即刻App年轻人的同好社区
下载
App内打开
PatrickKavanagh
612关注2k被关注5夸夸
Old Friendship 创业投资人 ChatCut, Wanaka, Manus, Plaud, Robinhood
置顶
PatrickKavanagh
4月前
与许多中国风险投资人不同,我和我的合伙人克里斯不会投资于竞争对手。投资竞争对手是风险投资人能做的最不忠诚和不专注的事情。它显示出完全缺乏信念,并且是懦弱的行为。
一旦我们选择了创始团队,我们就会完全承诺给他们,并确保他们有赢得的最佳机会。对我们来说,忠诚和完全专注是世界一流伙伴关系的基础。即使我们最终错了,错失了赚钱的机会。我们不想投资于Robinhood、Manus、Plaud、Qbit或Wanaka的竞争对手。当我们给你开支票时,那就是我们的承诺。坚持到底。(Grok translation)

Unlike a lot of VCs in China, my partner Chris and I do not invest in competitors. Investing in competitors is the most disloyal and unfocused thing a VC can do. It shows a total lack of conviction and is spineless behavior.

Once we pick a founding team, we commit to them entirely and make sure they have the best chance of winning. To us, loyalty and total focus is the foundation of a world-class partnership. Even if we end up being wrong and leave money on the table. We don’t want to invest in Robinhood, Manus, Plaud, Qbit, or Wanaka competitors. When we write you a check that’s our promise. In it until the end.
817
PatrickKavanagh
5天前
数学创新即将变得疯狂。
Harmonic 在不断进步的过程中,将像撕裂纸张一样横扫那些未解之谜——完全形式化,且可证明。
我的朋友 Boris(可能是当今活着的最聪明数学头脑之一),用 Harmonic 的数学 AI「Aristotle」只花了 6 个小时就解决了一道 Erdős 问题。

Math innovation is about to get crazy. Harmonic is going to rip through unsolved problems as it gets better; fully formalized and provable. My friend Boris, probably one of the smartest mathematical minds alive, solves an Erdos problem in 6 hours with Aristotle, Harmonic’s math AI.
23
PatrickKavanagh
9天前
Old man’s been running this bar for 56 years and has built quite the music collection.
00
PatrickKavanagh
27天前
好吧,我显然错过了这个,但谷歌正在训练大型语言模型与海豚交流——这太酷了!!

Ok I clearly missed this but Google is training LLMs to talk to dolphins - which is so cool!!

DolphinGemma: How Google AI is helping decode dolphin communication

00
PatrickKavanagh
28天前
Big rice terrace week
50
PatrickKavanagh
1月前
今天午餐时,我和别人聊起了中国科技圈里的英雄崇拜现象。创始人常常想成为某个成功人士的样子,并且总是挂在嘴边。
在英语中,有一句谚语“永远不要见你的英雄”,这绝对是真理!著名创始人的公关宣传叙事隐藏了太多人性缺陷。研究这些童话故事是非常不健康且误导性的。
我的建议是从你自己生活中的那些有才华且有动力的人开始!如果你没有这样的人,就去寻找他们——但要找到真实的东西。
著名创始人童话故事的知识对你不会有任何影响,而你自己生活中的才华和灵感,如果你仔细观察并认真把握,几乎肯定会影响到你。

I had a conversation at lunch today about hero worship in the Chinese tech community. It is common for founders to want to be like someone successful and talk about it all the time.

In English we have a saying “never meet your heroes” and it’s absolutely true! There are so many human flaws that a PR-driven narrative of famous founders hides. It is deeply unhealthy and misguided to study these fairytales.

My advice is start with the talented and driven people in your own life! Seek them out if you don’t have any - but find something real.

The knowledge of famous founder fairytales will in no way affect you, whereas the talent and inspiration in your own life almost certainly will if you look closely and take the opportunity seriously.
56
PatrickKavanagh
1月前
I feel lucky to have a lot of older sisters and brothers in Beijing who look out for us and are true friends. Yangjie, Xiaochao, Cassie, and the rest of the team are always a pleasure to spend time with.
20
PatrickKavanagh
1月前
We are so early to doing full work tasks, but it’s exciting to see a smaller team staying competitive with the big guys.
31
PatrickKavanagh
1月前
Pretty crazy that 1X spent so much time on a hype video for a teleoperated robot. Not only did they invest a lot in the launch but X sentiment was super negative when people found out it wasn’t AI and a remote operator would have access to the privacy of their home.
30
PatrickKavanagh
1月前
最近一周,我作为数据科学家更多地使用Claude的代码,它自从今年夏天我尝试以来,取得了巨大的进步:
很少出错——干净地将数据加载到数据库中,并且能够进行数据科学工作。
在丰富数据方面也非常出色。我的使用案例是将信用卡对账单与旅行收据邮件(包含旅行日期、城市等)结合,生成大量关于我的消费习惯的统计数据,还将会计风格从现金制转向权责发生制。我整个周末都花在这上面!
还没有尝试在它上面使用Python的机器学习库,但可能接下来会试试。

I’ve been working more with Claude code as a data scientist this last week and it has made a ton of improvements since I tried this summer:

Rarely gets confused - loads things cleanly into a DB and can do data science.

Also great at enriching data. My use case was combining credit card statements with travel receipt emails (with travel dates, cities, etc) to generate lots of statistics about my spending habits and also shifting accounting style from cash to accrual method. Spent all weekend on it!

Haven’t tried using python ML libraries on it yet but might try that next.
00
PatrickKavanagh
1月前
Anyone wanna build GPU-native software instead of CPU’s?

Christian is one of the main Rust library contributors for CUDA and is looking for engineers experienced in compilers, GPU/Linux kernels, Vulkan/ROCm/CANN.

He was head of Facebook mobile release tools and worked with me at Robinhood. Hiring globally.

Announcing VectorWare

00