Two Gen Z tech entrepreneurs refused a big pay package from xAI CEO Elon Musk, but this rejection had an entirely unorthodox reason behind it. William Chen and Guan Wang, founders of Sapient Intelligence, have been working on developing a unique kind of large-language model (LL) that challenges the ongoing trends of AI technology established by the likes of OpenAI and Anthropic.
Sapient Intelligence came into being just this year. It stemmed from the idea of building an AI model that would behave and respond to problems like a human being. This was in the context of knowledge generation and implementation, which would simultaneously learn from its mistakes and make improvements accordingly. The only other known company that uses such a kind of technology is DeepSeek, which has already rattled the AI tech industry once.
Sapient Inc. Founders William Chen and Guan Wang reject Elon Musk’s offer
During his appearance at the Fortune Brainstorm AI event in Singapore in July this year, co-founder William Chen spoke about the AI technology they building at Sapient Intelligence. Throwing light on their product HRM (Hierarchical Reasoning Product), Chen shared the attributes of this open sourced software. At the event, he said, “I want to make an algorithm that solves every single problem in the world.”
In Chen’s opinion, the LLM would “do different things, tackle different tasks, has the adaptability…has continuous learning ability.” Most importantly, he highlighted that this model would also be able to “autonomously improve on itself.”
In a report by Fortune dated November 28, the founders shared their experience of being approached by xAI CEO Elon Musk to be a part of his team. The report stated, “Musk sent an email through what, at the time, was his new company, xAI, which wanted to recruit the students in a multi-million dollar pay package, Chen says. It was the kind of offer young founders dreamed of.” However, the duo chose to refuse the offer.
William Chen spoke to Fortune about the limitations of LLMs that are yet to be learnt about. He said, “We decided that large-language models have their limitations.” Talking about his company, Chen said, “We want a new architecture that will overcome the structural limitation of [large-scale machine learning].”
Related: Cathie Wood Hails Tesla CEO Elon Musk as “Thomas Edison of our age”
Sapient Inc. prioritizes good data over more data
One of the biggest factors that has helped gain Sapient Intelligence significant clout in the AI tech industry is their commitment towards research. They have been working on their LLMs and publishing their findings in academic journals, thus sharing any progress or challenges they are facing in their work. At the same time, the team at Sapient Intelligence have also prioritized the utility of “good data” over “more data.”
The founders’ initial project, titled OpenChat, was trained on a specific amount of data that consisted of high-quality interactions. With the help of Reinforcement Learning or RL, Chen and Wang developed OpenChat to learn to improve itself by making decisions, understanding feedback, responding to it by modifying itself.
Launching it as an open-sourced project, OpenChat became a big sensation in the AI tech industry. As per the report, researchers from Stanford University and UC Berkeley “pulled the code, built on top of it, and began citing the work.“
Given the amount of sheer potential Sapient Intelligence is showing already, it is only a matter of time before Chen and Wang become of two of the most prominent names in the industry.
Also Read: Elon Musk Believes Programming Morality Into AI Can Be Detrimental
