The late Doug Lenat, who founded the company Cyc, dreamed of advancing artificial intelligence, and particularly symbolic artificial intelligence, to the point where the system could make novel discoveries, for example in the area of mathematics. The goal is that the system can propose theorems and not just prove them. Advances were made and continue to be made by Cyc and others. In the meantime, AI has focused on statistical connectionist networks that work through methods of large language models. Recently, the Shanghai-based NovelSeek company, renamed GeniX, announced a unified closed-loop multi-agent system that assists researchers to develop discoveries in 12 different scientific research areas. While a human still is in the loop to guide the discovery process (and we might speculate that human intuition might always be required), the system is said to inspire innovative research through expert insights, combining large language models and domain knowledge with multiple agents orchestrating the data flow and task scheduling. Between the DeepSeek announcements in early 2025, and now the NovelSeek announcement, there is concern that China is excelling in AI technology development. Policy-makers in the US are aware that US AI researchers need up the game. The US has a lead, but it is not as large as many believe. Going forward, the job of scientists may not be replaced, the process of discovery might become more efficient with multi-agent technology.