Taeyoung Kim

CEO of AIFactory, Keras Korea, Kaggle Korea

[Taeyoung Kim @tykimo]

We used an artificial intelligence model to generate the backside magnetogram of the Sun and our work was listed in Nature Astronomy. Of course, the model is implemented as Keras. Thanks to @fchollet and Keras that helped me get into deep learning easily.

[Khairallah AL-Awady @eng_khairallah1]

Eric Schmidt (ex-Google CEO): “if you really want to make money, it’s actually easy. found an agentic AI company.”

If I had only 30 days to do that , I'd begin here and save this:

Agent Architecture
https://t.co/Xyy3e9AjAQ

Claude Code 101:
https://t.co/tZbHeRDWkj

Claude Code in Action:
https://t.co/RDYEVbydhW

Prompt engineering (official):
https://t.co/aYQzAWmObh

Interactive prompt tutorial (hands-on):
https://t.co/5k9My0hYgY

CLAUDE.md & how to give Claude memory:
https://t.co/gtmOGKAvDe

Skills, teach Claude reusable workflows:
https://t.co/DJFqh3E6OB

MCP, time connect Claude to Slack, GitHub, Drive:
https://t.co/XbRdmmcYmP

Routines (automate tasks 24/7):
https://t.co/LGbhOeWWdJ

Claude Code Ultimate Guide (community):
https://t.co/56DAmEuqH8

Awesome Claude Code (skills, hooks, plugins):
https://t.co/jUIBuxvV5K

All 13 Anthropic Academy courses (free certs):
https://t.co/rHn0gDmtGH

Claude Code full docs:
https://t.co/KYHnapDdHG

All of this is for free at $0/month

Then read this guide by this builder

[Anatoli Kopadze @AnatoliKopadze]

Anthropic engineers finally showed how they actually use Claude Code internally

31 minutes of internal workflow that most Claude users will never see on their own

here's what they cover:
> how to set up project context files the right way
> custom commands that save hours of repeated work
> hooks that make Claude behave exactly how you need
> subagents and how to actually spec them properly

"your agent isn't the problem, your spec is"

the people who understand how Claude Code actually works inside Anthropic are shipping things everyone else thinks requires a whole team

that's exactly why I put together a breakdown of Claude features most people have never discovered

you can find it below

[Anthropic @AnthropicAI]

We've raised $65 billion in Series H funding at a $965 billion post-money valuation, led by @AltimeterCap, Dragoneer, @Greenoaks, and @sequoia.

This investment will help us advance our research and expand our capacity to meet growing demand for Claude.

[Pankaj Kumar @pankajkumar_dev]

Claude Sonnet 4.8 Leaks

- Anthropic accidentally shipped a massive 512,000-line internal debugging source map through a Claude Code npm update on March 31, 2026
- The leaked source code references Sonnet 4.8 inside unreleased keyword filters, strongly suggesting Anthropic is skipping Sonnet 4.7 entirely
- Sonnet 4.8 is expected to inherit vision upgrades from Opus 4.7, reportedly pushing visual accuracy above 98% for UI mockups and messy architecture diagrams
- A major coding boost with cleaner one pass completions and much more literal instruction following
- A new "X high" reasoning effort tier may also be coming for stronger logic without extremely long generation times
Updated tokenizer may use 30% more tokens on the same prompts compared to older Sonnet versions
- Currently expected around mid to late June

[lucas @lucas_flatwhite]

👩🏻‍💻 Hermes Agent: 실무자용 참고서 (한국어)
https://t.co/H2COIn0zPr

오.. 너무 좋네요! Hermes Agent를 실무에서 제대로 써보고 싶다면, 이 한국어 가이드를 꼭 한 번 들여다보세요.

이 가이드는 설치부터 시작해서 tools, skills, multi-agent Kanban, 지속 메모리, cron 작업, MCP 통합까지 거의 모든 주요 기능을 섹션으로 나누어 설명하고 있어요.

heartbeat, multi-agent Kanban, skill 시스템, 다양한 provider 지원, 보안 이슈까지 꽤 깊이 있게 다루고 있어요.

실제로 Hermes를 오래 쓰면서 생기는 질문들,, 인증은 어떻게 관리하지, config는 어떤 계층으로 쌓이는지, skill은 어떻게 발전시키는지.. 이런 것에 대한 답이 정리되어 있죠.

Hermes는 여러 플랫폼과 동시에 대화할 수 있는 완전한 에이전트 런타임..!

저도 매일매일 Hermes와 함께 합니다~~

[rewind @rewind02]

Andrej Karpathy spent 4 minutes in an interview explaining a single idea

about how most people haven’t even started learning how to use AI

and everyone paying $20/month for a subscription.. that's not really using Claude at all

his point is that the real skill gap is the ability to build with AI

he identified 4 behaviors that break Claude Code and put them all into one file

a developer expanded it into 21 rules and published it - 82,000 stars and #1 on GitHub Trending

coding accuracy jumped from 65% to 94%

here's what these 21 rules actually are and why most developers using Claude every day have never configured them

the full breakdown is covered in the article below 👇

[Avi Chawla @_avichawla]

Karpathy's prediction about RL is coming true now!

He called reward functions unreliable and argued that a single reward number is too low-dimensional to teach an agent what "good" means for complex tasks. To solve this, Agents need a knowledge-guided review as a higher-dimensional feedback channel.

Every major AI lab trains models with RL today (OpenAI, Anthropic, DeepSeek).

And their key bottleneck has always been the reward functions.

GRPO by DeepSeek worked well for math and code because the environment gave a binary signal.

But for real agent tasks, someone still has to hand-code the scoring function. That takes days and breaks every time the pipeline changes.

RULER (implemented in OpenPipe ART, 10k stars) addresses the exact problem Karpathy identified.

The reward criteria are defined in plain English, and an LLM evaluates each trajectory against that description to provide feedback for training.

I trained a Qwen3 1.4B agent that plays 2048 using GRPO with this exact workflow.

In this case, the agent saw the board, picked a direction, and RULER evaluated the outcome, all from this natural language definition.

You can see the full implementation on GitHub and try it yourself.

Here's the ART Repo: https://t.co/fsoLXDK4Zu

(don't forget to star it ⭐ )

Just like RLHF replaced manual rankings and GRPO replaced the critic model, natural language rewards are replacing hand-coded scoring functions.

RL reward engineering is now prompt engineering.

I wrote a full walkthrough covering RL for LLM agents, from RLHF to GRPO to RULER, in the article below.

[ClaudeDevs @ClaudeDevs]

Happy Friday! We've reset everyone's 5-hour and weekly rate limits.

[astaxie @astaxie]

今天群里面讨论怎么样学习 Harness，Harness 工程我学习这两个：
1. https://t.co/l2EySwL22O 通过这个了解每一个 Harness 的核心机制
2. https://t.co/eaucFDhbyp 学习这个框架的各个模块设计实现，不懂的就让 AI 去解读实现逻辑

[CG @cgtwts]

Anthropic’s Claude Code team just showed how to automate your engineering workflow in under 30 minutes.

straight from the people who built it and it’s completely free.

[Khairallah AL-Awady @eng_khairallah1]

ANTHROPIC JUST RELEASED THE OFFICIAL PLAYBOOK FOR BUILDING A COMPANY WITH CLAUDE CODE.

30 minutes. free. from the engineers who built it.

Bookmark this before you forget.

CEO: 1 human. Employees: AI agents. Operations: fully automatic.

The zero-headcount company is no longer a joke.

[Andrej Karpathy @karpathy]

LLM Knowledge Bases

Something I'm finding very useful recently: using LLMs to build personal knowledge bases for various topics of research interest. In this way, a large fraction of my recent token throughput is going less into manipulating code, and more into manipulating knowledge (stored as markdown and images). The latest LLMs are quite good at it. So:

Data ingest:
I index source documents (articles, papers, repos, datasets, images, etc.) into a raw/ directory, then I use an LLM to incrementally "compile" a wiki, which is just a collection of .md files in a directory structure. The wiki includes summaries of all the data in raw/, backlinks, and then it categorizes data into concepts, writes articles for them, and links them all. To convert web articles into .md files I like to use the Obsidian Web Clipper extension, and then I also use a hotkey to download all the related images to local so that my LLM can easily reference them.

IDE:
I use Obsidian as the IDE "frontend" where I can view the raw data, the the compiled wiki, and the derived visualizations. Important to note that the LLM writes and maintains all of the data of the wiki, I rarely touch it directly. I've played with a few Obsidian plugins to render and view data in other ways (e.g. Marp for slides).

Q&A:
Where things get interesting is that once your wiki is big enough (e.g. mine on some recent research is ~100 articles and ~400K words), you can ask your LLM agent all kinds of complex questions against the wiki, and it will go off, research the answers, etc. I thought I had to reach for fancy RAG, but the LLM has been pretty good about auto-maintaining index files and brief summaries of all the documents and it reads all the important related data fairly easily at this ~small scale.

Output:
Instead of getting answers in text/terminal, I like to have it render markdown files for me, or slide shows (Marp format), or matplotlib images, all of which I then view again in Obsidian. You can imagine many other visual output formats depending on the query. Often, I end up "filing" the outputs back into the wiki to enhance it for further queries. So my own explorations and queries always "add up" in the knowledge base.

Linting:
I've run some LLM "health checks" over the wiki to e.g. find inconsistent data, impute missing data (with web searchers), find interesting connections for new article candidates, etc., to incrementally clean up the wiki and enhance its overall data integrity. The LLMs are quite good at suggesting further questions to ask and look into.

Extra tools:
I find myself developing additional tools to process the data, e.g. I vibe coded a small and naive search engine over the wiki, which I both use directly (in a web ui), but more often I want to hand it off to an LLM via CLI as a tool for larger queries.

Further explorations:
As the repo grows, the natural desire is to also think about synthetic data generation + finetuning to have your LLM "know" the data in its weights instead of just context windows.

TLDR: raw data from a given number of sources is collected, then compiled by an LLM into a .md wiki, then operated on by various CLIs by the LLM to do Q&A and to incrementally enhance the wiki, and all of it viewable in Obsidian. You rarely ever write or edit the wiki manually, it's the domain of the LLM. I think there is room here for an incredible new product instead of a hacky collection of scripts.

[Taeyoung Kim @tykimo]

Add one prompt to your AI assistant and every conversation becomes a passive language lesson. No textbooks, no flashcards, no study time. Just copy one file to your Claude Code setup and learn while you work. 23 languages, 506 pairs, all open source. https://t.co/FGGnP79rBp

[Taeyoung Kim @tykimo]

🦀 Cheliped Browser — open-source browser automation for AI agents. Converts any webpage into Agent DOM. 3,512 avg tokens · 40ms extraction · 88.9% accuracy. Works as a Claude Code & OpenClaw skill
👉 https://t.co/uf11hou0b2

[Harrison Chase @hwchase17]

a lot of engineering orgs (Stripe, Ramp, Coinbase) are building internal cloud coding agents

we're releasing a fully OSS one today - every company should have the power of cloud agents at their fingertips

[Evan Luthra @EvanLuthra]

I literally can’t believe Anthropic released this for FREE.

A 33-PAGE CHEAT SHEET on how to master Claude skills.

Bookmark this or get left behind in AI: https://t.co/RIcsy4tWrN

[Sydney Runkle @sydneyrunkle]

🤖 So you think you need a multi-agent system? This week I'll cover deep dives into 4 architectures:

1. Subagents: a main agent coordinates specialized workers as tools
2. Skills: load specialized capabilities on-demand without switching agents
3. Handoffs: agents pass control to each other for different workflow stages
4. Router: classify queries and route to the right specialists in parallel

Each pattern excels with different constraints: parallel execution, conversation state, and context isolation.

Full breakdown: https://t.co/s3Qqnayrpa

[Sydney Runkle @sydneyrunkle]

Most projects don't need a multi-agent system. But when you do...

I spent the last few weeks researching 4 architectural patterns to find out when you actually need each one.

👥 Subagents (centralized control)
💡 Skills (progressive disclosure)
🔄 Handoffs (sequential workflows)
🧭 Router (parallel synthesis)

Check out my blog post with a decision framework, benchmarks, and tutorials for each pattern!

📖 https://t.co/UWy14K6OiH

[Taeyoung Kim @tykimo]

I tried folding dishcloths, socks, and towels using two LEGO Technic–based 6-DoF robot arms. #PhysicalAI #Robotics #LEGO