A couple of weeks ago, I spent one afternoon at this year’s edition of VivaTech in Paris, Europe’s largest startup and tech event. It wasn’t something I had planned. A friend of mine gave me a ticket at the last minute, so I arrived with little (no) preparation. Thousands of people, hundreds of companies (all sizes, niche and mainstream), in the one building I managed to visit.

I’m not a great fan of big crowds, so it didn’t take long before I was already feeling somehow overwhelmed. So, I decided to take it easy: instead of trying to come up with an optimal plan in a hurry, I decided to just wander around or, as the French would say, flâner. And so I did.

From this, I got a general sense of what was going on there:

  • A strong presence from the main French industries: both tech companies and more traditional industries that invest heavily in innovation and in adopting AI. The beauty and luxury industries, for example.
  • A strong European presence: there were many pavilions from countries all around the world, but especially from Europe. This gave me glimpse into where each country’s talent and investment are heading.
  • Global tech companies and AI. Completely expected. State-of-the-art AI, robotics, infrastructure, frontier models, and so on.

Since we are only a prompt away from knowing what the main highlights of the event were, I’m simply going to give an honorable mention to a few things that stood out to me.

Multiverse Computing

I had already heard about Multiverse Computing. I think what they do is pretty brilliant. Their technology, called CompactifAI, uses quantum-inspired tensor networks to transform the MLP and self-attention layers of transformer-like architectures. This results in a reduction in the model’s memory footprint and number of parameters while maintaining accuracy.

They had a modest stand where they were showing one of their small models running on a Raspberry Pi. The model was able to describe, the image captured by a webcam in real time. I believe they were running their 0.3B LittleLamb model, probably in a pipeline with a small video-language model.

I also had the chance to exchange a few words about these topics with Michel Kurek, CEO of Multiverse Computing in France.

colormass

I was happy to come across colormass’ stand because one of its founders is the sibling of a friend of mine. The product they displayed was a material scanner that captures surfaces with impressive detail. Beyond their appearance, fine properties such as texture and reflection are captured and can then be reproduced in rendered images, for example in digital environments.

Although I don’t understand all the underlying technology, the current and potential applications of this kind of innovation are pretty exciting.

Healthcare

Innovation in healthcare is something that my brain simply cannot ignore. In the age of AI, this translates into wearables with great capabilities, patient monitoring, and frontier models for the analysis of all sorts of data, to name a few.

One thing stood out to me, though: Physiologas Technologies’ in-home hemodialysis machine. Yes, you read that right.

If you’ve ever met someone who suffers from kidney failure, you know how revolutionary this kind of technology could be. Hemodialysis is a complicated procedure that needs to be performed frequently: I’ve met people who need it almost daily. And it is expensive. From their website: “In the United States alone, there are approximately 550,000 dialysis patients, and annual healthcare costs for dialysis exceed $23 billion.”

The ways in which this kind of technology could impact so many people are simply amazing: potentially expanding access to dialysis, reducing costs, and improving patients’ quality of life.

To be honest, the major players in cloud and platform services had very crowded stands. But I believe that, as someone who works in software infrastructure, it is perhaps just as important to be aware of the products being developed by all kinds of companies (their scale, needs, and implementations) so that we can enable them, as it is to get an overview of the technologies offered by a major provider.

By the end of the afternoon, I was pleased with my visit. I think that, after all, wandering around without a plan wasn’t such a bad strategy.