Maybe

B

[Claude @claudeai]

Our latest Claude Code hackathon is officially a wrap.

500 builders spent a week exploring what they could do with Opus 4.6 and Claude Code.

Meet the winners:

[Maybe @Bug45585917]

@aiedge_ @grok 总结下文章发给我

[Andrej Karpathy @karpathy]

There's a new kind of coding I call "vibe coding", where you fully give in to the vibes, embrace exponentials, and forget that the code even exists. It's possible because the LLMs (e.g. Cursor Composer w Sonnet) are getting too good. Also I just talk to Composer with SuperWhisper so I barely even touch the keyboard. I ask for the dumbest things like "decrease the padding on the sidebar by half" because I'm too lazy to find it. I "Accept All" always, I don't read the diffs anymore. When I get error messages I just copy paste them in with no comment, usually that fixes it. The code grows beyond my usual comprehension, I'd have to really read through it for a while. Sometimes the LLMs can't fix a bug so I just work around it or ask for random changes until it goes away. It's not too bad for throwaway weekend projects, but still quite amusing. I'm building a project or webapp, but it's not really coding - I just see stuff, say stuff, run stuff, and copy paste stuff, and it mostly works.

[Naval @naval]

Vibe coding is the new product management.

Training and tuning models is the new coding.

[歸藏(guizang.ai) @op7418]

之前提过一嘴国内没办法上 ChatGPT 的朋友我一般会推荐用月之暗面的 Kimi Chat。
我最近也开始用的比较多了，尤其在一些我的日常任务上面，它比 ChatGPT 还方便一些，尤其是不用担心网络问题打开就用，最重要的是目前所有功能都是免费的。

Kimi 的表现从侧面上说明一个 AI 产品模型本身能力是一方面，对整个 LLM 生态其他能力的打磨和探索也很重要。

下面从我日常用的最多的两件事情来举例子：

首先是论文的总结和阅读，之前 PDF 格式的论文文件总结起来很费劲因为涉及到很多 OCR 的事情即使是 GPT-4 完成的也不太好，总结的不够详细总是忽略关键信息。

刚好最近arxiv默认支持了网页格式的论文，理论上应该更好总结了，直接丢网页版本链接就行，Kimi 在这件事情上做的非常好，比如昨天字节的这个很复杂的论文，非常详细而且结构很清晰，我没有特别优化提示词。

反观 GPT-4，直接告诉我没办法访问这个链接，太离谱了，之前 PDF 的时候虽然不输出关键信息，起码还会写，这下直接不访问了。

我日常最多用的第二个事情是翻译，所以第二个测试是用宝玉的翻译提示词分别让 ChatGPT 和 Kimi 翻译同一段比较复杂的 LLM 论文的简介，首先翻译内容上两者没有出现幻觉和丢失的情况。

ChatGPT出现问题的方面主要是体验和行文，Code 组件无论多长的内容都不会换行不管输出的内容是不是代码，所以导致我根本无法在界面上完整预览，输出结果。 Kimi 则会在大段文本内容输出的时候正确的换行，起码我可以看全。

另外在意译结果上，我感觉 Kimi 更符合中文的书写和阅读直觉，看起来更加流畅，可能还是因为 ChatGPT 主要照顾的还是英文用户，所以中文语料相对较少，导致写的中文即使通过提示词优化过也还是有翻译腔。各位也可以对比一下。

这里使用 Kimi，现在是免费的：https://t.co/7DpYpX180A

[Paul Graham @paulg]

"I think a life properly lived is just learn, learn, learn all the time."

— Charlie Munger

[Vincent Niu @VincentNiu222]

把一些之前在twitter上发过的碎片内容展开讲讲 -
1.在Web2.0里，你的数据有价值，但不属于你，价值也不分给你。
2.链上数据是一座金矿，应用越多，数据越多，金矿越大
3.在Web3.0里，你的价值行为终将获得回馈，价值的载体是你的数据，也属于你自己。
https://t.co/C1dH7m01q4

[6529 @punk6529]

1/ On NFT Twitter

We have a lot of newcomers to NFT twitter. This is a thread to teach them our ways.

6529 will start this list today, but will add to it w/suggestions.

NFT twitter has a great culture, very positive. Let's keep it this way.

[Mason Nystrom @masonnystrom]

What is Web3?

Web3 is a shift in internet culture.

Empowered by new technologies, the Web3 movement is spearheaded by a shift in how we, the collective, view and value the Internet.

1/

[Eshita @eshita]

One of my favorite quotes from @ljin18 in this interview:

"Creators are building their businesses on rented land when they are building on web2 platforms"

[Naval @naval]

Just as you travel so that you can miss your home, you socialize so that you can miss your self.

[Naval @naval]

In the modern world, don’t minimize your risks - maximize your opportunities