This article is a machine translation from Japanese. It may contain translation errors.

PredNext Inc. Has Been Established

We are pleased to announce the establishment of PredNext Inc. (Representative Director: Takayuki Tokunaga).

Message from the Representative Director

I am delighted to announce the establishment of PredNext.

The company name “PredNext” is inspired by a book called “On Intelligence” written by Jeff Hawkins. When I first picked up this book, we were in the midst of the neural network winter. Neural networks were considered obsolete technology, and there was an atmosphere of “long live SVMs” and “long live convex optimization.” At the same time, there were faint concerns about the future—whether there might be limits to what could be achieved with just convex optimization and feature engineering. It was purely by chance that I picked up this book, but it became a catalyst for my interest in neural networks as next-generation technology. However, neural networks at that time were not particularly practical, and after looking at RBMs and Deep Belief Networks, I gave up, thinking they were incomprehensible and didn’t seem practical.

In “On Intelligence,” several hypotheses about intelligence were proposed, one of which was that the brain is constantly making predictions and learning based on the differences from those predictions. When I read this hypothesis, it was like scales falling from my eyes. Looking back, unsupervised pre-training through next token prediction is one of the key technologies enabling today’s large language models, and this is precisely the hypothesis described in “On Intelligence.” I was once again impressed by Jeff Hawkins’ foresight.

When thinking about the company name, I considered various keywords. Reflecting on my career, one of the books that had the greatest impact on me was “On Intelligence.” Once I realized that, it didn’t take long to arrive at this company name.

PredNext is a company focused on AI. While AI is a powerful tool, it is not omnipotent at this point. Appropriate problem formulation, data quality, implementation methods, and above all, the understanding and cooperation of stakeholders are key to success. We at PredNext understand these complex factors and are committed to supporting AI utilization that brings true value to our clients’ businesses.