Franz
1.8K posts
Franz
@curliph
ASIC designer, Technophile, CTO of me, speak for myself here.
22.563561,113.872353 Katılım Kasım 2009
830 Takip Edilen93 Takipçiler
Zeyad S. designed a fully custom ATmega328PB board to serve as a compact, flexible development platform for prototyping embedded projects: linkedin.com/feed/update/ur…
English
Franz retweetledi
@RJDAIGOGO 诺奖有一项考量是要bring greatest benefit to humankind)。但这一项也不是必须条件，基础科学本身就会对人类了解自身的世界带来正向的利益。但诺奖确实要看到这种基础理论所带来的实际影响。邓的证明，其实假定了一个非常理想的初始条件，还有对粒子物理属性的限制。但我相信，邓也不会就此打住
中文
我有一个关于开源ai的问题，苦思不得其解，只好发出来让大家一起苦思：
如果很多人都能通过来源模型拥有制造生物武器、3d打印攻击无人机、网络攻击、或其他大规模损害/杀伤的能力，世界将变成怎样？
（btw，我并非支持闭源，闭源的结果同样甚至更加恐怖，不过希望了解大家对这个问题的的看法？）
中文
@Mononofu @JensenHuang it's free for everyone to embrace open source world. the spirit of open source is freedom. open or not is free for everyone. it's welcome for everyone to join open source community, but we should never ask someone to open source. So, Welcome Anthropic! 🥳
English
I’m so excited that @JensenHuang is a believer in open source now, looking forward to the CUDA and GPU driver open source release!
Jensen Huang@JensenHuang
For my first post, I’m sharing a letter @NVIDIA signed on why open models matter. AI will transform every industry, power every company, and be built by every country. Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty. The world needs both frontier closed models and frontier open models. images.nvidia.com/pdf/Open-Weigh…
English
“I want to understand, not just in a mathematical way, the ideas in all branches of theoretical physics” - Feynman.
Mathematics is beautiful, give you faith. But that is a very different way to understand the real physics. If you just keep that in mind, you will get far away.
English
@7011Huazhong @_FORAB shoot from our tmsc 16nm design rule, unit (um). And more, Quantum Tunneling take effect under 5nm. for 7nm~16nm, that not main problem, even w/o high-k
English
中文
英文区研究机构，发现华为麒麟 9030 芯片，经电子显微镜测量，其最小金属间距为 32.5 纳米，该间距比英特尔最新 18A 工艺节点，约 36 纳米紧密约 10%。
他们产生疑问，在没有 EUV 光刻机的情况下，它们能把这些芯片量产出来，究竟发生了什么？
SemiAnalysis@SemiAnalysis_
We cut open the Kirin 9030 and put it under an electron microscope. The smallest metal pitch measures 32.5 nanometers. That is tighter than Intel 18A, their brand new leading edge node. A Chinese fab with no EUV is out-pitching Intel's EUV node by roughly 10 percent. What's going on here? 🤯 "This is the HiSilicon Kirin 9030, the chip inside Huawei's newest flagship phone. A few weeks ago, we cut it open, put it under an electron microscope, and measured the smallest wires inside the chip, the metal pitch. And what we saw was unexpected. The smallest metal pitch inside the Kirin 9030 measures only 32.5 nanometers. That’s smaller than the metal pitch in Panther Lake, which is based on Intel's brand-new 18A node. A Chinese fab, cut off from the most advanced tools, without EUV, is packing its wires about ten percent tighter than Intel's leading edge EUV node. What’s going on here?"
中文
不用任何语法知识，单靠电脑里最普通的压缩软件（gzip），就能直接算出人类语言的进化树。这是2002年一篇著名论文《语言树与压缩》里的真实实验。机器是怎么靠打包文件摸透语言结构的？拿两份不同语言的文件A和B。把B的一小段文字掐下来，拼在A的结尾，然后执行压缩。接着算一算它跟单独压缩A的体积差。如果这两种语言结构相似，压缩软件用处理A的规律去对付B依然会极其顺手，新增体积就会非常小。反之，如果两种语言八竿子打不着，压缩算法碰壁，新增体积就会激增。没有任何预设字典，连主谓宾是什么都不懂。仅仅根据两段文本拼在一起能不能被压得更小，机器就自动把全球文件分了类，连谁和谁是语言近亲都排得明明白白，甚至还能顺手鉴定出文章是谁写的。在大语言模型横空出世的二十年前，普通的压缩软件就已经在做这种原始的相似性判断了。机器不需要理解人类的情感，它只需要想尽办法把文件变短，就能被迫摸到语言的骨架。这里面藏着一个极其核心的数学概念：交叉熵（Cross Entropy）。它最初就是在衡量针对A环境优化的压缩方案，扔到B环境里有多费劲。而今天，所有顶尖大语言模型的训练底座用的核心指标全都是它。表面上，千亿参数的模型是在玩预测下一个词的文字接龙。但在数学的深处，只要想尽办法把庞大的文本压缩到极致，智能自己就会浮出水面。
中文
Now it is fully integrated and officially supported 😎
github.com/Lakr233/vphone…
zqxwce@_zqxwce_
vphone 27.0b3 😁 (with 26.4 kernel 😜)
English
Nangate45 is not a manufacturable technology. I wait for the first real working silicon.
English
@fiapp_pro Just consider data/computing markets, Alibaba should become the Oracle of China.
English
