Mango Du

This account is verified. Learn more

@mangodu

Co-founder @XmindHQ. 20 years of thinking tools. Studied fashion at Marangoni. Giving shape to things that don't have one yet.

SYD / SG / HK xmind.com Joined December 2008

119 Following 152 Followers 68 Posts

Latest posts 5

Pinned post

Mango Du @mangodu Oct 11

Just showed @tim_cook how @XmindHQ brings visual thinking to life on the new iPhone Duo! 📱✨

From a non-CS team inspired by Mac design back in 2006 to building a native experience for millions worldwide, Apple’s standards have always guided us.

Grateful for the amazing chat and encouragement, Tim! Onward! 🚀

Mango Du @mangodu 2h

Agents can take notes for you. They can’t understand for you.

That part still happens on paper for me. Slow hand, eyes on the page, brain catching up.

Then I snap a photo, Xmind turns it into a map, and the paper version becomes something I can keep building on.

Mango Du @mangodu Oct 22

A few days after demoing this to @tim_cook, woke up to incredible news: Xmind is officially featured in Apple’s Global Newsroom ahead of tomorrow's #iPhoneDuo launch! 🚀

Apple noted how Xmind "brings mind maps and outlines together in a side-by-side view, letting users explore the big picture while working through the details."

We are Day 1 ready. But this dual-screen canvas is just the beginning. We built this space for something even bigger:

The all-new Xmind Agent is landing on iOS very soon. 🧠✨

Massive congrats to the @XmindHQ team!

0:03

Mango Du @mangodu Oct 6

What makes us irreplaceable in the age of AI agents?

I’ve been thinking about this a lot lately.

As AI gets better at doing things, I believe our real value will increasingly come from knowing what’s worth doing.

Agency. Judgment. Taste. Context. Trust.

These may matter more than ever.

Maybe the biggest shift isn’t about what AI can do for us, but how we think, decide, and choose what matters.

Just a few thoughts I’ve been reflecting on lately.

0:45

Mango Du @mangodu Oct 5

A few hundred people replied to this. Almost nobody argued, they just went and built something. Three days of it in one picture.

Andrej Karpathy Verified account @karpathy Oct 2

We'll be spending a lot more time trying to understand the outputs of language models. A few thoughts, tips & tricks:

Writing. Something I've had success with: Ask your LLM to explain something in ASD-STE100, it's a controlled language specification originally developed for aerospace maintenance documentation. LLMs well-versed in this language and it comes with heavy constraints on clean writing style that I often find a lot more readable. Sometimes I've tried to soften it a bit e.g. ask for "80% of the way to ASD-STE100" because the spec is quite stringent. But even better:

Diagrams / images. Instead of writing, ask your LLM to create a diagram. These can be a lot easier to process, parse, and understand. But even better:

Web pages. Ask for output "in HTML" to get a beautiful, interactive webpage. LLMs are getting really good at frontend and can create beautiful experiences, animations, etc. But even better:

Explainer videos. The output format I am most bullish on is fully custom / bespoke explainer videos generated on any arbitrary topic. Experiment with things like "Create a 3b1b style video explainer on X. Use my ElevenLabs API key for audio narration". (you'd need an API key for the latter or you can ask your LLM to find you decent free alternatives that use your local compute). This is actually starting to work!

In summary:
- As LLMs get better, they will do more and more of the legwork autonomously, and a lot more of our work will rise up the abstractions into oversight and understanding.
- Luckily, LLMs can help here too because as intelligence and code are increasingly abundant, you can ask for large, custom, discardable software artifacts (e.g. web apps, video explainers) that would have never made sense to create before. Push the boundaries here and you'll be surprised.

View quoted post on X