相关标签。

This seems like a retrieval problem, so I tried an encoder model + bi-encoder stack (BAAI BGE encoders and MS marco cross-encoder). However they don't perform well because of the following problems:

这看起来是个检索问题,所以我试了编码器模型+双编码器堆叠(BAAI BGE 编码器和 MS marco 交叉编码器)。但效果不咋样,问题如下:

>they don't know MTG terminology

>它们不懂MTG术语

the embedding for "deal damage" will not be close enough to "burn" or "bolt"

"deal damage" 的嵌入向量跟 "burn" 或 "bolt" 的向量距离差太远

>they depend too much on matching words

>太依赖词面匹配

"deal damage" will end up similar with "scales with damage dealt", which is semantically different

"deal damage" 会和 "scales with damage dealt" 搞混,实际这俩语义差远了

"hate indesctructible" will end up similar with "gives indesctructible" although they're wildly different effects

"hate indesctructible" 会和 "gives indesctructible" 相似,但效果天差地别

"synergy-X" will end up similar with "synergy-Y" even if X and Y have little to no similarity

"synergy-X" 会和 "synergy-Y" 差不多,即使 X 和 Y 八竿子打不着

I'm a bit lost on how to move on from here. Do I finetune the models and hope for the best?

我现在有点不知道咋走下一步,是微调模型然后赌一把吗?

An alternative idea is to do "in-context retrieval" aka throw every tag at an LLM's system prompt and ask it to find relevant ones. However when attempting to refine it by moving some tags behind tool functions, the agent/LLM does not really understand when it actually needs them. For example, there's ~300 tags relating to synergy with specific creature types (tribal tags). I gate them behind a tool called get_tribal_tags and explain that they're only needed if the user asks for synergy with a specific creature type, but the agent will frequently use that tool after every prompt in order to "check which tags are available" (even if the user's prompt doesn't mention specific creature types).

另一个思路是"上下文检索",即把所有标签塞进 LLM 的 system prompt 里让它找相关的。但想优化一下把部分标签挪进工具函数后,agent/LLM 就搞不清啥时候真的需要它们。比如,有大概300个跟特定生物类型配合相关的标签(部族标签)。我把它们放到一个叫 get_tribal_tags 的工具后面,解释只有用户提到跟特定生物类型配合时才需要,但 agent 几乎每次提问后都会调这工具去"看看有哪些标签"(哪怕用户没提任何特定生物类型)。

#177 No.109446522
Anonymous · 2026-08-03 11:54

>>109445825

What exactly do you want us to tell you? That LLMs are unreliable pieces of junk?

你到底想让我们说啥?说LLM就是不靠谱的垃圾?

图片: https://i.4cdn.org/g/1785758055820643.jpg

#178 No.109446597
Anonymous · 2026-08-03 12:05

>>109446522

this will happen with the AI photo apps for sure, if not already happened.

这种破事肯定也要发生在AI照片应用上,没准已经发生了。

But it also has a good thing. We try to fight natural selection with medicine, but AI is bringing it back on track

但这也是好事。我们用药物对抗自然选择,AI正把它拉回正轨。

#179 No.109446633
Anonymous · 2026-08-03 12:09

>>109446597

>will

Aren't we optimistic.

我们可真是乐观啊。

图片: https://i.4cdn.org/g/1785758999343840.png

#180 No.109447014
Anonymous · 2026-08-03 13:06

>>109435851

>what are you working on, /dpt/?

>你在搞啥,/dpt/?

learning c64/basic

在学c64/basic

Making small steps. I try wading through the shit and do as much as possible on the c64 itself. Clicked together the sprite on the desktop tho

慢慢来。我在烂泥里摸爬滚打,尽量在c64本机上搞。虽然精灵图是在桌面端拼的

图片: https://i.4cdn.org/g/1785762379301074.png

#181 No.109447152
Anonymous · 2026-08-03 13:27

>>109446522

junk or not there's no way that you can't solve the problem I'm describing with them in 2026. I'm just looking for the easiest way to do so

不管垃圾不垃圾,2026年了不可能解决不了我描述的问题。我就是在找最省事的办法。

#182 No.109447161
Anonymous · 2026-08-03 13:31

>>109447152

Good luck, see you in 2027, and in 2028, and 2029 ...

祝你好运,2027年见,然后2028年,2029年……

#183 No.109447279
Anonymous · 2026-08-03 13:50

>>109441697

Are you being sarcastic or did you not read the LLM response? The answer is correct but the explanation is way off. 1000 is a 4 digit number. The actual rigorous proof would probably require dozen of pages, but intuitive human explanation is that 1062 further to the right in the number line. The LLM could have just omitted the explanation and that would have been acceptable too. But no, it's probably badly parroting a portion of some obscure theorem.

你是故意的还是没读LLM回复?答案是对的但解释扯远了。1000是个四位数。真正严谨的证明可能得写几十页,但直觉人话就是1062在数轴上更靠右。LLM完全可以不发那段解释,那也能接受。可不,它八成是在乱转述某个冷门定理的一部分。

#184 No.109447411
Anonymous · 2026-08-03 14:03

>>109447279

>The LLM could have just omitted the explanation and that would have been acceptable too.

>LLM完全可以不发那段解释,那也能接受。

$ perl -e 'if(1000 <= 1062){print "Christ are you retarded!\n"}'

$ perl -e 'if(1000 <= 1062){print "天哪,你是不是傻!\n"}'

#185 No.109447643
Anonymous · 2026-08-03 14:35

>>109447161

I get hating on the endless AI push but that's just silly. Did you read my post or did seeing "LLM" in it trigger you? Or do you think a model suggesting "tag: card advantage" when you ask it for card draw is some unsolved problem?

我理解对AI这股永无止境的浪潮有反感,但那也太蠢了。你读了我的帖子吗,还是看到“LLM”这个词就触发了你?又或者你以为一个模型在你问它抽牌时建议“标签:卡差优势”是什么未解难题?

#186 No.109447677
Anonymous · 2026-08-03 14:39

... and 2030 and 2035 and 2050 ...

……还有2030、2035、2050……

#187 No.109447689
Anonymous · 2026-08-03 14:40

>>109445825

my first instinct would be to do your alternative strategy but I'm fairly sure you don't even need a full-blown LLM for that

我的第一反应是做你说的备选方案,但我相当确信那根本不需要一个完整的LLM。

I was thinking of doing something similar but categorizing used car sales (they usually contain lots of vague/not very precisely written information)

我也想过做类似的事,不过是用在二手车销售分类上(那些信息通常很含糊、写得不够精确)。

#188 No.109447720
Anonymous · 2026-08-03 14:44

>>109447643

there's no point rationalizing with them anon.

匿名网友,跟他们讲道理没用。

#189 No.109447867
Anonymous · 2026-08-03 15:00

>>109446522

if you read his problem it is exactly the kind of thing where LLMs being unreliable is not a massive hinderance, as long as it can do something like 80-90% somewhat correctly.

你要是读了他描述的问题,就会发现这正是LLM不可靠并不是大碍的场景,只要能做到大概80-90%基本正确就行。

#190 No.109447928
Anonymous · 2026-08-03 15:08

>he believes it's 80-90%

>他居然信能到80-90%

#191 No.109447954
Anonymous · 2026-08-03 15:10

>>109437048

me too, I’m just a bit conflicted about how to release it

我也是,只是有点纠结怎么发布它。

#192 No.109448188
Anonymous · 2026-08-03 15:35

>get ticket assigned to me

>工单分派到我头上

>need to add new feature to old file

>要给旧文件加新功能

>open file

>打开文件

>1000 lines

>1000行

>extension is .ts but was just copy pasted from old js code

>扩展名是.ts,但就是从旧js代码直接复制粘贴来的

>none of the dependencies have type declarations

>所有依赖都没有类型声明

>1: //@ts-no-check

I want to die

我想死

pic unrel a forum I'm working on as a form of therapy

发个无关图,是我当疗愈在做的一个论坛。

图片: https://i.4cdn.org/g/1785771304748049.png

#193 No.109448590
Anonymous · 2026-08-03 16:16

why is /dpt/ autistic while /vcg/ is not? This is very counter-intuitive

为什么/dpt/是自闭症而/vcg/不是?这太反直觉了。

#194 No.109448716
Anonymous · 2026-08-03 16:27

>absence of evidence == evidence of absence

>证据缺失就等于证据不存在

Well, at least we got you confirmed.

嗯,至少确认了你就是这样的。

#195 No.109448825
Anonymous · 2026-08-03 16:37

>>109448716

>>absence of evidence == evidence of absence

>>证据缺失就等于证据不存在

https://www.youtube.com/watch?v=RxMYvPy8QwA

FOR AN AUSTRALIAN THIS IS HEAVY DUTY SHIT

图片: https://i.4cdn.org/g/1785775020664680.jpg

#196 No.109448887
Anonymous · 2026-08-03 16:42

>>109448825

I like how he's manipulating the crowd into not thinking, otherwise they'd run him out of the city.

我喜欢看他怎么操控群众不去思考,不然他们早就把他赶出城了。

#197 No.109449244
Anonymous · 2026-08-03 17:19

>>109439877

Does rolling even make sense for 145 options?

145个选项还用滚骰子有意义吗?

#198 No.109449596
Anonymous · 2026-08-03 17:56

>>109444744

>>109449244

>fizzbuzz two times in a row

>连续两次fizzbuzz

KEK

#199 No.109449714
Anonymous · 2026-08-03 18:10

Hello, fat hairy manchildren.

你们好啊,又肥又毛茸茸的巨婴们。

#200 No.109449765
Anonymous · 2026-08-03 18:16

>>109449596

Maybe next time I'll get something actually fun.

也许下次我就能遇到点真正好玩的了。

图片: https://i.4cdn.org/g/1785781001712793.png

#201 No.109450241
Anonymous · 2026-08-03 19:07

>>109449765

Registry dumper.

注册表导出工具。

#202 No.109450284
Anonymous · 2026-08-03 19:14

>>109450241

Sorry, I don't use windows.

抱歉,我不用Windows。

#203 No.109450322
Anonymous · 2026-08-03 19:17

>>109450284

>not doing a prototype in Wine

>不在Wine里做原型

>or in a VM

>especially since images are free to get these days

>尤其是现在图片都免费随便拿了

Just stop making excuses.

别找借口了。

#204 No.109450502
Anonymous · 2026-08-03 19:34

>>109448590

Autists come in two varieties: obsessed with highly technical stuff (math, coding, physics, etc), and obsessed with artistic stuff (history of the roman empire, science fiction/fantasy, etc) and sometimes there's a cross between the two (obsessed with trains: feats of mechanical engineering which are also beautiful big pieces of metal).

自闭症患者分两种:一种痴迷于高度技术性的东西(数学、编程、物理等),另一种痴迷于艺术性的东西(罗马帝国史、科幻/奇幻等),偶尔也有两者交叉的(痴迷火车:既是机械工程的杰作,又是漂亮的大块金属)。

People who actually know how to code are likely somewhere on the spectrum - to get REALLY good at coding requires an autistic level of obsession with things that others find mind-numbingly boring. Autists were looked down on because of their obsessions, until about the 2010s when coding websites became a fashionable job skill (and paid fuckloads of money, too)

真正会写代码的人多半在自闭谱系上——要真正精通编程,就得上瘾一样痴迷于那些别人觉得无聊到爆的东西。自闭症患者因为痴迷而被瞧不起,直到2010年代左右编程网站变成时尚职业技能(还赚大钱)才翻身。

The normies, having been filtered by basic math skills, seethed as they tried to make a living doing office bitchwork like sales, marketing, human resources, etc. They coveted the job security and permissive workplace culture that the coders enjoyed, but they were too stupid to be coders. Javascript allowed some actual troglodytes to squeeze by and make poorly performing react sites, putting themselves on the very shores of the dark continent of autism spectrum disorder.

那些被基础数学能力筛掉的普通人,气得直跳脚,只能靠做销售、市场、人事这类办公室杂活混口饭吃。他们眼馋程序员的工作稳定和宽松的职场文化,但脑子不够用当不了程序员。Javascript倒是让一些真·山顶洞人勉强挤进来,做出性能稀烂的react网站,把他们自己推到了自闭症谱系这片黑暗大陆的岸边。

Then, LLMs came out. Now, actual fucking idiots no longer needed to learn to code to make websites. They get to continue being retards (note: autist != retard) but still can shit out a crappy website occasionally.

然后,大语言模型出来了。现在,真正的蠢货不用学编程也能做网站了。他们继续当他们的脑残(注意:自闭症患者≠脑残),但偶尔也能拉坨屎一样的破网站出来。

You see, the biggest defenders of LLM usage are usually the least capable people on a coding team. They have an inherent imposter syndrome because they actually ARE imposters. If you challenge the way they use AI, they get very defensive because they know they're frauds who didn't do any of the work to get where they are.

你看,最力挺LLM用法的往往是编码团队里能力最差的那帮人。他们天生就有冒充者综合征,因为他们确实就是冒充者。你要是质疑他们用AI的方式,他们会特别防御,因为他们知道自己就是骗子,根本没付出任何努力就混到了今天的位置。

Ultimately, both the autists and the retards will probably end up getting paid minimum wage to prompt claude to make websites, and the field of software engineering will be completely eviscerated as trillionaires make more trillions of dollars off of human suffering.

最终,不管是自闭症患者还是脑残,大概都会拿着最低工资去提示Claude做网站,而软件工程这个领域会被彻底掏空,亿万富翁们靠着人类的苦难赚更多的亿万。

图片: https://i.4cdn.org/g/1785785694258944.jpg

#205 No.109450570
Anonymous · 2026-08-03 19:40

>>109450502

>note: autist != retard

>备注:自闭症患者≠脑残

Oh yes, they are.

哦对,他们就是。

I've written an entire fucking book about why the do-one-thing-and-do-it-well philosophy that is still present in most programmer's minds is nothing more than autism being made manifest. If we had culled the autists in the 1980s we wouldn't be on the path of a complete civilizational collapse right now, because there wouldn't be any autists that design interfaces completely counter-intuitive to how hardware has been working since at least 1989.

我写过一整本他妈的书,讲为什么那种“只做一件事并做好”的哲学——至今还残留在大多数程序员脑子里——不过是自闭症的具象化。如果80年代我们就把自闭症患者消灭了,现在根本不会走在文明全面崩塌的路上,因为就不会有自闭症患者去设计那些完全违背1989年以来硬件运作逻辑的界面了。

But then again I'm arguing for autists to be culled in a general inhabited by autists. Of course my message is ignored.

但话说回来,我在一个自闭症患者聚集的版块里主张消灭自闭症患者,我的观点当然没人理。

#206 No.109450606
Anonymous · 2026-08-03 19:43

>>109448590

/dpt/ is not autistic it's schizophrenic.

/dpt/不是自闭,是精神分裂。

#207 No.109450610
Anonymous · 2026-08-03 19:43

>>109450570

>inb4 what's the name of the book

>我知道有人要问书名叫啥

Not released yet. Might not ever possibly. My father's death a couple months ago has seriously derailed my life.

还没出版。可能永远也不会出版了。我父亲几个月前去世了,把我的人生彻底搞乱了。

#208 No.109450613
Anonymous · 2026-08-03 19:44

>>109450570

>Giving a fuck about UI design

>居然有人在乎UI设计

图片: https://i.4cdn.org/g/1785786247682666.gif

#209 No.109450649
Anonymous · 2026-08-03 19:48

>>109439877

these challenge sheets were always made in a retarded way, ie. someone asked the thread for suggestions, was very slow to figure out the maybe 20 or so replies wasn't nearly enough, then just slapped on literally whatever just to get to 100

这些挑战表从来都是用脑残方式做的,比如有人来帖子里要建议,过了半天才反应过来那大概20条回复根本不够,然后随便乱凑数硬凑到100条

that's why these "challenges" are always all over the place

所以这些“挑战”从来都是乱七八糟的

then eventually the next retard would notice the challenge sheet wasn't really good and decided to make a "better" one, repeating the exact same mistakes as the last guy

然后下一个脑残发现挑战表不太好,决定做一个“更好”的,结果把上一个家伙犯的错全重复一遍

#210 No.109450669
Anonymous · 2026-08-03 19:49

>>109450613

>UI design

>UI设计 ❞ 重度自闭症:

Holy autism:

我的天,这也太自闭了吧:

>>109444220

>>109444299

Although, one could argue that MS should've looked more into batching kernels even back then. After all, why can't messages be retrieved in bulk?

不过,有人可能会说微软当时就该更深入地研究内核批处理。毕竟,消息为什么不能批量拉取?

图片: https://i.4cdn.org/g/1785786593095747.gif

#211 No.109450733
Anonymous · 2026-08-03 19:58

>>109450669

Oh, and while we're at it:

哦,顺带一提:

>did you know there's no resource file parser in the kernel?

>你知道内核里没有资源文件解析器吗?

>it's all userspace-based

>全部基于用户态

It's like ... the ONE reason why you'd want a resource file with all the resources and UI elements is to pass this shit over to the kernel, and it's not even userspace-based. It's why I can laugh at complete fucking retards like Raymond Chen. You either need to be a genius or a complete retard to laugh at retards like Chen, and I have debugged way too much MS code to be called "complete retard".

这就像……你之所以想要一个包含所有资源和UI元素的资源文件,唯一的原因就是为了把这玩意儿传给内核,结果它甚至都不是用户态实现的。这就是为什么我能嘲笑像Raymond Chen那样的纯傻逼。你要么是个天才,要么是个彻头彻尾的傻逼才能嘲笑Chen这样的傻逼,而我调试过太多MS的代码,配不上"彻头彻尾的傻逼"这个称号。

图片: https://i.4cdn.org/g/1785787138457697.jpg

#212 No.109451046
Anonymous · 2026-08-03 20:33

>>109450322

>something fun

>有点意思的事

>dealing with windows

>跟Windows打交道

Pick one and only one.

二选一,只能选一个。

#213 No.109451047
Anonymous · 2026-08-03 20:33

>>109450669

>batching kernels

>批处理内核

Man, that's one batchin' kernel there!

卧槽,这内核可真够"批"的!

#214 No.109451070
Anonymous · 2026-08-03 20:35

>>109451046

>he thinks there's one, and only one

>他觉得只有一个,且仅有一个

Autism incarnate.

自闭症本症。

>>109451047

There's none.

一个都没有。

#215 No.109451644
Anonymous · 2026-08-03 21:32

do any uni classes actually teach you shit about systems design or am I supposed to learn how to set up load balancers and cdns and auth servers and stateless shit on the cloud from experience and youtube tutorials

大学课程到底有没有教系统设计这种实打实的东西,还是说设置负载均衡、CDN、认证服务器以及云上无状态那堆玩意儿,都得靠经验和YouTube教程自学?

图片: https://i.4cdn.org/g/1785792768438349.png

#216 No.109451725
Anonymous · 2026-08-03 21:39

>>109451644

>how to set up

>如何设置

yes, universities will teach theory and well established standards that are at least 20 years old

是的,大学会教你理论以及那些至少二十年历史的成熟标准。

#217 No.109451868
Anonymous · 2026-08-03 21:52

I'm getting good (this is almost identical to the previous potds, so I already knew the solution)

我做得挺好(这个跟之前的potd几乎一模一样,所以我早知道解法)。

https://leetcode.com/problems/stone-game-iii/

图片: https://i.4cdn.org/g/1785793962200382.png

#218 No.109452027
Anonymous · 2026-08-03 22:10

>>109451868

This shit bores me.

这玩意儿让我犯困。

I might solve it in a few days.

我可能过几天再解。

#219 No.109452037
Anonymous · 2026-08-03 22:11

>>109452027

that is like one of the easiest leetcode hards on the platform, you just need to know basic 1 state top down dp

这算是平台上最简单的LeetCode hard之一了,你只需要知道基础的1状态自顶向下DP就行。

#220 No.109452045
Anonymous · 2026-08-03 22:12

>>109435851

I've been working in my automata/FSM experimentation sandbox. Not really doing anything in particular at present. Last week I actually spent some time making manual coroutines though, it was pretty interesting.

我一直在搞我的自动机/FSM实验沙盒。目前没在做什么特别的事。不过上周我花时间手写了协程,还挺有意思的。

I also experimented with gate guarded edges for state transitions too.

我还试了用门卫边来做状态转换。

#221 No.109452267
Anonymous · 2026-08-03 22:33

>>109444744

i did it

我搞定了

(3‿5 (0=|)◶[⊢‿"fizz","buzz"‿"fizzbuzz"] ⊢)¨ 1+100

这个代码用于解决FizzBuzz问题,输出1到100范围内的相应结果。

#222 No.109452723
Anonymous · 2026-08-03 23:18

>>109452037

I don't know. Nqueens was more interesting.

不知道。N皇后更有意思。

#223 No.109452731
Anonymous · 2026-08-03 23:19

>>109452723

nqueens is backtracking doe, this is dp

N皇后是回溯啊,这个是DP。

#224 No.109453834
Anonymous · 2026-08-04 01:35

>>109452731

Still more interesting.

还是更有意思。

DP is weird. Sometimes the solution is just intuitive and other times it just doesn't seem to be.

DP这东西很怪。有时候解法就是直觉性的,有时候又完全摸不着头脑。

#225 No.109453971
Anonymous · 2026-08-04 01:50

>>109436028

One thing I want to do is try and get Claude or some other agentic coding AI to build stuff while I go on rides

我想试试让Claude或者其他agentic编码AI在我不在的时候干活儿,我出去骑车。

#226 No.109454222
Anonymous · 2026-08-04 02:21

Java is the best.

Java是最好的。

图片: https://i.4cdn.org/g/1785810104220892.jpg

#227 No.109454285
Anonymous · 2026-08-04 02:30

>>109454222

the best at sucking big fat cock? Yes.

最擅长含大屌?没错。

I have to admit, it can be a fun hobby, but it's just not for everyone.

我得承认,它确实能是个有趣的爱好,但并不是每个人都适合。

Good three times repeating numbers

三个数码反复出现,挺好。

#228 No.109454340
Anonymous · 2026-08-04 02:38

>>109454222

C# is the better lang but

C#是更好的语言,但……

>Microslop

Can't believe Oracle, the holder of 6 million datacenters, is not enslopping the JDK.

不敢相信Oracle,拥有六百万个数据中心的主,居然不把JDK圈在自己怀里。

#229 No.109454736
Anonymous · 2026-08-04 03:32

>>109439877

Fine fuck it. I'm rolling it. But I'm going to do that in Unreal Engine because I'm trying to learn it.

行吧操了。我决定干了。但我要用虚幻引擎来做,因为我在学它。

also can't you anons update this with actually good challenges? Should it pretty easy to do with AI around.

还有,你们这些anon能不能更新点真正有挑战性的题目?有AI在应该很容易吧。

#230 No.109454741
Anonymous · 2026-08-04 03:33

>>109451868

Oh wow, the day before's stome game is fucking dumb where you just return True because Alice always wins.

哇哦,前天的石头游戏题蠢得一逼,直接返回True就行,因为Alice永远赢。

#231 No.109454893
Anonymous · 2026-08-04 03:55

>>109454736

>also can't you anons update this with actually good challenges?

>还有,你们这些anon能不能更新点真正有挑战性的题目?

You could do it too

你也能做啊。

#232 No.109454923
Anonymous · 2026-08-04 04:00

I want to learn how to program so I can eventually make a game in unity

我想学编程,这样以后能在Unity里做个游戏。

Are there any good C# learning resources? I liked learncpp and how that was set up, Microsoft’s resource for learning c# looks like it was designed by someone smoking meth

有没有好的 C# 学习资源?我挺喜欢 learncpp 那种设置方式,微软的 C# 学习资源看起来像是嗑药的人设计的。

#233 No.109454945
Anonymous · 2026-08-04 04:02

>>109454923

Why not ask claude?

为啥不问问 Claude 呢?

#234 No.109455260
Anonymous · 2026-08-04 04:42

>>109454923

You'll end up making a free porn game, and people will roast you for it, and then you'll hang from your ceiling.

你最后会做出一个免费的色情游戏,然后大家会把你骂得体无完肤,接着你就得在天花板上吊着了。

Happens every single time.

每次都是这样。

#235 No.109455337
Anonymous · 2026-08-04 04:54

>>109435851

Building a fitness app as a desperate play to get employed based on its awesomeness.

做个体脂管理应用,当作孤注一掷的求职手段,就靠它的牛逼来吸引人。

#236 No.109455370
Anonymous · 2026-08-04 04:58

>>109455337

>as a desperate play

>孤注一掷

Underrated statement of the century.

这世纪最被低估的一句话。

If you think that's what we need, then you're fucked and should commit suicide.

你要是觉得这才是我们需要的,那你完了,赶紧自杀去吧。

Sorry for your ego, but sometimes you just have to do things for the greater good.

抱歉伤了你自尊,但有时候为了更大的善,你就得这么做。

#237 No.109455406
Anonymous · 2026-08-04 05:04

>>109455370

lol haha

哈哈,呵呵。

#238 No.109455407
Anonymous · 2026-08-04 05:05

>>109455406

I'm not here to amuse you, I'm here to take out the trash.

我不是来逗你笑的,我是来清垃圾的。

#239 No.109455425
Anonymous · 2026-08-04 05:07

>>109455407

aha

#240 No.109455442
Anonymous · 2026-08-04 05:10

>>109455425

Yes.

#241 No.109455456
Anonymous · 2026-08-04 05:13

>>109455407

Waste bin’s in the corner janny, hop to. Yo wait, one more thing

垃圾桶就在角落里,保洁员,赶紧动手。哎等等,还有件事——

>You turn

>你转身

Bankshot!

>A wad of paper hits you on the forehead

>一团纸砸在你额头上

>Bounces off the rim of the basket in your hands

>从你手里的篮子边缘弹开

Ah damn close one, mind grabbing that for me? Thanks man!

哎哟差点进了,帮我捡一下呗?谢啦兄弟!

#242 No.109455473
Anonymous · 2026-08-04 05:17

>he can only think of one type of trash

>他脑子里就只想着一种垃圾

Remember: evidence of autism => refusal of direct engagement.

记住:自闭症证据 => 拒绝正面交锋。

#243 No.109455495
Anonymous · 2026-08-04 05:21

Hey I’m engaged to you, relax mang it ain’t no thang to get twisted ‘bout

嘿,我可是跟你订婚的人,放轻松啦,没啥好纠结的。

Also I can also think of paper plates also

而且我还想得到纸盘子呢。

#244 No.109455498
Anonymous · 2026-08-04 05:22

>>109455473

#245 No.109455501
Anonymous · 2026-08-04 05:22

>>109455495

#246 No.109455508
Anonymous · 2026-08-04 05:23

>>109455498

#247 No.109455511
Anonymous · 2026-08-04 05:24

>>109455508

#248 No.109455518
Anonymous · 2026-08-04 05:25

Exactly!

#249 No.109455529
Anonymous · 2026-08-04 05:27

Indeed.com

#250 No.109455535
Anonymous · 2026-08-04 05:28

>>109455518

#251 No.109455552
Anonymous · 2026-08-04 05:31

Please try to say on topic. sir. What kinda programming u working on? I’m building a fitness app myself it’s awesome but I digress

请尽量别偏题,先生。你在搞哪种编程?我自己也在做个健身应用,超级棒,不过跑题了。

#252 No.109455556
Anonymous · 2026-08-04 05:32

>>109455370

#253 No.109455563
Anonymous · 2026-08-04 05:34

>>109455556

Whoah ur making a fitness app too?? Any pointers?

哇,你也在做健身应用??有啥指点不?

#254 No.109455567
Anonymous · 2026-08-04 05:34

>>109455556

#255 No.109455578
Anonymous · 2026-08-04 05:36

>>109455567

Don’t be so hard on urself u have value and r loved

别对自己那么狠,你有价值,有人爱你。

#256 No.109455581
Anonymous · 2026-08-04 05:37

>>109455370

>you're fucked and should commit suicide.

>你完了,赶紧自杀去吧。

#257 No.109455588
Anonymous · 2026-08-04 05:38

>>109455581

No don’t say that baby u got lots to live for who’s gonna finish ur app otherwise

不,宝贝别这么说,你有大把活下去的理由,不然谁来完成你的应用呢?

#258 No.109455594
Anonymous · 2026-08-04 05:39

>>109455581

Actually, y'know what, I take it back, the government should publicly, painfully, and brutally cull you in front of national TV, and not just you, but your entire family. We literally CANNOT let scum like yourself roam the Earth; we have to make merely BIRTHING defects like that a crime deserving of the death penalty for at least two generations.

其实,知道吗,我收回那句话。政府应该在全国电视直播前,公开、痛楚、残忍地把你处决了,不光是你,还有你全家。我们真的不能让你这种渣滓在地球上乱晃;我们得让“生出这种缺陷”本身就成为死罪,至少惩罚两代人。

#259 No.109455609
Anonymous · 2026-08-04 05:42

I’m thinking the twist, cuz u always need a twist see, the twist is what gets people interested, the twist for my app is gonna be that it’s got like an internal database of all sort of exercises and have room for custom ones and there’s be internal weighting attributes to each for what muscle groups they target to what intensity so that ultimately u can drag n drop exercises u like n the app will recommend others to fill in gaps and hit all ur target groups so you got more control and variability and choice to ur workouts while keeping it more relevant to what u want to build rather than just what’s recommended to u u know?

我在想那个转折点,因为你总得有个吸引人的点,对吧,转折点才是让人上头的地方。我的应用的转折点就是它内置一个运动数据库,各种动作都有,还能自定义添加,每个动作都有内部权重属性,对应练哪些肌肉群、强度多高,最后你可以拖拽你喜欢的动作,应用会推荐其他动作来补缺口,覆盖所有目标肌群,这样你就能更自由地控制、有更多变化和选择,同时更能贴合你想练的东西,而不是只按推荐来,你懂吧?

#260 No.109455628
Anonymous · 2026-08-04 05:45

I also wanna do a model view of a human body n when u add exercises it blights the muscles being targeted with a scaling hue determined by intensity of ur exercise stack so u can visualize what ur hitting and immediately eyeball untouched areas but that’s currently outa scope

我还想做个 3D 人体模型视图,你添加动作的时候,目标肌肉会按你运动强度的比例变色高亮,这样你能直观看到在练哪儿,一眼看出没练到的地方,不过这个目前超出范围了。

#261 No.109455634
Anonymous · 2026-08-04 05:46

>implying I'm gonna read these worthless walls of text

>暗示我会去读这些毫无价值的废话长文

Guess I'm gonna take care of the dishes.

看来我得去洗碗了。 ┐ 你还没捡那团纸呢

#262 No.109455637
Anonymous · 2026-08-04 05:47

You still haven’t picked up that wad of paper tho

可你还是没捡起那团纸啊

#263 No.109455645
Anonymous · 2026-08-04 05:49

>>109455473

#264 No.109455747
Anonymous · 2026-08-04 06:10

>>109436028

Damn that felt good

妈的,那感觉真爽

God said sports are homo. Tackle a horse.

上帝说了,运动挺基的。去撞匹马试试。

图片: https://i.4cdn.org/g/1785823859949798.png

#265 No.109455769
Anonymous · 2026-08-04 06:15

>>109455747

>doesn't say what amount of calories are lost

>没说消耗了多少卡路里

#266 No.109455785
Anonymous · 2026-08-04 06:17

>>109455769

Why would I care about that? I am not a professional. I don't care about my watts or calories. I might get a heart rate monitor soon so that I do not overpace as often, but in general I just ride so that it feels good and is fun

我干嘛要在乎那个?我又不是专业的。我不关心我的功率或卡路里。我可能很快会弄个心率监测仪,免得老是起速太快,但总的来说我就是骑到爽、骑到好玩

#267 No.109455792
Anonymous · 2026-08-04 06:18

>>109455769

>>109455785

PS: I just calculate with rougly 500 calories per hour when I do long tours. For 2 hours riding it doesn't fucking matter

补充:我骑长途的时候就大概按500卡每小时算。骑两小时根本无所谓

#268 No.109455804
Anonymous · 2026-08-04 06:20

>>109455785

>I am not a professional.

>我不是专业的。

Not about being a professional; it's about being fucking overweight.

跟专不专业没关系;是跟胖得要死有关系。

#269 No.109455815
Anonymous · 2026-08-04 06:21

>>109455563

*foo;

#270 No.109455833
Anonymous · 2026-08-04 06:25

>muh pointers

>我的指针

Holy.

图片: https://i.4cdn.org/g/1785824719180631.png

#271 No.109455938
Anonymous · 2026-08-04 06:51

>>109455833

Never understood why people hate them. Pointers are super simple to understand. There is really nothing complex about it

一直不懂为啥有人恨它们。指针超简单的好吧。真没什么复杂的

>but muh pointer arithmetic.

>但我的指针运算啊。

It's the same fucking thing

不都一个屌样

A + B = C means something entirely different depending on what you add. Is it scalars, vectors, matrices or just fucking pointers. It's the same operator for fundamentally different operations. Deal with it

A + B = C 根据你加什么,意思完全不一样。是标量、向量、矩阵,还是就他妈是指针。本质不同的操作共用一个运算符。学着接受吧

#272 No.109455953
Anonymous · 2026-08-04 06:55

>it thinks this is about hate

>它以为这是仇恨的问题

>and not about autismic interpretation of words

>而不是对字词的自闭式解读

>inb4 I don't have a diagnosis I'm not autistic

>先说了我没诊断过,我不是自闭

#273 No.109455967
Anonymous · 2026-08-04 06:58

>>109455953

I think you forgot to quote me, but I guess you mean my forum post.

我想你忘引用我了,但我猜你指的是我的帖子。

You're absolutely right! I am not autistic

你说得完全对!我不是自闭

#274 No.109455971
Anonymous · 2026-08-04 06:59

>>109455967

>67

SIX SEVEN!!!! SIX SEVEN!!!!

HAHAHA SIX SEVEN

图片: https://i.4cdn.org/g/1785826771910428.gif

#275 No.109455986
Anonymous · 2026-08-04 07:03

Y'know what? At this point I'm almost convinced that >>109450502 is right.

你知道吗?到这一步我几乎相信 >>109450502 说得对了。

#276 No.109456047
Anonymous · 2026-08-04 07:18

>>109455986

He's partly right. But there are also autists that are straight up retarded and cannot get jack shit done in their entire life.

他说对了一部分。但也有些自闭患者纯粹是脑子不行,一辈子啥事都干不成。

Maybe life's a bit more complex, huh?

也许生活复杂一点,是吧?

#277 No.109456071
Anonymous · 2026-08-04 07:23

>life's a bit more complex

>生活复杂一点

Nah, they're literally enabled for the worst of humanity.

拉倒吧,他们基本上是被纵容成人类最坏的那一面了。

#278 No.109456094
Anonymous · 2026-08-04 07:30

>>109455938

the general concept of "it's just an address" is simple, sure

“它就是个地址”这个总体概念当然简单

the problem is how much unnecessary complexity it can add to a program once you start writing anything non-trivial - it's as bad as a wrongly applied design pattern in Java, except with pointers the level of abstraction goes lower, so it's more difficult to figure out what was the purpose of the code

问题是一旦你开始写任何不琐碎的东西,它能给程序加多少不必要的复杂度——就跟Java里乱用设计模式一样糟,只不过指针把抽象层级拉得更低,所以更难搞明白代码的意图是什么

#279 No.109456102
Anonymous · 2026-08-04 07:32

>>109449244

No it doesn't. These challenges never made sense. It was just for pure rolling and spaming this thread

不会的。这些挑战从来就没道理。纯粹就是为了刷帖和灌水

#280 No.109456172
Anonymous · 2026-08-04 07:47

>>109456102

>>109450241

#281 No.109456533
Anonymous · 2026-08-04 09:07

>>109439887

Ok, I am literally doing it now. Learning all of it again, even I heard most of it already. Found a book on my hard drive and will now work through it

行,我他妈的现在就在做。从头再学一遍,就算大部分都听过。在硬盘上找到一本书,现在开始啃

> The function in the preceding equation is the activation function we are applying on

> 前面方程里的函数是我们应用的激活函数

> top of the summation so that we attain non-linearity (we need non-linearity so that our

> 加在求和之上,以获得非线性(我们需要非线性,这样我们的

> model can now learn complex patterns).

> 模型才能学习复杂模式)。

This is new to me for example. Neat.

这个我以前没接触过。挺酷的。

#282 No.109456671
Anonymous · 2026-08-04 09:46

>complains about the font

>抱怨字体

>cannot make up the mind about what font does it

>连用什么字体都拿不定主意

You retarded worthless autists are so fucking worthless holy shit. We, as in, neurotypicals and those who can be treated into neurotypicality, really have to extinguish your ilk. Forget race wars, welcome brain wars.

你们这些毫无价值的自闭症废物是真他妈没用,天哪。我们,也就是神经典型者以及那些能通过治疗变成神经典型的人,真的得把你们这类人灭掉。别扯什么种族之战了,欢迎来到大脑之战。

图片: https://i.4cdn.org/g/1785836784304891.png

#283 No.109456701
Anonymous · 2026-08-04 09:54

>>109456671

I have no idea what the author wants to say with those lyrics. They sound kinda pointless

我完全不懂作者想用那些歌词表达什么。听着挺没意义的。

#284 No.109456737
Anonymous · 2026-08-04 10:03

That's because you're autistic and don't deserve further consideration.

那是因为你有自闭症,不值得再被考虑。

#285 No.109456776
Anonymous · 2026-08-04 10:16

>>109456737

No for real. You keep spaming some weird compiler outputs here, but what's your goal behind it? What point are you trying to make?

说真的,你在这儿不停刷一些古怪的编译器输出,但你背后的目的是啥?你想论证什么点?

Compilers are not perfect? I think we knew that before. What's your goal for self-improvement?

编译器不完美?这我们早知道了吧。你自我提升的目标是什么?

#286 No.109456786
Anonymous · 2026-08-04 10:19

>>109456776

>but what's your goal behind it?

>但你背后的目的呢?

Finally!

What's my goal is for people to be *able* to demand better, with actual arguments. And I don't just mean

我的目的是让人们有能力,用实际论据去要求更好的东西。我指的不只是

>programmers

, I mean actual fucking users. Users should be able to expect that everyone between the source code writing and them executing the code have done their literal best to make sure that as little lifetime is wasted as possible.

,我指的是真正的用户。用户应该能期待,从写源码到他们执行代码之间的每个人,都尽了自己最大的努力,确保尽可能少浪费生命。

>I think we knew that before.

>这我们早知道了吧。

Who is "we".

#287 No.109456829
Anonymous · 2026-08-04 10:30

>>109456786

Also this is my strongest argument for why I'm not autistic, aside from everything else. I *used* to be on low-budget hardware. I *know* what it feels like. And understanding *why* yesteryear's software was so slow - i.e. the autistic, senseless pointlessness of it all - is what's making me so angry that I am totally fine sacrificing all of autisty.

这也是除了其他一切之外,我证明自己不是自闭症的最有力论据。我曾经用过低端硬件。我知道那是什么感觉。而理解为什么过去的软件那么慢——也就是说,那种自闭症式的、毫无意义的蠢——正是让我愤怒到完全愿意牺牲所有自闭特征的原因。

>in contrast to humanity

>对比人类

#288 No.109456904
Anonymous · 2026-08-04 10:45

>>109456786

>with actual arguments.

>用实际论据。

What arguments do you mean? Just look at the code, that's just some constructed piece of garbage. No user would write that

你指的是什么论据?直接看代码,那不过是堆构造出来的垃圾。没有用户会写那种东西。

#289 No.109456930
Anonymous · 2026-08-04 10:51

>>109456904

>What arguments do you mean? Just look at the code

>你指的什么论据?直接看代码

...

Yes. Yes, just look at the code. You don't have to be a brain genius; just go around pester the retards responsible for your compiler output, and publicly boycott them if they refuse. We've got Twitter/formerly X, for fucks sake. We can *shame* companies into bankruptcy.

对。对,就看看代码。你不需要是天才;就去烦那些该为你的编译器输出负责的白痴,如果他们拒绝改,就公开抵制他们。咱有推特/前X,妈的。我们能靠羞辱让公司破产。

#290 No.109456940
Anonymous · 2026-08-04 10:53

>>109456930

So what's your point? Too many assembly instructions? The """wrong""" ones? What do you want to express with your art?

那你到底想表达啥?汇编指令太多?还是“错误”的那些?你想用你的艺术表达什么?

#291 No.109456944
Anonymous · 2026-08-04 10:54

>>109456940

>why not both? ™

>为啥不两者都要?™

#292 No.109456957
Anonymous · 2026-08-04 10:58

>>109456944

Because almost everything has a trade-off? So again: What's your point?

因为几乎每件事都有权衡取舍?所以再问一次:你的点是什么?

#293 No.109456970
Anonymous · 2026-08-04 11:02

>>Because almost everything has a trade-off?

>>因为几乎每件事都有权衡取舍?

>he doesn't know

>他根本不懂。

Go learn assembly, then come back.

去学汇编,然后再回来。

Alternatively, just don't argue against your betters, like normies.

或者,就像普通人那样,别跟比你强的人抬杠。

#294 No.109456972
Anonymous · 2026-08-04 11:03

>>109456970

Just assume that I know assembly. So now the hurdles are cleared and you can explain what your point is

就假设我懂汇编吧。那障碍清除了,你现在能解释你的点了吗?

#295 No.109456981
Anonymous · 2026-08-04 11:05

>>109456970

Also picrel.

还有附图。

>but muh tradeoffs

>但我的权衡取舍啊

have been the autistic excuses of autistic autists for literal decades, and I'm tired of pretending otherwise. Learn how fucking hardware works, or be MADE to stop programming; I don''t care which.

这话都成了自闭症患者几十年来最经典的自闭借口了,我懒得再装作不是。要么学学硬件到底怎么运作,要么就别再编程了;哪个我都不在乎。

图片: https://i.4cdn.org/g/1785841509293545.jpg

#296 No.109456988
Anonymous · 2026-08-04 11:06

>>109456972

>Just assume that I know assembly.

>就假设我懂汇编。

If you did you'd know that shit like, I dunno, loop-invariant code hoisting was autistically retarded. Now don't speak to me or my son ever again: >>109456981

你要真懂的话,就会知道像循环不变代码提升那种东西,简直是自闭得离谱。现在别再来骚扰我或我儿子了:>>109456981

#297 No.109457021
Anonymous · 2026-08-04 11:15

>>109456988

>loop-invariant code hoisting was autistically retarded.

>循环不变代码提升简直是自闭得离谱。

Why do you think that and what is your solution to that?

你为啥这么想?你又有啥解决办法?

>>109456981

Ok and what if some person doesn't want to know how the hardware works, because they just want to ship an android app? Why does everyone has to know everything? Do you know all of maxwell's equations and what they do?

行,那要是有人压根不想懂硬件咋回事,就想着赶紧把安卓应用发出去呢?为啥每个人都得啥都懂?麦克斯韦方程组你全会吗,知道它们干啥用的不?

If not, then you should shut down your computer and rip all the electricity out of your house

要是不会,那你该关电脑,把家里电线全扯了

#298 No.109457042
Anonymous · 2026-08-04 11:17

>>109457021

>Why do you think that

>你为啥这么想

>Learn how fucking hardware works, or be MADE to stop programming; I don''t care which.

>学学他妈的硬件怎么运转的,不然就逼你停止写代码;我不管你选哪个。

I'll give you one hint though: line-fill buffers.

不过我给你提个醒:行填充缓冲区(line-fill buffers)。

>what if some person doesn't want to know how the hardware works

>那要是有人不想懂硬件怎么工作咋办

>>109456970

>Alternatively, just don't argue against your betters, like normies.

>或者,别跟比你强的人抬杠,像普通人那样就行。

I cannot believe how I'm pants-on-head´drunk for this interaction, during a heatwave, without AC, and still winning: >>109450733

我真不敢相信,我热浪天没空调,脑子都快糊了,跟人吵这架还能赢:>>109450733

#299 No.109457071
Anonymous · 2026-08-04 11:22

>>109457042

>I'll give you one hint though: line-fill buffers.

>我给你提个醒:行填充缓冲区。

I don't need hints. I want an explanation from you. Why are you so frustrated and so desperate about finding super tiny errors in assembly code that have no measurable effect for the very most applications?

我不用提示。我要你解释清楚。你为啥这么急眼,非得揪着汇编代码里那点小到对绝大多数应用毫无影响的错误不放?

#300 No.109457081
Anonymous · 2026-08-04 11:24

>super tiny errors in assembly code

>汇编代码里那点小错误

Dude.

Duuuuuuuuuuuuude.

>just don't argue against your betters, like normies.

>别跟比你强的人抬杠,像普通人那样就行。

#301 No.109457163
Anonymous · 2026-08-04 11:37

>>109457081

What do you mean? Can you elaborate?

你啥意思?能展开说说吗?

#302 No.109457197
Anonymous · 2026-08-04 11:40

How. Much. More. Literal. Do. You. Need. Me. To. Be.

我.还.要.说.得.多.直.白.你.才.能.懂.

>just don't argue against your betters, like normies.

>别跟比你强的人抬杠,像普通人那样就行。

Normies can accept the reality lined out by experts, so why can't you.

普通人能接受专家划定的现实,你咋就不行呢?

No, that's not a question.

不,我不是在提问。

#303 No.109457296
Anonymous · 2026-08-04 11:56

do not listen to theorists. they always criticize everything but never produce or deliver anything

别听那些理论派的。他们啥都挑刺,可啥也做不出来,啥也交付不了

even a pajeet writing overengineered AI-slop is a better programmer than someone who "knows how things work" but never produces anything used by anyone

就算是个写过度工程AI垃圾的outsource佬,也比那种“懂原理”但做出来的东西谁都用不上的人强

shitty code that actually does something productive is always better than the most beautiful and efficient code of a program that has no practical use for anyone

能出活儿的烂代码,永远比写得再漂亮再高效但没人用得上的程序强

#304 No.109457307
Anonymous · 2026-08-04 11:57

>they always criticize everything but never produce or deliver anything

>他们啥都挑刺,可啥也做不出来,啥也交付不了

Agreed. So where's the registry dumper?

同意。那注册表导出工具呢,在哪儿?

#305 No.109457313
Anonymous · 2026-08-04 11:58

You see? This is what algol and algol accessories will do to your brain. Endless fights in this autistic thunderdom over inefficiencies introduced by inadequate languages.

你看吧?这就是algol和algol衍生语言对你脑子干的好事。在这破圈子里为了语言低效引起的破事没完没了地吵。

#306 No.109457332
Anonymous · 2026-08-04 12:02

>muh algol

>我的ALGOL啊

I don't even know what that is, but nice revealing your autism-projecing autim

我压根不知道那是啥,不过你倒挺会暴露自己自闭还投射别人的毛病

#307 No.109457439
Anonymous · 2026-08-04 12:15

>>109457197

I don't understand what the artist is trying to say here. What's your point and what is your exact problem?

我不懂这位老哥想说啥。你啥观点,具体问题又是啥?

can't you just be happy?

你就不能消停点高兴会儿吗?

>>109457296

Yes, I know, but they must have some reason to get THIS mad, no??

行吧我知道,但他们总得有理由气成这样吧,对吧??

>>109457307

What registry dumper? What do you want to achieve with that?

啥注册表导出工具?你想用它干啥?

>>109457332

>I don't even know what that is

>我压根不知道那是啥

See. I think that is your problem. Your are so deep in gcc's asshole, that you don't see anything else. If you'd open your eyes (and ignore javascript), then you'd realize that the world ain't THAT bad really. Compilers work just fine. They get improved bit by bit, but in general they do acceptable work

看吧。我觉得这就是你的问题。你陷在gcc的菊花里太深了,别的啥都看不见。你要是睁开眼(别管javascript),就会发现世界也没那么糟。编译器跑得好好的。一点一点在改进,但总体干得活是能用的

#308 No.109457783
Anonymous · 2026-08-04 13:05

(无正文)

图片: https://i.4cdn.org/g/1785848704457172.png

#309 No.109457891
Anonymous · 2026-08-04 13:22

>>109457783

Wow, AI is very good these days

哇,现在AI真厉害了

#310 No.109458640
Anonymous · 2026-08-04 15:07

Good morning~

早上好~

…I see you’ve still left that paper on the floor.

我看你还是把那张纸留在地上了。

#311 No.109458815
Anonymous · 2026-08-04 15:34

>>109457296

"productive" is a fictional concept.

“高效”是虚构的概念。

if you aren't catching fish, you're literally just a parasite on society no matter what you do.

你要是不抓鱼,不管干啥,说白了就是社会的寄生虫。

← 上一篇返回列表下一篇 →

(如果你觉得这篇文章有启发,可以点击这里付费