wang

Creating the computing engine for next era. Cloud | Backend | Frontend | AI Infra

[wang @wangzhr4]

@feneko_san Vibe coded?

[wang @wangzhr4]

Claiming "AI has killed software engineering" is just a hilarious exposure of one's ignorance.

[wang @wangzhr4]

Shanghai, Nanjing Road.

[wang @wangzhr4]

To anyone who claims "UDP is not about connection", and TCP is "connection-oriented":

You are not understanding networking yet. Please dive deeper.

[wang @wangzhr4]

"Connection" is just an illusion. Because under the hood, nothing persists. You do not actually hold a dedicated connection in the transport even with TCP which has been claimed as "connection-based". What you hold in TCP peers are contexts.

And you can do that within UDP too.

[SudoX7 @sudox7]

i made a UDP client-server in C using raw sockets

and it cleared up so much confusion about what "connectionless" actually means

[wang @wangzhr4]

没必要听 A 厂虚头巴脑的说什么“AI coding”已经全面替代了软件工程师。AI 如果能自主开发以下任何一个即可：

1. 一个可用的浏览器内核
2. 一个可用的 C 语言编译器
3. 一个可用的 Python 解释器
4. 一个可用的 Java 运行时

这个要求对于鼓吹“AI 已经让程序员下岗”的人，不管高吧？

[wang @wangzhr4]

Anthropic 是懂营销的。一会号称自主发现了数十个 Linux 内核漏洞，一会儿又号称自主发现了数十个致命的网络攻击漏洞。那么，请你公布漏洞啊，这又不是什么新鲜事，本来科技行业就是一直在修漏洞，CVE 这么多年来也是公开的。与其神神秘秘的说“我的模型能威胁全世界的安全”，不如按照行业规矩，把漏洞、发现过程和修复建议都公布出来，也好让大家看看成色如何。

[wang @wangzhr4]

Metadata can be expensive. This is not magical, it created 2 source of truth: the medatata and the actual body - the string array.

This is not a silver bullet, and there are tradeoffs.

[trish @TrisH0x2A]

Redis does not use C strings

antirez wrote a custom string library called SDS: Simple Dynamic Strings

it later became a standalone library

C strings had three problems Redis could not accept

- strlen is O(n) because it scans until the null terminator
- strcat needs manual capacity management and often scans the string again internally
- and C strings cannot safely store arbitrary binary data because null bytes end the string

SDS fixes all of this by storing metadata in a header placed right before the string buffer

[wang @wangzhr4]

Anthropic 是懂营销的。一会号称自主发现了数十个 Linux 内核漏洞，一会儿又号称自主发现了数十个致命的网络攻击漏洞。那么，请你公布漏洞啊，这又不是什么新鲜事，本来科技行业就是一直在修漏洞，CVE 这么多年来也是公开的。与其神神秘秘的说“我的模型能威胁全世界的安全”，不如按照行业规矩，把漏洞、发现过程和修复建议都公布出来，也好让大家看看成色如何。

[wang @wangzhr4]

It is because there is no trailing byte or signal for arrays. Array is a simple space in the memory, and there is no way to get their tails.

[はるはる @haruharu_RQA]

C言語、配列の長さを返す関数がないのガチか

[wang @wangzhr4]

Open source is dead. AI is able to generate thousands lines of codes and flood into any project. How come the classic open source project governance handles this?

[wang @wangzhr4]

@skywind3000 @Poison_2_ 但是 Apple 是 OpenCL 的发起者啊

[wang @wangzhr4]

Yuanshen Stadium.

[wang @wangzhr4]

night at Shanghai.

[wang @wangzhr4]

@schteppe Overloading is a sin. Thank god C doesn't support that.

[wang @wangzhr4]

@brankopetric00 @SaadInCyber Whenever I hear anyone who claims "I have stored the credentials in the K8s secrets", I always get pissed off. How come one is not aware of the difference between encryption and encoding?

[wang @wangzhr4]

This is not even an apple-to-apple comparison.
Before posting shit, you really need to understand the very basic difference between stack and heap, between static and dynamic.

[SudoX7 @sudox7]

your "clean" C++ is slower than this "ugly" C

[wang @wangzhr4]

以前叫：中间人攻击
现在叫：Token 中转站

[wang @wangzhr4]

@goodhunt @RelVM2nd 这个翻译可真是太“地道”了。

[wang @wangzhr4]

@tison1096 我觉得结构化输出反倒是不必要的，毕竟这些 utils 将来都是 AI 在用，结构不结构化的无所谓的。只要在语义上是合理的，就行了。

[wang @wangzhr4]

@KELMAND1 吃相太难看，水平太低。