Weekly Retrospective — June Week 2, 2026

I visited the COEX Smart Tech Expo 2026. There were really a lot of people, and since it was an expo covering not just AI but also a wide range of themes like robotics and the multiverse, it was a large exhibition that used all of Halls A, B, C, and D. Jensen Huang had stopped by, and in line with the trend of moving beyond AI toward robotics, you could see that the number of companies working in the robotics field has grown enormously. I'm not in robotics myself, but I could feel why so many people around me dream of starting robotics companies.

At the AI expo, most companies were running their services through LLM-based API connections, and there were no cases of companies that had their own AI models. And that makes sense—rather than building your own model, what should come first is whether the service is one that's actually needed and whether there are customers for it!

It also seems like what's now leading the truly massive market has shifted from simple AI services to Physical AI. The trend was moving toward robotics, manufacturing process automation, and facility/disaster safety. And the exhibition halls for the companies in those areas were larger, too. You could really feel that the market for simple AI services has tipped over into saturation.

I had assumed these were just companies that had applied simple AI to their services, but it turned out that most of those companies had already secured large B2B customers, such as major corporations, for over 10 years. The fact that they were leading the market with features that looked simple was actually the result of long years of accumulated effort. It was wonderful to see that Korea has so many great companies, and that every time I look, there are so many new and impressive services and products.

I did hear someone say that people who work at startups should focus on product development and meet more customers rather than going around to expos like this... but either way, going occasionally just to get some motivation doesn't seem like a bad thing to me.

Seeing the shifts in the industry became an occasion for me to think about the service I should be building. It made me realize that the service I'm working on doesn't have any remarkable technical edge, nor is it targeting a niche customer. I'll need to review and adjust further—checking whether this is the right direction and whether there are people who actually need it—by doing the 5 Ways.