"Go all the way until it hurts. If you're doing something and it's easy, it's not valuable." - @travisk
"If anyone says a strategic thing was easy, I'm like, 'You messed up. You could have gone way further. More competitive advantage. More differentiation. Get it together.'"
I've built a Claude Code viewer so you can see what's happening under the hood.
Gona use it to learn what are the agentic patterns that made claude code sooo darn good.
Will defn use it to deep dive:
1. Subagents
2. Architecting tool calls and tool results
3. In general how the claude code harness is architected
4. Auto-compacting
I've open-sourced the github repo so grab it in the thread to visualise your own claude code traces 🧵
Would love to hear what new tricks you've learnt from Claude Code that is useful for building effective agents!
“Why don’t you guys have a fancy demo video?” For starters we’re solving unsexy problem in the construction industry. But rather than a fancy demo, what we have instead is deep understanding of our customer’s workflow and how our AI Agents can help teams unlock real productivity!
The total value of all public companies in the tiny Bay Area (population 8M) is greater than India, Japan and Germany (population: ~1680M) combined.
There’s a reason why it’s called the global hub of innovation.
We use Devin with Claude code and it’s truly game changing. We have a few services that interact with each other and Devin’s ability to simultaneously update both is helping to speed up our deployment cycle.
Talked to a founder whose workflow is multiple AI coding agents running in the background for each of the core services of his app and marketing site. His time is spent on prompt design, code review, fixing issues, and integration. Says he executes 10X faster. Wild future.
Untold - The Immortal Blades Saga
So beyond excited to share this concept trailer - Untold is a story I have worked on for over 8 years and next year, we will be making it into a reality with season 1 of the series. This project is probably the one I hold most close to my heart - and this journey we have been on has been beyond words.
Discover your Destiny 2026
when you can quickly implement and test ideas, you learn faster. when you learn faster, your intuition improves. when your intuition improves, you choose better experiments. when you choose better experiments, you get better and more interesting results. when you get better results, you have more confidence to try more ambitious ideas.
this holds true both for product, research, org design work
Zuck poached 4 elite Chinese researchers from OpenAI:
- Jiahui Yu: led o3, o4-mini, GPT-4.1
- Hongyu Ren: creator of o3-mini & o1-mini, core contributor to o1
- Shuchao Bi: head of post-training multimodal org
- Shengjia Zhao: key contributor to GPT-4 & o1
These are undeniably some of the best researchers at OpenAI.
Llama 5's about to flex hard.