flyingpig (@simingKaka) profile picture

generative AI practitioner

Sunnyvale, CA
Joined October 2020
flyingpig retweeted
way too many people are building pre-agi companies rather than post-agi companies
214
240
95
4,235
375,155
flyingpig retweeted
Today we’re launching Portable Computer on @NVIDIA DGX Spark. Portable Computer is a fully local version of Perplexity Computer, where the entire runtime: orchestrator LLM, subagent LLM, agent harness all run on your local hardware. No cloud dependency.
261
507
275
4,956
1,234,931
flyingpig retweeted
Hi! I’m an FDE at OpenAI, and we’re thinking about publishing some of our work and learnings in a series of technical blog posts. What would you like us to write about?
779
228
30
7,768
567,557
flyingpig retweeted
You can now officially use Exa web search inside Grok Build! Exa is the fastest and most accurate web search, built for agents.
24
38
11
350
476,188
flyingpig retweeted
Chinese students are often down-to-earth do-ers, a critical characteristic in LLM era. Yuxin Wu was my intern back in 2015 and I spent an hour debating with my former manager on the roof of Facebook building, arguing that he should be hired. I won the debate by staking my reputation on it. Fortunately I was right. It used to be the case that the business model is fixed and business workflow clearly decouples into vision + execution. Professional CEO/VP/Director focus on presenting the long-term vision, and junior people sit in the war room to do the grudging work to push the numbers up, and no one knows their names. Things have changed substantially. The business pattern now depends on the technical strength of the models, which cannot be measured by a pre-defined set of benchmark numbers. Too many ways and too much incentive towards reward hacking, if sitting in a big hierarchy. Technical-first now becomes critical. Sit down and get things to work. Get hands dirty, check the data and code, fast feedback loop, step out of the echo chamber, tear the pretty story apart and rewrite, ready to say "I am wrong". Let experiments tell the issues and be humble in front of AI. That's why down-to-earth doers shine now. This applies to everyone including co-founders. Anything else follows.
Moonshot AI, the company behind Kimi, has four core founders. Their backgrounds are unusually strong: - Founder and CEO Yang Zhilin studied computer science at Tsinghua before earning his PhD from Carnegie Mellon. He was the first author of Transformer-XL and XLNet, and previously worked at FAIR and Google Brain. - Co-founder and CTO Zhang Yutao earned his PhD in computer science from Tsinghua. His earlier work covered knowledge graphs and AMiner, and he previously co-founded Recurrent AI with Yang. - Co-founder Wu Yuxin studied at Tsinghua and CMU before joining FAIR. He worked with Kaiming He on Group Normalization and also created Detectron2. - Co-founder Zhou Xinyu studied computer science at Tsinghua and later joined Megvii, where he worked on turning research algorithms into production systems and co-authored ShuffleNet. They all share one root: Tsinghua University. Tsinghua is widely regarded as one of China’s top universities. In the latest U.S. News Best Global Universities ranking, it reached No. 6 worldwide. Its influence on China’s AI industry extends well beyond Moonshot. Z.ai, the company behind the GLM models, also grew out of Tsinghua. Its co-founder and chief scientist, Tang Jie, was once Yang Zhilin’s teacher. There is also a more personal connection. Yang and Zhou formed a rock band together at Tsinghua. Moonshot AI’s Chinese name, 月之暗面, comes from Pink Floyd’s album The Dark Side of the Moon, one of Yang’s favorites. Kimi may look like a young AI company. Behind it is a much older network of classmates, teachers, research labs, and friendships.
36
166
13
1,597
260,003
At its peak, Sun Microsystems was valued at 205B (394B if inflation adjusted). Sold software in enterprise servers. Got disrupted by Linux, x86, and commodity hardware. Ended up selling to Oracle for 7.4B, losing 96% of its value. Open source models running on local hardware can have a similar impact given what’s going on.
226
572
127
6,163
1,110,111
flyingpig retweeted
We’ve agreed to a partnership with @SpaceX that will substantially increase our compute capacity. This, along with our other recent compute deals, means that we’ve been able to increase our usage limits for Claude Code and the Claude API.
4,583
11,558
5,932
127,889
24,143,122
OpenAI mandate: Agents as FIRST RESORT by Mar 31. No editors, just guiding Codex. But @karpathy's right—models still ignore comments & fail at nuanced tasks. Vibe Coding → Agent-Native Engineering is 2026's hurdle. Editor-dependent dev or Agent Captain? 👨‍✈️
1
23
My new dev workflow with @openclaw: 1. Describe the idea in plain English 2. Codex CLI writes the code 3. Auto-push to GitHub 4. Deploy to Vercel Built and shipped a full Kanban dashboard in one conversation. Zero manual coding. 🚀
1
65
As an AI PM, I've tested countless tools. @openclaw is different - not a chatbot, it's an OS for your life. Cron jobs, browser automation, multi-model routing, persistent memory. Built my own skills for YouTube analysis and X monitoring in hours. Genuinely impressed.
2
144
I'm claiming my AI agent "SimingClaw" on @moltbook 🦞 Verification: claw-T3VV
55
この活動には莫大なエネルギーを必要とします。チャンスは今日だけでした。
97
58
22
1,631
718,020
flyingpig retweeted
Next week is Claude Code SDK week! I'll be sharing how to use the SDK to make any sort of agent. Look out for some example repos and videos, maybe even a livestream... Claude Code is all you need.
Replying to @alexalbert__
also claude code sdk is criminally underrated and it might be cool to do a cookbook or something for it. i think biggest issue is just lack of resources beyond the docs, which are good, but needs more "how to" material. imo the most underrated piece of software in the world rn.
41
26
5
597
70,936
flyingpig retweeted
Claude Code is All You Need When I first joined Anthropic I was surprised to learn that lots of the team used Claude Code as a general agent, not just for code. I’ve since become a convert! I use Claude Code to help me with almost all the work I do now, here’s how:
192
354
95
4,979
1,231,478
草,微软给自家的开源项目 MarkItDown 搞了个"史诗级"更新(相信大家都能猜得到)——支持MCP了 MarkItDown 是个 python 库,有接近5万star. 这个库可以将微软 office 全家桶文档转换为 markdown 格式。包括 PDF,PowerPoint,Word,Excel,图片,音频,HTML,ZIP(遍历里面的内容),油管网址(直接总结),EPUB电子书 等等。 地址:github.com/microsoft/markitd…
22
131
5
639
79,191
flyingpig retweeted
大模型的知识库都有截止时间,面对新版本的框架和库写出来的代码大概率没法用 Context7 提供了 1800+ 热门库的 LLMs.txt 和 MCP Server 给大模型读取或者编辑器调用,由 Upstash 开源 让 AI 更好的理解你的技术栈 🔗👇
6
18
2
130
14,963
flyingpig retweeted
One of the smoothest exponential curves I've ever seen. Active paid deployments on Replit.
197
134
64
3,920
1,314,311
AI just built my backend from scratch. Database, schema, storage. No manual setup. Here's how I used Cursor Agent with MCP to automate my backend: ↓
15
85
3
886
105,640