Recursive has unveiled Flint, an AI model with open weights designed for coding tasks. The project utilizes the Eureka Machine system to implement recursive self-improvement, allowing the AI to automate most stages of its own development.
According to CEO Richard Socher, the Eureka Machine independently developed new training methods, neural network architectures, and hardware-optimized code. Benchmarks indicate that Flint outperforms similar models in programming tasks.
The company has secured $650 million in funding, reaching a valuation of $4.65 billion. Investors include Google Ventures, Greycroft, Nvidia, and AMD Ventures. The model contains 40 billion parameters and was tested on a Mac with 48 GB of RAM.
While some experts express concerns regarding the safety of recursive self-improvement, leading to proposed legislative restrictions in the US Congress, Recursive maintains that the process remains under human supervision and control.
