Sakana AI's approach is unique in the AI field, which has been focused on scaling up models. The startup argues that smaller, specialized models can collaborate to match the capabilities of larger models, likening it to how groups of people with distinct skills outperform lone polymaths at complex jobs. The funding will be used to further develop its nature-inspired AI techniques and expand its team.
Key takeaways:
- Sakana AI, a Tokyo-based artificial intelligence startup co-founded by two ex-Google engineers, has raised $30 million in seed funding from investors including Lux Capital and Khosla Ventures.
- The company aims to develop smaller, more efficient AI models inspired by nature, such as the collective behaviors of schools of fish and flocks of birds.
- Sakana's approach differs from the current trend in AI, which focuses on scaling up models to beat benchmarks, a tactic that has led to backlash over the computing resources and environmental impact required.
- The funding and partnerships with major companies like NTT will allow Sakana to further develop its nature-inspired AI techniques and potentially pioneer a new AI paradigm from Asia.