Aei

每日保持大脑的主动安全触发

LuciusAI 架构工程师

[Aei @5cdcmkbk]

@LuciusHQ The Context Layer allows an Agent to move beyond being stuck in a single snapshot state, enabling it to evolve continuously and sustainably support enterprise scenarios.
It really feels like hiring an intern from scratch and gradually training them into a professional employee.

[Aei @5cdcmkbk]

Context Layer 和传统 RAG 最大的不同在于，后者更强调知识系统的完整性，而 Context Layer 则将所有知识变成了可复利的企业资产。

这意味着哪怕 1% 的知识完整性都会立刻见效，而随着越来越多的 Context 涌入，AI 会真正融入工作流，无感参与人类协作，持续产生价值。

[Lucius @LuciusHQ]

We raised $3M to build Lucius AI - the Context Layer for Your Organization. Backed by Future Capital Discovery Fund, we’re tackling a problem we kept running into ourselves:

Individuals ship 10× faster with AI. Organizations don't.

Over 30% of your team's time is spent rebuilding context someone already had. It shows up everywhere a decision was already made but can't be found again - community operations, customer support, pre-sales reception, sales research, project management, internal collaboration.

We're building Lucius to close that gap.

[Lucius @LuciusHQ]

We raised $3M to build Lucius AI - the Context Layer for Your Organization. Backed by Future Capital Discovery Fund, we’re tackling a problem we kept running into ourselves:

Individuals ship 10× faster with AI. Organizations don't.

Over 30% of your team's time is spent rebuilding context someone already had. It shows up everywhere a decision was already made but can't be found again - community operations, customer support, pre-sales reception, sales research, project management, internal collaboration.

We're building Lucius to close that gap.

[Aei @5cdcmkbk]

Codex on mobile 之后 belike：

[Aei @5cdcmkbk]

@xuanwo GPT: 这其实不是充满高端概念的意淫，而是我在里面拉屎作者都不愿看一眼的垃圾堆。让我稳稳接住你～

[Aei @5cdcmkbk]

提一个问题：

如果之前说 AI 无法完全代替“高分人群”，现在一个高认知个体+旗舰模型 > 10+高认知个体，不也是完全替代了剩余 9 个人吗？

[Timothy Gowers @wtgowers @wtgowers]

I've recently got in on the act of getting AI to solve open problems in mathematics. More precisely, I gave some questions asked by Melvyn Nathanson to ChatGPT 5.5 Pro, to which I have been given access, and it answered them. 🧵

[Aei @5cdcmkbk]

其实我挺奇怪一些云平台交互设计的，特别是 Token 创建的时候，来一句 Copy place. You will not be able to see it again.

我创建出来 Token 肯定是存在别的地方了，然后就创建的地方看不见，也太反人类了。

[Aei @5cdcmkbk]

@AlchainHust Antigravity 尝试过一些，也挺好用的，模型影响 Agent 的能力比例会更大一些

[Aei @5cdcmkbk]

@ryolu_ Back to the best era.

[Aei @5cdcmkbk]

Codex 最新版本上线了 Hook 能力，突然想到一些内容。

Agent + 工程化（包括 Hook 和 Harness）都是在尝试做大模型的灵活度和工程化准确控制的 trade-off。

回到第一性原理，这是可预测和不可预测的混合调剂。在某个位置可预测，某个位置有惊喜这件事，会是后续 Agent 应用一直要关注的事情。

[Aei @5cdcmkbk]

@Russell3402 打粉是什么？10 年程序员在线炒粉吗😂

[Aei @5cdcmkbk]

@amehochan 想问一下在 Vibe 过程中是纯通过 Feature 来驱动大模型开发的吗？还是先设计了一些数据模型，再让 AI 开始写。

好奇这种需要大量数据驱动的项目，纯靠 AI 能否做到很好的状态流转

[Aei @5cdcmkbk]

所以在 AI 时代的新范式下，工程思维或者架构思维依旧应该是高阶工程师的基础素养？

还是说新时代会有新的架构思维需要被整理并出现？

[Aei @5cdcmkbk]

@Russell3402 现在大部分大学生毕业就是高不成低不就的状态，学点技术挺好的。
再过两天我被写成 Skill 就去学个电焊 hhh

[Aei @5cdcmkbk]

@Jiaxi_Cui 可是现在让 AI 出方案 80% 还是使用 nginx

[Aei @5cdcmkbk]

@jakevin7 cool！

[Aei @5cdcmkbk]

@ai_xiaomu 希望视频里的女生在 Vibe Coding 出一坨屎的时候，也能引导 AI 情绪，让 Agent 攻击代码

[Aei @5cdcmkbk]

@Russell3402 “你这个网站为啥和我的一模一样”😆

[Aei @5cdcmkbk]

用 Codex 帮朋友做了一个四级单词学习表，用 cloudflare page 服务部署，从开始做到发给朋友链接 20 分钟不到。

小需求的 AI 时代红利

[Aei @5cdcmkbk]

Codex 一个界面上三个输入口：
1. 前端批注
2. 主输入框
3. 宠物小人

没起到什么负面影响，分别解决了我的细节要求，主线程任务和情绪价值🤣