Hugo_Liu

Done is better than perfect.
Done is better, then perfect.

[九原客 @9hills]

个人学习工作流

一、网上��兴趣的信息顺手采集到自托管的 Karakeep。

二、有意识采集的，以及从Karakeep 中过滤出需要精读或保存到的，保存到 Zotero（支持网页快照）。

Zotero 使用免费10GB的webdav 网盘，也可以自建。
使用多种pdf双语对照翻译服务出对照PDF。

三、使用Claude code + Obsidian 整理研��内容。基本到这步的量很少。

所有工具都有 iOS 版本，保证移动工作流顺畅。

[郭宇 guoyu.eth @turingou]

One Nature Nyaruswiga 一处真正的奢华营地！家人一般热情又完美的服务，独特的用餐体验，近在咫尺的大象🐘猎豹🐆狮子🦁和其他动物们居住的小山谷，我们在此住了五天，这是我目前为止心中评价最高，世界上最完美的度假村！

[AIGCLINK @aigclink]

一款论文转多模态海报工具：Paper2Poster，给它一篇论文，可自动生成一张学术海报，生成质量高制作成本低

生成的海报可读性较好，结构清晰、用词精简，比GPT-4清晰可读，比PPTAgent布局合理

输入论文PDF全自动处理，可以自动提取重点，进行智能���版设计，自动调整布局，维持论文逻辑顺序并控制信息密度

#论文转海报 #AI学术海报

[Pavan Jayasinha @pavanjayasinha]

I implemented an LLM end-to-end in hardware, and ran it on an FPGA.

Zero Python. Zero CUDA. Just pure SysVerilog.

All my progress + everything I learned from 200h of LLM chip design (demo at the end)👇

[Andrew Ng @AndrewYNg]

I’ve noticed that many GenAI application projects put in automated evaluations (evals) of the system’s output probably later — and rely on humans to judge outputs longer — than they should. This is because building evals is viewed as a massive investment (say, creating 100 or 1,000 examples, and designing and validating metrics) and there’s never a convenient moment to put in that up-front cost. Instead, I encourage teams to think of building evals as an iterative process. It’s okay to start with a quick-and-dirty implementation (say, 5 examples with unoptimized metrics) and then iterate and improve over time. This allows you to gradually shift the burden of evaluations away from humans and toward automated evals.

I wrote previously in The Batch about the importance and difficulty of creating evals. Say you’re building a customer-service chatbot that responds to users in free text. There’s no single right answer, so many teams end up having humans pore over dozens of example outputs with every update to judge if it improved the system. While techniques like LLM-as-judge are helpful, the details of getting this to work well (such as what prompt to use, what context to give the judge, and so on) are finicky to get right. All this contributes to the impression that building evals requires a large up-front investment, and thus on any given day, a team can make more progress by relying on human judges than figuring out how to build automated evals.

I encourage you to approach building evals differently. It’s okay to build quick evals that are only partial, incomplete, and noisy measures of the system’s performance, and to iteratively improve them. They can be a complement to, rather than replacement for, manual evaluations. Over time, you can gradually tune the evaluation methodology to close the gap between the evals’ output and human judgments. For example:
- It’s okay to start with very few examples in the eval set, say 5, and gradually add to them over time — or subtract them if you find that some examples are too easy or too hard, and not useful for distinguishing between the performance of different versions of your system.
- It’s okay to start with evals that measure only a subset of the dimensions of performance you care about, or measure narrow cues that you believe are correlated with, but don’t fully capture, system performance. For example if, at a certain moment in the conversation, your customer-support agent is supposed to (i) call an API to issue a refund and (ii) generate an appropriate message to the user, you might start off measuring only whether or not it calls the API correctly and not worry about the message. Or if, at a certain moment, your chatbot should recommend a specific product, a basic eval could measure whether or not the chatbot mentions that product without worrying about what it says about it.

So long as the output of the evals correlates with overall performance, it’s fine to measure only a subset of things you care about when starting.

The development process thus comprises two iterative loops, which you might execute in parallel:
- Iterating on the system to make it perform better, as measured by a combination of automated evals and human judgment;
- Iterating on the evals to make them correspond more closely to human judgment.

As with many things in AI, we often don’t get it right the first time. So t’s better to build an initial end-to-end system quickly and then iterate to improve it. We’re used to taking this approach to building AI systems. We can build evals the same way.

To me, a successful eval meets the following criteria. Say, we currently have system A, and we might tweak it to get a system B:
- If A works significantly better than B according to a skilled human judge, the eval should give A a significantly higher score than B.
- If A and B have similar performance, their eval scores should be similar.

Whenever a pair of systems A and B contradicts these criteria, that is a sign the eval is in “error” and we should tweak it to make it rank A and B correctly. This is a similar philosophy to error analysis in building machine learning algorithms, only instead of focusing on errors of the machine learning algorithm's output — such as when it outputs an incorrect label — we focus on “errors” of the evals — such as when they incorrectly rank two systems A and B, so the evals aren’t helpful in choosing between them.

Relying purely on human judgment is a great way to get started on a project. But for many teams, building evals as a quick prototype and iterating to something more mature lets you put in evals earlier and accelerate your progress.

[Original text: https://t.co/V3BZe8sRWE ]

[Ying Sheng @ying11231]

Hi everyone in the community. We expect SGLang to achieve the best performance. If you meet any problem or see a not high enough performance when using SGLang, please let us know. Your feedback is the key to make the product better.

[Hugo_Liu @Hugo_Liu_Dev]

这就是我为什么那么不看好前司的点🤪🤪🤪

[magnolia1234_bpc @Magnolia1234B]

Bypass Paywalls Clean new release v4.0.2.0

+ Bhaskar
+ History Today (app only)
+ StatNews
+ Telerama-fr
- LaTribune-fr
- Law-com
Fix Aachener-zeitung-de
Fix LeFigaro-fr
Fix Lequipe-fr (video)
& more

Chrome: https://t.co/RGoTmtqTgb
Firefox (delayed): https://t.co/YM7UNcVp8q

[HPCA @HpcaArchConf]

Distinguished Artifact Runner Up:
Title: "An Optimizing Framework on MLIR for Efficient FPGA-based Accelerator Generation"
Authors: Weichuang Zhang, Jieru Zhao, Guan Shen, Quan Chen, Chen Chen, Minyi Guo

[宝玉 @dotey]

新手 Cursor 用户：给我写一个博客系统
Cursor：一个简陋博客首页
用户：太烂了，没卵用，再也不用了

专业 Cursor 用户：这是博客日志列表的截图，请参考截图的样式使用Tailwind+ShadcnUI做一个UI组件显示日志列表，数据访问参考 @/app/posts.ts 里面的方法

Cursor：你要的日志列表组件 /components/post-list.tsx
用户：链接不对，我期望的是 /posts/$id 但是你用的 /post?id=$id

Cursor：这是修改后的列表组件

新开会话继续写分页功能

专业 Cursor 用户：这是列表组件代码 /components/post-list.tsx，现在请给它加上分页功能，数据库访问代码在 @/app/posts.ts ，需要添加相应的分页访问代码代码，分页组件使用shadcnui的Pagination组件，样式使用Tailwindcss

Cursor：这是更新后的数据库访问代码，这是添加分页组件后的日志列表，请确保你安装了Pagination组件，这是安装方法

用户：请做以下调整：
- 修��分页链接为 /posts/page/$page
- 只保留上一页、下一页链接

Cursor：这是修改后的代码

[Tony出海 @iamtonyzhu]

下载社媒资源是刚需。
这个居然还是开源的！github 地址： https://t.co/Ree0LzRQFt

[Tigris 会讲课教授是好老师 @tig88411109]

🎓斯坦福推出了一款令人惊叹的AI驱动的研究工具！

它叫 STORM，你只需输入一个主题，它就会从数百个网站中搜索信息，为你撰写一篇全面的研究文章。从格式/主题/内容/引用/绝对专家水准。

关键是：完全免费对公众开放！

这简直是Google Research的有力挑战者，赶紧试试吧！ #AI工具 #AI研究

[Orange AI @oran_ge]

Trancy 这个插件确实很厉害
把观看外文YouTube视频的体验提升了一大截儿
而且还有多种练习模式

[Levi Yuan @levinomad]

昨天用了下v0. dev, 一键copy一个网站，效果很惊艳，感觉是cursor和windsurf后，最被忽视的一个AI神器。

1.找到你想要网站，截图（或者自己figma设计）
2.图片导入v0. dev
3.简单描述你的要求，比如按钮是否用组件
4.点击生成

最后效果，几乎98%以上还原了对标网站，无论排版，颜色，字体，布局等等。
最大的颠覆是，他还能自动识别网站那些数据是动态的，哪些模块是可以抽离的，微调下可以直接使用。

以下是我用adsense教程网站截图生成的的效果

[超级个体 Austin @austinit]

分享一个开源的 PC 微信聊天记录数据导出工具。

导出后数据可以做永久化保存，即使微信停止支持，聊天记录也可以随时查看。
前端界面尽量与微信界面保持一致，减少使用成本。
。
下面是效果图，不是微信界面。链接👇
https://t.co/vvWRdQ3FKE

[Josh @josh_bickett]

I’ve been working on something like this with https://t.co/IzDRBN5AnD, the idea is that an audiobook plays and you can interrupt to ask question.

The challenge is getting copyrights to real books so I’ve been trying to prompt LLM to generate the audiobook on the fly.

I may try adding some public domain books

[Jiayuan (JY) Zhang @jiayuan_jy]

下午的时候充分体验了一下 ChatGPT o1 pro mode。

直观的感觉是：

在复杂代码的能力上超过了 Claude 3.5 Sonnet。

借助 o1 pro mode，把 Lucene 引擎中的核心部分用 Python 重写了一下。

puceny - a simple local search engine

支持全文 index 一个本地的文件目录，并提供了一个简单的 web UI 来搜索。

https://t.co/M3kgiYZKij

[Hugo_Liu @Hugo_Liu_Dev]

#llm_application

[张XX @alexu19049062]

发现一款开源项目，是使用 Flutter 开发的第三方 Bilibili 客户端。
功能包括推荐视频列表、热门直播、动漫系列、离线缓存、回复评论、子弹聊天和搜索功能，好评的是还支持画中画，目前已有 7.8k star。

项目源码：https://t.co/224m74pv9g

[Hugo_Liu @Hugo_Liu_Dev]

#独立开发

[0xinhua @0xinhua_zh]

现在开展自己的软件生意真的极其地便宜。

你唯一需要的就是 20 美元左右域名费用👇：

🎨 设计：Figma $0
💾 数据库：Supabase $0
🚀 部署：Vercel $0
🔐 用户：Auth0 $0
📧 邮件：Resend $0
💻 编码：V0 + Cursor $0
📊 统计：Mixpanel $0
🔗 短链：Dub $0
📬 公司邮箱：Lark $0
🤖 顾问：ChatGPT $0