🚨 Building AI Agents on-chain?
Welcome to the Skill-to-Agent Dual Cascade Hackathon AMA! A 6-week hackathon with a 50,000 $PROS prize pool hosted by @pharos_network and @AnvitaFlow.
📅 June 8 (Mon) 20:00 UTC+8
👇Set your reminder
https://t.co/OTji6FrPj8
Join us as we break down:
🔹 Hackathon tracks & judging criteria
🔹 Key timelines & submission requirements
🔹 How to build Skills and AI Agents on Pharos
🔹 Live Q&A with the team
Guest Speakers
Wishlonger @wishlonger Co-founder & CEO of Pharos Network
Cobe Zhang @hszcg Founder of AnvitaFlow
Host qiuqiu @0xqiuqiuu BuilderHero of @OpenBuildxyz
🎁 Giveaway
Follow @pharos_network@AnvitaFlow@OpenBuildxyz and join the AMA, 5 lucky participants will share 50 USDT!
💬More Questions
https://t.co/3VQgQafZz0
The next build starts here. ⚡
Monad Blitz is landing in Beijing!
Join us for a one-day hackathon where you'll build, ship, and demo your ideas on Monad. @monad@monad_zw@Monad_APAC
Whether you're exploring AI Agents, programmable payments, DeFi, consumer apps, infrastructure, or any other Monad-native innovation, this is your chance to turn ideas into working products.
🏆 $1,500 Prize Pool
🗓 August 15, 2026
📍 Beijing, China
Bring your ideas. Find your teammates. Build fast.👉
Register now: https://t.co/ukWJYdkShX
#Monad #MonadBlitz #Hackathon #AIAgents #Web3 #BuildOnMonad #OpenBuild
5/5
A few things stood out this week:
🚩AI demand is driving growth in China’s chip industry.
🚩Open models are moving closer to the capital market.
🚩AI agents are becoming a major source of token usage.
That’s it for this week.
Follow @OpenBuildxyz for more updates on China’s AI and innovation ecosystem.
See you next Monday.
1/5
What did you miss in China’s AI last week?
CXMT made a major market debut, Moonshot AI fully opened Kimi K3, and AI agents pushed token usage to new highs.
Time for last week’s recap 👇
4/5
China’s AI model token usage has now ranked first globally for 13 weeks in a row.
The gap is still growing.
The main driver is no longer chatbot conversations.
It is AI agents running tasks on their own. In finance, government services, and manufacturing, agents are working around the clock.
Why is asking good questions becoming more valuable than writing code?
A few years ago, being able to code already made you stand out.
Now AI can write a lot of the code for you. The harder part is knowing what to ask it to build in the first place.
🚩1. More people can build. Not everyone knows what is worth building.
AI is pretty good at answering “How do I make this?”
But it cannot fully decide what the real problem is, who you are solving it for, or whether the idea makes sense at all.
When the question is unclear, faster answers do not help much.
🚩2. Asking a good question means you have already done some thinking.
You need to know what you want, what matters, what the limits are, and what you are willing to leave out. That is less about execution and more about judgment.
Once those decisions are clear, the code is often the easier part.
🚩3. AI can also help you go in the wrong direction faster.
Give it a vague or badly framed problem, and it may still produce something that looks polished and convincing.
That is the dangerous part.
The output can look good even when the direction is wrong.
🚩4. The tool is usually not the biggest difference.
Two people can use the same model. One gets what they need in five minutes. The other keeps rewriting prompts for hours.
Most of the time, the difference is whether they understand the problem well enough to ask the right question.
Coding is becoming easier to access. But knowing what to ask is becoming harder to replace.
When you use AI, do you struggle more with how to ask, or with knowing what to ask?
#AI #PromptEngineering #FutureOfWork
Proud to see AnySearch's first developer meetup come to life in Beijing 🎉
The room was full of developers, founders, and AI practitioners all gathered around one question: what should search infrastructure look like when the primary consumer shifts from humans to AI Agents?
Huge thanks to @AnySearchAI , @datawhale2018, and everyone who joined us.
#AnySearch #AIAgents #SearchInfrastructure #AIInfra
That’s a wrap on AnySearch’s first developer meetup! 🎉
Developers, founders, and AI practitioners gathered in Beijing for lively discussions on how search infrastructure is evolving in the Agent era. A huge thank you to @OpenBuildxyz , @datawhale2018, our speakers, and everyone who joined us.
Since launch, AnySearch has grown to 𝟐𝟎𝟎,𝟎𝟎𝟎+ 𝐝𝐞𝐯𝐞𝐥𝐨𝐩𝐞𝐫 𝐮𝐬𝐞𝐫𝐬 and processed 𝟐𝟎 𝐦𝐢𝐥𝐥𝐢𝐨𝐧+ 𝐬𝐞𝐚𝐫𝐜𝐡 𝐜𝐚𝐥𝐥𝐬 worldwide. A huge thank you to every developer who has supported us—we look forward to seeing you again at our next event!
Explore AnySearch: https://t.co/JNUGTMrsAB
#AnySearch #AIAgents #Developers #AISearch #AIInfra#Beijing
The EAG 2026 Global Application & Builder Initiative is well underway🧑💻
As the Ethereum ecosystem enters its next phase of application innovation, EAG is expanding structured regional builder programs that connect developers, universities, founders, and local communities to a global application acceleration pathway.
Here's how builders can get involved. 👇
Is “anyone can build a product” actually a lie?
This phrase has been everywhere lately. On one side, it’s true. Building has become dramatically easier. AI can help people write code, design interfaces, create content, and turn ideas into working products faster than ever before.
But there’s a difference between being able to build something and being able to make something people actually want.
That’s where the phrase gets misunderstood.
🚩 The barrier dropped. The success rate did not.
AI lowered the cost of execution, but it didn’t solve the harder problems: finding real demand, understanding users, making the right decisions, and staying long enough to improve.
A product doesn’t fail because someone couldn’t ship it. It fails because nobody needed it.
“Anyone can build” describes the starting line, not the finish line.
🚩Before, technical ability was often the first barrier. Today, more people can get past that stage.
But the difficult parts remain:
-Finding a problem worth solving
-Getting the first users
-Learning from feedback
-Making decisions with limited information
- when growth is slow
🚩Shipping is only the beginning.
The real competition starts after launch.
When everyone can build faster, the difference comes from judgment: knowing what to build, who to build for, and how to keep improving.
So no, “anyone can build” isn’t a lie.
Anyone can start building is true.
Anyone can create something people need and make it succeed has never been true.
1/6
What did you miss in China's AI last week:
Time to for lats week's recap! Kimi K3 confirms its scale, Alibaba enters the trillion-parameter race, and AI is starting to take a bigger role in scientific research.
👇
5/6
A few things stood out this week:
🚩Open-source models are moving closer to the frontier.
🚩Large-scale models are making a comeback after months of focus on efficiency.
🚩AI is moving beyond generating content and becoming a tool for research and experimentation.