hidecloud
@hidecloud
Co-founder of @ManusAI_HQ
See you in SF!
What if better work didn’t mean more work? Join us in San Francisco on July 18th for an evening with @hidecloud , co-founder of Manus AI, as we explore how AI-native workflows are helping teams move faster, focus deeper, and build better. This isn’t just about automation. It’s…
Working with data often means juggling multiple spreadsheets and spending hours turning them into clear, compelling charts. It takes time, the right tools, and know-how. For event marketers, that challenge starts right after the event: Turn raw event data into structured…
Forgot to include a key lesson in the blog: Context engineering can also overfit - not just to specific model families, but also to today's model limitations via premature optimization. At Manus, we never commit to an architecture based on static benchmarks. Instead, we treat…
After four overhauls and millions of real-world sessions, here are the lessons we learned about context engineering for AI agents: manus.im/blog/Context-E…
My point isn't that having deep pockets automatically guarantees a win. Burning through a large volume of tokens is a result, not a prerequisite for success. What often goes unnoticed is the capability required to even be able to consume so many tokens in the first place. It's…
Money is the best A/B test: Those who have managed advertising budgets of over 100 million have a completely different perception of growth. AI also needs similar ‘tuition fees of hundreds of millions of tokens’ - without extensive hands-on experience with large volumes of…
Money is the best A/B test: Those who have managed advertising budgets of over 100 million have a completely different perception of growth. AI also needs similar ‘tuition fees of hundreds of millions of tokens’ - without extensive hands-on experience with large volumes of…
Great to host our best San Francisco meetup ever. We evolved so much since four months ago. And our users also are getting more and more out of Manus. They just shocked me with their use cases tonight!
After four overhauls and millions of real-world sessions, here are the lessons we learned about context engineering for AI agents: manus.im/blog/Context-E…
This is our most expensive lesson building Manus. Cost us tens of millions to learn. Hope it saves you some pain.
After four overhauls and millions of real-world sessions, here are the lessons we learned about context engineering for AI agents: manus.im/blog/Context-E…
That’s true. In last December when we had the first working prototype of @ManusAI_HQ. Watching it solved complex tasks with self planning, reflection, tool-use, is the moment I actually cried. I think that’s also the reason why many users love Manus when it came out on March 5th.…
watching chatgpt agent use a computer to do complex tasks has been a real "feel the agi" moment for me; something about seeing the computer think, plan, and execute hits different.
I use a metric called PPD (Product Manager to Prompt Distance) to evaluate if a company is ready for AI products. It measures how many people a PM needs to talk to before changing a prompt. Ideally, it should be zero - where the PM writes the prompt directly. An acceptable…
My two favorite studio together! What a day.
At Bilibili headquarters with Yang Qi and Feng Ji from Game Science, the studio behind Black Myth: Wukong 📸
Had a great time sharing Manus’s latest updates and thoughts at the @RaiseSummit yesterday. Been to the Louvre many times but never knew there was such a massive conference space underneath! The evening got even better when Eurotech invited me to speak about personal growth and…
