Creating humanity's intellectual peer.
Millae is an independent research team in Seoul, South Korea. We are scientists working towards autonomous, self-directed intelligence.
Millae is an independent research team in Seoul, South Korea. We are scientists working towards autonomous, self-directed intelligence.
Most companies and labs split their attention between research and product, or crowd into the same handful of trendy approaches. We do neither. Millae is a research effort focused purely on fundamental questions — deliberately working the underinvested, unfashionable corners of AI research, where we believe the highest-leverage progress is still waiting to be found.
We build on what we learned in the past from researching reinforcement learning and LLMs. We also draw wisdom and inspiration from our past work experience in software engineering, non-profit leadership experience, and therapy, as well as Buddhism and Taoism.
We're currently selecting our next research topic. We're exploring:
Good research takes time, and we're not interested in a publication treadmill to prove otherwise. We work at the pace the problem actually requires, and we'd rather walk in the right direction than risk sprinting in the wrong direction.
We're philosophically aligned with Yoshua Bengio's call to rethink the pace of ML publication and the Slow Science manifesto. A tree that grows quickly may catch the eye, but it can be easily shaken when the wind comes. The more we rush to achieve, the more easily life is consumed; a life patiently built endures.
We believe the field goes further in the long run through open collaboration, rather than siloed proprietary efforts. We see ourselves as just one team contributing to the broader collective effort to reach general intelligence. If your lab is interested in collaboration, please email us.
"If you want to go fast, go alone. If you want to go far, go together."
We're looking for:
Seoul, South Korea
If this sounds like you, reach out directly. Tell us about your research and why you want to work with us.
Email UsWe share updates on our research rarely — only when it matters.