Wahchi

I was a developer and am becoming into builder.

[Wahchi @WahchiYu]

每个人都将拥有自己的拉斐尔老师

[Wahchi @WahchiYu]

When developing agent skills, I ran into a design question:
If an agent skill can execute Python scripts, and those scripts depend on third-party packages, where should those dependencies be managed?
Should they simply go into the main project's requirements.txt, or is it better for each skill to maintain its own dependency file/manifest?
Curious how others are structuring this in real-world agent/skill systems.

[Wahchi @WahchiYu]

@elonmusk would you try to create a trusted social app?

[Wahchi @WahchiYu]

@yangyi 🤔看管理多少agent，有没有带来效益？

[Wahchi @WahchiYu]

Before, we’d pray nothing broke at deployment. Now, we pray nothing breaks right after we finish the prompt.

[Andrej Karpathy @karpathy]

It is hard to communicate how much programming has changed due to AI in the last 2 months: not gradually and over time in the "progress as usual" way, but specifically this last December. There are a number of asterisks but imo coding agents basically didn’t work before December and basically work since - the models have significantly higher quality, long-term coherence and tenacity and they can power through large and long tasks, well past enough that it is extremely disruptive to the default programming workflow.

Just to give an example, over the weekend I was building a local video analysis dashboard for the cameras of my home so I wrote: “Here is the local IP and username/password of my DGX Spark. Log in, set up ssh keys, set up vLLM, download and bench Qwen3-VL, set up a server endpoint to inference videos, a basic web ui dashboard, test everything, set it up with systemd, record memory notes for yourself and write up a markdown report for me”. The agent went off for ~30 minutes, ran into multiple issues, researched solutions online, resolved them one by one, wrote the code, tested it, debugged it, set up the services, and came back with the report and it was just done. I didn’t touch anything. All of this could easily have been a weekend project just 3 months ago but today it’s something you kick off and forget about for 30 minutes.

As a result, programming is becoming unrecognizable. You’re not typing computer code into an editor like the way things were since computers were invented, that era is over. You're spinning up AI agents, giving them tasks *in English* and managing and reviewing their work in parallel. The biggest prize is in figuring out how you can keep ascending the layers of abstraction to set up long-running orchestrator Claws with all of the right tools, memory and instructions that productively manage multiple parallel Code instances for you. The leverage achievable via top tier "agentic engineering" feels very high right now.

It’s not perfect, it needs high-level direction, judgement, taste, oversight, iteration and hints and ideas. It works a lot better in some scenarios than others (e.g. especially for tasks that are well-specified and where you can verify/test functionality). The key is to build intuition to decompose the task just right to hand off the parts that work and help out around the edges. But imo, this is nowhere near "business as usual" time in software.

[Wahchi @WahchiYu]

It seems learning GO or Rust would be still necessary in recent years since most CLI is written by them. We will still need those two languages to build tools for AI.

[Andrej Karpathy @karpathy]

CLIs are super exciting precisely because they are a "legacy" technology, which means AI agents can natively and easily use them, combine them, interact with them via the entire terminal toolkit.

E.g ask your Claude/Codex agent to install this new Polymarket CLI and ask for any arbitrary dashboards or interfaces or logic. The agents will build it for you. Install the Github CLI too and you can ask them to navigate the repo, see issues, PRs, discussions, even the code itself.

Example: Claude built this terminal dashboard in ~3 minutes, of the highest volume polymarkets and the 24hr change. Or you can make it a web app or whatever you want. Even more powerful when you use it as a module of bigger pipelines.

If you have any kind of product or service think: can agents access and use them?

- are your legacy docs (for humans) at least exportable in markdown?
- have you written Skills for your product?
- can your product/service be usable via CLI? Or MCP?
- ...

It's 2026. Build. For. Agents.

[Wahchi @WahchiYu]

🤔Maybe openai is working on enterprise version?

[Peter Steinberger 🦞 @steipete]

Since I spend my night again sifting through security advisories, folks, security researches, slop clankers, PLEASE - read https://t.co/OUA1Z4GWMd and https://t.co/uUitTHjBqq

The security model of OpenClaw is that it's your PERSONAL assistant (one user - 1...many agents).

IT IS NOT A BUS. If you want to have multiple users that are adversarial to each other, use on VPS per gateway and user. (or Mac Minis, if you like spending money)

I closed like 20 reports today that try to force it into something it's was never designed for and that would just add loads of needless complexity and would introduce unnecessary bugs that won't benefit the wast majority of users.

[Wahchi @WahchiYu]

It seems now when learning a new coding language, we may write a tiny GPT instead of "hello world"😀

[Dr Milan Milanović @milan_milanovic]

I wanted to understand how GPT works, so I ported Karpathy's https://t.co/5FK9fuhDBF to C# from scratch. No frameworks and NuGet packages, just plain math in ~600 lines of code.

It builds a tiny GPT that learns from 32K human names and invents new ones. Every piece is there: autograd, attention, Adam optimizer, the works. Just at a scale you can actually sit down and read.

I also wrote a prerequisites guide that walks through all the math and ML you need, starting at a high school level. If you've ever wanted to peek under the hood of ChatGPT without drowning in linear algebra textbooks, this might help.

https://t.co/rwAB7pswzY

[Wahchi @WahchiYu]

最近在学假名，想到小时候我们学英语但没学音标前，会在英语单词标拼音，那么日本人学英语时会标假名么🤔

[Wahchi @WahchiYu]

Programmers after 2025: Writing .md files while Claude codes the app.

[Wahchi @WahchiYu]

My mind during runs has changed: it used to be about breaking problems into steps. Now it’s about how to write the right prompt so the problem can be understood clearly.

[Wahchi @WahchiYu]

@op7418 感觉可以画漫画

[Wahchi @WahchiYu]

@sama I think $AAPL could work with it, and the stock would be great again.

[Wahchi @WahchiYu]

@dotey 感觉可以用agi生成动图素材

[Wahchi @WahchiYu]

@elonmusk @grok Kong

[Wahchi @WahchiYu]

@sciencegirl512 same

[Wahchi @WahchiYu]

@elonmuskADO sure

[Wahchi @WahchiYu]

@elonmuskADO 9

[Wahchi @WahchiYu]

@elonmuskADO 18

[Wahchi @WahchiYu]

@elonmuskADO 18.125