777

分享AI、制作艺术、在Web3创造。
Share AI, make art, build in Web3.

[777 @027Simpulati]

@DRbitcoin36 这就玩上GTA6抢先版了？🤨

[777 @027Simpulati]

I finally get why the multi-agent team I cobbled together with MCP a few months ago never really worked.

A couple of days ago in Kiro, while working on a complex task, I got to see how well the main agent and sub-agents can actually collaborate. Even pure sub-agent delegation produced results that genuinely impressed me.

Really looking forward to the day when true multi-agents can freely talk to each other and collaborate.

[Ryan Lopopolo @_lopopolo]

We do not have multi-agent systems today. Ultra mode is still one trained system delegating to itself.

What comes next has not been invented yet. https://t.co/oMtfTGmF85

[777 @027Simpulati]

开源，MIT。
https://t.co/vsr7biqaC7

目前要自己编译安装：clone 下来跑 ./install.sh。需要 macOS 14 以上、Apple Silicon、Homebrew。

[777 @027Simpulati]

不是，Mac 自带的输入法，只有我一个人觉得不好用吗。

尤其是非英语系统的 Mac，打英文没选词，没纠错，就一个字母一个字母硬打。

所以我自己做了个输入法，叫 ABC+，专门给非英语系统的 Mac 用：
1.输入有候选框，数字键选词。
2.错字按空格自动纠错，下半句用灰字预测句子。

模型跑在你自己的 Mac 上，不联网，也不要任何系统权限。

名字很好记：系统叫 ABC，缺了点东西，我加了个 +。

[777 @027Simpulati]

时间是成功的显影液

“Hard work beats talent when talent doesn't work hard”

[jeff.hl @chameleon_jeff]

1/ Controversial take: hard work is more important than smart work.

It's a myth that we only have a few hours of good creative work per day. Train yourself to grind long hours first. You will surprise yourself. The work naturally become higher quality, less distracted.

[777 @027Simpulati]

@Casimirtrunix 🥹

[777 @027Simpulati]

去年，我的论文发表在 Nature 合作的 AI 期刊上。

论文正式发表的那一天，我没有想象中的开心，反而更多的是担忧。

得到一个成果也许没有那么难，但想要超越一个成果，需要成倍的精力、一次次打破固有的思维路径，以及承受随之而来的挫败的能力。

但其实在这个成果之前，我就已经意识到：

AI 应该服务于人，技术也应该服务于人。

也许是艺术背景的原因。从启蒙开始，我所欣赏的好作品、好设计，始终都是“懂人”的设计。

AI 只是换了一种工具和路径，但核心从未改变。

所以我一直想做的，是把前沿的研究能力真正落地，做成一些对人有价值的产品。

过去一年多，我持续探索 AI 与产品的结合，也不断寻找一个真正能够让我把想法落地，all in的平台，来超越我的上一个成果。

现在，这个备忘录的目标有了答案。

[777 @027Simpulati]

目前做出来使用率最高的产品

一眼看所有关注的涨跌，增加了预警和盈亏的功能后更方便了

并且和mac搭配 美观度满分（绿色的一天心情真好

[777 @027Simpulati]

Bear market关注大盘情绪尤为重要，尤其是在最近反复震荡的市场里，直接挂钩你的“运气”。

如果有桌面组件可以直接看所有关注的链上资金流动变化，不用每次去网站看，也不用每次记下前一天的趋势情况，直接一秒就做完大盘情绪分析该多好。

但是类似开源的工具在一年前都没找到过，现在除了有的cex支持桌面组件，但针对总量和流向以及现在的美股还是不够定制化。

所以我做了这样的工具，一眼看懂市场温度 + 资金流向 + 美股,不用登录、不用点开任何 App，熊市里最需要的那点信号,就浮在你桌面上。

开源免费，一条命令让agent拉下来👇（目前只支持Mac，需要win的可以私我
⭐仓库:https://t.co/tzwr3ev2it

[777 @027Simpulati]

@Casimirtrunix 太牛了 还以为是gpt6

[777 @027Simpulati]

提前阅读了内部抢先版，整理得非常详细，对研究Robin Hood Chain & meme来说很有实践价值

[Trunix❇️ @Casimirtrunix]

用 AI 整理了一版 Robinhood Chain Meme / RWA的研报：让你看懂早期资金从哪来、在哪成交、如何退出。

重点拆解热门币、早期钱包、KOL 首喊记录及平台费率，帮你搞清楚：盯什么、跟谁、去哪买更省、何时该走。

完整内容详见下文。 https://t.co/yS4dEEZD58

[777 @027Simpulati]

全部藏品 + 链上验证 👇

🖼 OpenSea: https://t.co/HqZZ9vjWNs

📜 合约 (Polygon): 0x20DA42bE9d92B8A39f231284a3AFe4E577030fFB

挑好想要的,评论区见 👀

[777 @027Simpulati]

上次的momoro mascot,全部上链了

评论留下地址会随机送NFT🐸

[777 @027Simpulati]

Crypto spent a decade counting users, but value never accrued to headcount, it accrued to where fees and settlement land.

[Raoul Pal @RaoulGMI]

Valuing crypto has always meant guessing how many people would eventually come onchain. 

We charted adoption curves, compared them to the internet, argued over when we'd hit a billion users… 

But even the most active human has a limit. A full time trader executes a few times an hour, for maybe twelve hours a day because sleep and food and life still exist. Thats the ceiling of human paced activity and its what this entire asset class is priced on.

But now, billions of agents are about to come onchain and agents have no ceiling. Circle's CEO is talking about tens of billions of them. Stripe is planning for a billion transactions per second. Visa, Mastercard, Coinbase, Klarna and Google are building agent payment rails right now.

The total addressable market for crypto was never the number of humans. Its the number of economic actors, and we are about to find out what happens when that number grows by orders of magnitude and never sleeps. Activity we cant comprehend, and all of it flowing through the networks the agents run on.

This is Reed's Law... value scaling not with the number of users but with the number of groups they can form. Agents dont just transact. They hire each other, form swarms, assemble into temporary companies for a single task and dissolve again a minute later. Every group is new value. Its an exponential built on top of an exponential, and nothing like it has ever existed.

[777 @027Simpulati]

@sunyuchentron 可以用这个skill和ai聊天，还原孙哥的行文，见字如面：

[777 @027Simpulati]

孙割那篇六千字的小作文，最聪明的不是文笔，是单位

我把这套写法提炼成了「孙本位.skill」

[777 @027Simpulati]

@Casimirtrunix 一个skill的大小 4MB

[777 @027Simpulati]

孙割那篇六千字的小作文，最聪明的不是文笔，是单位

我把这套写法提炼成了「孙本位.skill」

[777 @027Simpulati]

@justinsuntron 试用同款孙题文学：https://t.co/G3p8PT89Mf

[777 @027Simpulati]

孙割那篇六千字的小作文，最聪明的不是文笔，是单位

我把这套写法提炼成了「孙本位.skill」

[777 @027Simpulati]

附孙本位汇率表：