Recursive releases Flint, a laptop-based coding model built by AI
Recursive, an AI startup led by Richard Socher, released Flint, a free coding model designed to run on laptops. Developed in under four months using the company's automated Eureka Machine system, Flint represents an early test of recursive self-improvement. The model allows offline, private coding without cloud fees. Backed by $650 million in funding, the release highlights the growing debate over AI building better AI and its potential for autonomous development.
Key points
- Richard Socher, CEO of Recursive and former Salesforce chief scientist, announced the release of Flint, the company's first AI model.
- Flint is a free coding model optimized for laptop hardware, enabling offline use and code privacy without cloud processing fees.
- The model was developed in under four months by a small team, utilizing Recursive's automated research system, the Eureka Machine, for nearly every development step.
- Recursive emerged from stealth in May with $650 million in funding at a $4.65 billion valuation, backed by investors including GV, Nvidia, and AMD Ventures.
- The release serves as a practical test of recursive self-improvement, a concept where AI builds better AI, which remains a debated topic among researchers regarding human control.
- Flint is available for download on Hugging Face under the Apache 2.0 license, allowing developers to integrate the model freely.
Recursive, an artificial intelligence startup founded last year, has released its first model, Flint, marking a significant step in its mission to build AI that can improve itself. The company, led by CEO Richard Socher, formerly Salesforce’s chief scientist, developed the model in under four months using its proprietary automated research system, the Eureka Machine. This system assisted in nearly every stage of Flint’s creation, allowing a small team to build the model from scratch.
Flint is a coding model designed to run on standard laptop hardware rather than cloud servers. This architecture allows developers to use the model offline, keeping their code private and avoiding the usage fees associated with cloud-based AI services. The model is available for free on Hugging Face under the Apache 2.0 license. The release represents an early public test of recursive self-improvement, the concept that AI can build better AI, which in turn builds even better AI. While the idea is widely debated in the industry, with some researchers concerned about losing human control, Recursive’s current implementation remains a human-supervised research assistant.
The startup emerged from stealth in May with $650 million in funding at a $4.65 billion valuation. Backers include GV, Nvidia, and AMD Ventures. The release of Flint highlights the increasing capability of AI systems to contribute to their own development, potentially accelerating the pace of innovation in the sector while raising questions about the long-term implications of autonomous model building.
Sources
The WireByte editorial team synthesises technology news from multiple primary sources, verifies the facts, and links every source. Articles are produced with AI assistance and reviewed under our editorial policy.
