In my first investor meetings, I was the quiet one in the room. I didn't bring commanding energy, nor rehearsed one-liners. Just someone trying to be straight about what I was building and why. What I didn't expect was that this would actually work. What showed was the ability to grasp complex topics fast and explain them back clearly. When a hard question came in, I didn't get flustered but I broke it down and answered it straight. Over a few meetings, I think they started to see something else too: that I was an operator. Someone who would organize, execute, and not leave loose ends. But the thing that closed it was commitment. I showed I would go the extra mile because I was performing conviction and because I genuinely would. They could feel the difference. I've since read a lot about what makes a great fundraiser. The charisma, the timing, the nervous system regulation when the room is ready to write a check. All real. But there's another path nobody talks about enough. Investors don't just back ideas. They back people they believe will figure it out. And "will figure it out" doesn't always look like confidence. Sometimes it looks like depth. The person who does the homework, follows through on the small things, shows up the same whether there's a deal on the table or not. I wasn't the loudest founder in the room. But I think they felt like no obstacle between now and the outcome they wanted would stop me from grinding through it. That's its own kind of conviction. Quieter than most people describe it. But the investors who've worked with me longest recognized it the moment they saw it. You don't have to change who you are to raise. You have to make the thing that's true about you impossible to miss.

A turtle “claw fluttering”, a mating ritual where a male rapidly waves his front claws near the female’s face to attract her. Who says romance is dead. Instead of butterfly kisses, they should be called turtle kisses. "Girl, my man, he fluttered me so good yesterday, I couldn't see straight for an hour!"

Today’s #China# news had many layers. @realDonaldTrump in Beijing; Xi hosting a high-stakes summit; @elonmusk making faces... But for me, there was another very human moment. I spotted a familiar face on TV: Cheng Lei. We used to do TV crossovers back in the CCTV days, when I was at @Reuters which had a partnership with CCTV for financial news. (That kind of partnership is almost unimaginable today.) And Cheng Lei was one of the most recognizable English-language business anchors in China. Sharp, professional, fluent across cultures. A face of China’s opening to the world. Then one day - she was arrested as a spy! In 2020, Cheng Lei was detained in China. She was accused of supplying state secrets overseas and spent more than three years in custody before being released in 2023. By her own account, the ordeal was not only legal or political. It was deeply human: isolation, uncertainty, separation from her children, and the mental pressure of being caught inside something much larger than herself. She lost 1,154 days. And today, there she was again — on Australian television, analyzing Trump’s China visit. That image gave me déjà vu — and goosebumps. Because U.S.-China relations are not only about tariffs, semiconductors, Taiwan, AI, supply chains, or state banquets. They are also about people. Journalists. Translators. Entrepreneurs. Students. Families. Immigrants. People who once believed they could move between worlds, explain one side to the other, and build bridges through language, trust, and shared curiosity. Some of those bridges are now broken. Some became dangerous and some people paid a very real price. Cheng Lei’s story is a reminder of how quickly the world can change — and how personal geopolitics can become. The silver lining is - after years of silence, she can return to the screen with her own voice. That is resilience. Watching her today, I felt both a sense of sadness and admiration. Sadness for the world we lost. Admiration for a woman who survived, came home, rebuilt her life, and returned to journalism on her own terms. In the middle of another Trump-Xi summit, maybe this is the angle worth remembering: Great power politics is never abstract. Behind every diplomatic reset, every strategic rivalry, every headline about “China” and “America,” there are human lives being reshaped. And sometimes, the most powerful story is not only the leaders on the stage. It is the familiar face who made it back to the screen. 🫶

Germany is in panic over Chinese companies taking over its markets at home, abroad and in China. Having spent quite some time in China in the 2010s, I can’t help but thinking back to the arrogance and quite frankly racism of German expat businesses people. Chinese engineers with Ivy League degrees speaking fluent English and German looked down upon by German engineers with a degree from some German university, unimpressive English and less than two sentences of Chinese. The Germans didn’t see it coming because they couldn’t imagine Chinese people becoming better at what they are doing than themselves. An industry insider told me at the time how keen Chinese entrepreneurs were to collaborate with German car companies on EV development. But all the Germans worried about was they are going to steal our IP. Well, here we are.

Coding agents are accelerating different types of software work to different degrees. When we architect teams, understanding these distinctions helps us to have realistic expectations. Listing functions from most accelerated to least, my order is: frontend development, backend, infrastructure, and research. Frontend development — say, building a web page to serve descriptions of products for an ecommerce site — is dramatically sped up because coding agents are fluent in popular frontend languages like TypeScript and JavaScript and frameworks like React and Angular. Additionally, by examining what they have built by operating a web browser, coding agents are now very good at closing the loop and iterating on their own implementations. Granted, LLMs today are still weak at visual design, but given a design (or if a polished design isn’t important), the implementation is fast! Backend development — say, building APIs to respond to queries requesting product data — is harder. It takes more work by human developers to steer modern models to think through corner cases that might lead to subtle bugs or security flaws. Further, a backend bug can lead to non-intuitive downstream effects like a corrupted database that occasionally returns incorrect results, which can be harder to debug than a typical frontend bug. Finally, although database migrations can be easier with coding agents, they’re still hard and need to be handled carefully to prevent data loss. While backend development is much faster with coding agents, they accelerate it less, and skilled developers still design and implement far better backends than inexperienced ones who use coding agents. Infrastructure. Agents are even less effective in tasks like scaling an ecommerce site to 10K active uses while maintaining 99.99% reliability. LLMs' knowledge is still relatively limited with respect to infrastructure and the complex tradeoffs good engineers must make, so I rarely trust them for critical infra decisions. Building good infrastructure often requires a period of testing and experimentation, and coding agents can help with that, but ultimately that’s a significant bottleneck where fast AI coding does not help much. Lastly, finding infrastructure bugs — say, a subtle network misconfiguration — can be incredibly difficult and requires deep engineering expertise. Thus, I’ve found that coding agents accelerate critical infrastructure even less than backend development. Research. Coding agents accelerate research work even less. Research involves thinking through new ideas, formulating hypotheses, running experiments, interpreting them to potentially modify the hypotheses, and iterating until we reach conclusions. Coding agents can speed up the pace at which we can write research code. (I also use coding agents to help me orchestrate and keep track of experiments, which makes it easier for a single researcher to manage more experiments.) But there is a lot of work in research other than coding, and today’s agents help with research only marginally. Categorizing software work into frontend, backend, infra, and research is an extreme simplification, but having a simple mental model for how much different tasks have sped up has been useful for how I organize software teams. For example, I now ask front-end teams to implement products dramatically faster than a year ago, but my expectations for research teams have not shifted nearly as much. I am fascinated by how to organize software teams to use coding agents to achieve speed, and will keep sharing my findings in future posts. [Original text: ]