In this episode of The FlipSide Podcast, we open with an AI that almost started a war. From there, we lay out the latest from both sides:· The computer is creeping from your pocket to your face. AI wearables are going to turn us all into cyborgs. The iPhone may be no better than Xiaomi or Huawei, but people try everything and still end up back on Apple's ecosystem.· AI is eating the web’s traffic. In the future you won’t need a website at all, just information AI can read.· The US has Google. China never got any all-in-one search engine like that. But in China you can’t get by without WeChat, the super app even Elon envies.· Ads are learning to talk back, but ChatGPT’s ads only reach the people who get the least out of AI.· Both sides worry the other’s AI is stealing their data, and the US is starting to borrow China’s playbook by taking stakes in companies.Then we circle back to the AI that nearly caused all that trouble. Chris says it’s no different from how we use AI at work every day: AI is great at making crap that looks perfect. AI getting things wrong isn’t the scary part. The scary part is when no one can tell.Fifty minutes with FlipSide, from both sides: how much can we really trust AI at this stage?

This week on FlipSide: Galen Gao, co-founder of Linkloud.Linkloud is one of the largest communities in China for founders building global SaaS and AI products. Alongside the events, growth workshops, and market-entry work, Galen runs study trips — taking groups of founders and operators to Silicon Valley, and separately to Tokyo, a week at a time, with twenty-plus sessions packed in with local founders and partners.Three years of doing that gives you a particular view of things. He came on to lay it out.Some of what stuck with us:

·Services companies are turning into AI companies. Instead of selling the tool, sell the outcome.

·Founders going global have stopped chasing the general-purpose product. They're going vertical, and starting to fine-tune small models on their own data.

·Most brands are still underestimating GEO, generative engine optimization. Meanwhile their customers are already asking AI to compare them against a competitor.

·Japanese enterprises are moving on AI faster than you'd expect, and they want to own their data and their models outright.

·The hardest part of operating across regions is rarely the strategy. It's hiring your first local employee.Near the end, Galen said something that reframed the whole conversation: we all live in a bubble. Outside Silicon Valley, plenty of people are adopting AI at maybe a third of our speed. And in Brazil, Southeast Asia, and India, teams are building for markets nobody was watching. One cracked video translation across twenty-plus Indian languages and grew in a way nobody predicted.The places the models can't reach yet might be exactly where the opportunity is.

Eighty minutes outside the bubble.

For years, China proudly wore the "copycat" label. In this episode, we watch it start to flip.A phone company that began making cars two years ago just built something Ford never thought of in a hundred years. Chris explains why: Xiaomi lives next to the factory floor. An AI prodigy who had an Apple offer, with permission to stay in Beijing, chose to build at home instead, and made Kimi. Apple is only now getting to a folding phone with their iPhone Duo; Huawei, Xiaomi, and Samsung already make one. And the advice that startups should steal IP first and hire lawyers later? That didn't come from China. Those are words straight out of ex-Google CEO's mouth.We also look at the other side of the coin: which of Y Combinator's most-wanted startup ideas belong in China and which only work in the US; why Anker wants off Amazon, and why brands in China are always renting their traffic; and where does that traffic go once AI starts answering the questions for us.So who's copying whom now? Forty-five minutes, two sides, you decide.