Yesterday I went to VivaTech. It was my third consecutive year attending since moving to France, and each visit has given me a different perspective.

The first time, I went as a student. Everything seemed novel, and I even brought a stack of résumés with me to look for an internship. The second time, as a newcomer to the workforce, I had no particular goal and only took a cursory look around. The third time, as someone exploring career possibilities, I had a better understanding of the tech industry and how businesses operate, so I found everything fascinating. I spent six hours there and still felt it was nowhere near enough.

The biggest difference between this year and last year was the venue. Since this year marked the tenth anniversary, the exhibition space was expanded by 30% and covered three floors in total. There were noticeably more companies, and they were more international. Most strikingly, there were many more dedicated booths from China and South Korea, all on the ground floor—which I suppose counts as a prime location, since most people spend the longest time browsing there. For example, by the time I reached the third floor at the end, I only gave it a quick once-over before leaving. I do not know whether this was arranged by the organizers or because they paid the most for their booths.

Here are a few things that particularly impressed me:

No matter when you went, the Unitree Robotics booth was always surrounded by people. I imagine the robot dance at the Spring Festival Gala amazed everyone. So yesterday, one robot model—the premium version featured at the gala—was mainly responsible for dancing. Its moves looked a lot like hip-hop, and it could also perform a few martial arts moves. Another robot—the more affordable, cost-effective version—was mainly responsible for boxing. But yesterday, the staff only had them demonstrate their skills rather than actually fight each other.

There were also two kinds of robot dogs. One was relatively small and could imitate a real puppy by walking, rolling over, and greeting people; I had already seen this one three years ago. The other had wheels, and an exhibitor told me that factories mainly use it to conduct inspections for hazardous tasks.

Later, there was also a parade featuring robots from all the companies, which was quite fun. Volunteers held leashes while staff members inside the procession operated the robots. The different humanoid robots and robot dogs either walked or danced, while a large crowd followed alongside with cameras, filming videos.

I tried an AI translator made by a Chinese company. There was almost no delay, and it supported 140 languages. Best of all was its split-screen feature: each of the two people in a conversation could view a screen oriented toward them, which would be extremely helpful for communicating while traveling. They said they were currently discussing a partnership with Teams. Only one party would need to pay, and anyone in a meeting with the paying user would see subtitles in the target language on their computer, making international company meetings much more efficient.

As I tried it, I could not help marveling that we are only in the third year of the large language model era, yet the technology has already reached this level. In the near future, we may truly no longer need to suffer through the pain of learning a new language—or at least, we will no longer need to learn one for practical purposes such as work. Instead, people will be more likely to learn a language because they are interested in the culture it carries.

I came away with far too many thoughts yesterday to fit into a single post. I will publish the next part in a few days.