K.I.S.S

ZTNA developer & preacher; experienced in 2C&2B biz practice;
1000x engineering with supergrok heavy

[K.I.S.S @shangjiyu]

是的，所以我决定把IDEA GIT天赋带到 VSCODE 😂 https://t.co/WaMxcKsDhN

[Viking @vikingmute]

在经过了长期的思想斗争以后，现在正式不用 IDE 了，Cursor 也正式切换成了 Agent View，Zed 也很少打开了，代码只关注架构，每次简单浏览 Diff 了，Code Review 交给多个 Agent 互相 review，验证就是完备的 E2E 了，未来代码就是黑盒了，这是心酸但是没办法的事实了。不知道大家现在还看代码吗？

[Zartbot @zartbotF]

OpenAI Jalapeno deep div analysis , I believe someone should forward to @JensenHuang :)

https://t.co/XVAAPe0bNM

[K.I.S.S @shangjiyu]

@ClashbyHako 快点上流量统计桑基图 必爆😂

[K.I.S.S @shangjiyu]

@Hugo_Qin 自动驾驶一是双，一直自动一直爽？ GIT 特别是 DIFFER 能力还是特别需要的，现在完全还是alpha 阶段，大家一起来呀

[K.I.S.S @shangjiyu]

是的，所以我决定把IDEA GIT天赋带到 VSCODE 😂 https://t.co/WaMxcKsDhN

[Viking @vikingmute]

在经过了长期的思想斗争以后，现在正式不用 IDE 了，Cursor 也正式切换成了 Agent View，Zed 也很少打开了，代码只关注架构，每次简单浏览 Diff 了，Code Review 交给多个 Agent 互相 review，验证就是完备的 E2E 了，未来代码就是黑盒了，这是心酸但是没办法的事实了。不知道大家现在还看代码吗？

[Progyan @plugyawn]

I spent the last six months trying to deconstruct Taalas's patents. We think it can be better: 100x fewer memfetches than their bitROM, better software by better quantization than their hardware team could.

So we wrote a compiler that:
> takes a huggingface checkpoint,
> quantizes the model
> descends the weights down metal layers to RTL and GDS, do your DRC, Yosys, PEX, make the electrical waveform execute a mat-vec from your huggingface checkpoint (thanks Cambricon tech papers)

in ~7000 lines of human-readable code.

we're looking for someone with contacts with a foundry/access to 7nm PDKs or contacts at a cheap EuroPTW shuttle? I'm broke and unemployed.

@itsclivetime this is what i think is the future of perplexity per picojoule. @zerohedge @zephyr_z9 you guys wanna see 40,000 tokens/sec?

[K.I.S.S @shangjiyu]

@imsoff 现阶段卖任何年付都是不明智的😜

[K.I.S.S @shangjiyu]

@wquguru 辅助驾驶不是自动驾驶哦 😂

[Kimi.ai @Kimi_Moonshot]

Releasing the model weights and technical report of Kimi K3.

Kimi K3 is our most capable model: a 2.8T MoE model with native visual understanding and a 1M-token context window.

New model architecture: 2.5x the intelligence per unit of compute, not just more params.

Alongside Kimi K3, we're opening up more of the stack behind it — high-performance attention kernels, MoE communication library, and infrastructure for running agent environments at scale.

Model weights: https://t.co/7m7eEg6Y0B
Tech report: https://t.co/yeu6cjpMCT
Tech blog: https://t.co/YTfiMSNM1f

[K.I.S.S @shangjiyu]

@plantegg 18个小时 完全恢复

[K.I.S.S @shangjiyu]

@xiaohu 华为小艺也说他满分

[Elon Musk @elonmusk]

Grok 4.5 is not yet using our internally developed C/C++ inference software that exact maps to the GB300 hardware. Doubling or more of the current speed is probably achievable.

[K.I.S.S @shangjiyu]

@mranti 矛盾的存在是必然的，社会发展就是由矛盾推进的。

[K.I.S.S @shangjiyu]

@HiTw93 支持

[42HOW @42how_]

LI Auto's MindVLA-o1 model addresses VLA limitations with multi-modal MoE Transformer architecture, 3D ViT encoding, enhanced multi-modal reasoning, predictive modeling, and integrated hardware-software design for improved efficiency and adaptability.

[Elon Musk @elonmusk]

This will be big

[K.I.S.S @shangjiyu]

@KKaWSB 碳基智能 和 硅基智能 核心差异是良心

[K.I.S.S @shangjiyu]

@st7evechou vibe coding 之后，这些个软件开发门开被撸平了😂

[Geek Lite @QingQ77]

请全体起立，今天给你们看一个真正的硬核狠活，超级王炸的项目

ANE：绕过限制，直接用 Apple Neural Engine 跑神经网络训练（PoC）

苹果一直把 ANE 当作纯粹的推理加速器锁在线上，但这个项目通过逆向工程证明了一点：ANE 其实完全有能力跑反向传播。

这老哥直接暴力逆向了苹果的私有 API（_ANECompiler 是什么鬼？拿来吧你！），硬生生把反向传播（Backprop）塞进这块被封印的硅片里跑通了。没有 CoreML，没有 Metal，不带 GPU 玩，纯靠 ANE 搓出了一个 Transformer 训练循环。
https://t.co/5BCf9TiuPR

[K.I.S.S @shangjiyu]

@cccchuizi @kiipela 抽