Martin @MartinMaBuild

How to get promotion? 1. Take initiative in deciding what to work on — don’t wait for your manager to assign tasks. 2. When you identify problems, also propose concrete solutions. 3. Speak up in meetings — don’t wait until you’re called on to contribute. 4. Focus on managing outcomes, not just completing tasks. 5. Don’t always wait for permission — take ownership, move forward, and seek forgiveness if necessary. 6. Have the courage to confront difficult issues instead of avoiding discomfort. 7. Manage up effectively — proactively communicate your progress and impact to leadership. 8. Think bigger — at the organization or company level — instead of focusing only on your own tasks or team’s success.

AI Security is super important, here is what did to make the security: 1 run openclaw in sandbox or stand-alone Mac mini 2 review the skill code before download and use third party skill 3. the prompt instruct the AI does not delete anything resources 4. make sure the human in the loop

do you feel the AI bubble taste from this news: In March this year, Crypto.com founder Kris Marszalek purchased the domain name AI.com for $70 million, setting a new record as the most expensive domain transaction in history. The entire payment was made to the seller in cryptocurrency. According to recent reports, the seller of the domain was a Malaysian man named Arsyan Ismail. He originally bought the domain in 1993 for $100 using his mother’s credit card — at the time, he was only 10 years old. His mother was confused by the charge because she didn’t understand what a domain name was; she simply noticed a mysterious deduction on her credit card statement. Ismail said that his decision to purchase the domain was entirely personal: the letters “AI” happened to match his own initials. He had no foresight about the rise of artificial intelligence, nor did he have any long-term strategic plan.

After years in AI infrastructure and HPC, I’m now building my own AI-driven product. Reflecting on my experience optimizing large-scale compute systems, I see how accessible AI tools will shape the future. I’ll share insights on scaling, practical AI infra, and lessons learned building in public. Follow along if you’re interested in the intersection of AI, infra, and real-world applications!

I want to become people like @elonmusk, cracking the hardest problem, building the creative new world, bring the value to the human being. where can i start to achieve this goal?

How to Fix Your Entire Life in 1 Day (for AI Builders) — inspired by Dan Koe Most AI builders don’t have a motivation problem. They don’t have a discipline problem. They don’t even have a skill problem. They have a direction problem. We live in an era where tools are abundant, models are improving weekly, and information is infinite. And yet, many builders feel stuck, anxious, or burned out—despite “doing everything right.” Dan Koe makes a simple but uncomfortable point: You’re not failing because you’re lazy. You’re failing because you’re operating under the wrong default system. That system is built on constant input. Notifications. Feeds. Trends. Benchmarks. Other people’s success. You wake up and immediately consume before you decide. So the solution isn’t another productivity hack or AI workflow. It’s one intentional day of reset. First: Cut the inputs. No social media. No news. No tutorials. No “how others are doing it.” Not forever—just for one day. When you remove external noise, discomfort appears first. Anxiety. Restlessness. The urge to check. That’s not boredom. That’s withdrawal. And if you push through it, something rare happens: clarity. Second: Reset identity, not goals. Don’t ask, “What should I build?” Ask: Who do I want to become as a builder? If no one was watching, what would I still work on? What kind of work would I regret not pursuing in 10 years? Most people chase goals that don’t belong to them. Clear identity makes decisions obvious. Without it, even the best execution is wasted. Third: Subtract ruthlessly. This is where real leverage comes from. Not: Add more habits Add more features Add more side projects But: What 20% of my activities drain 80% of my energy? What work looks productive but doesn’t align with who I’m becoming? If I could only keep three priorities, what survives? Progress accelerates when misaligned work is removed. You don’t “fix your life” in one day by solving everything. You fix it by making one clear decision about direction—without distraction. For AI builders especially, clarity beats speed. A wrong direction with AI scales failure faster. A clear direction turns even simple tools into leverage. You don’t need more grind. You need one quiet day to choose deliberately— then let consistency do the rest.

I have coded 7 apps within these 3 weeks using cursor + Claude Code. I told my boss, in the future, nobody need "coding", only stupid people need learn the algorithms, just like learning to use paper and pencil to get the square foot of 2

today I did not build anything, sometimes we run all the way to somewhere, need stop and confirm the directions

we builders have to learn how to promote the product we build, we can use AI to build the promoting system

@saviomartin7 why suddenly this project form consulting business mode to "crypto" mode?

@YutaWeb33 hello

I saw a lot of people struggling to setup the OpenClaw, do you think it is good idea to help them run a docker on the VPS, charge them some fee. but how can they trust me, and it won't scale as they will go away after they realize the toy is not that fun (as I am feeling now)

what happening to the crypto and stocks today?

@saviomartin7 very nice. do you pre-create the docker for each user in VPS?

morning, i just created the to do list for myself and my 5 coder agents

built an app to verify your paper references and find citations to support your claim citetruth.com

Third thing: sometimes you just need to find papers on a topic. Describe what you're writing about, get papers ranked by how strong the evidence is. citetruth.com - free to try. Still early. If you use it, tell me what's broken.

The harder problem: citations that exist but don't support your claim. We've all gotten that reviewer comment. "Reference 14 does not support the assertion on page 3." So there's a second mode where you paste your claim + references and it checks whether they actually back you up.

Last year I almost submitted a paper with fake citations. A co-author used ChatGPT to "help" with the lit review. Three of the references didn't exist. Spent a weekend manually checking everything. Built a tool so I never have to do that again.