Farang Secures €1.5M to Transform AI with Brain-like Models
Farang, an AI research lab based in Stockholm, has successfully secured €1.5 million in seed funding to advance its innovative foundational Large Language Models. The company aims to transform how AI mimics human cognition, promising to generate intuitive responses while significantly cutting down on computational costs. Their unique architecture differs from the traditional transformer models currently employed by leading AI platforms like ChatGPT and Claude.
Rather than predicting responses one word at a time, Farang’s model comprehends complete answers first, akin to sketching an entire picture before filling it in with detail. This method allows for internal reasoning through non-textual processes, which results in more coherent outputs and a drastic reduction in required computational power. Emil Romanus, Farang’s Founder, emphasizes that they aren’t just building another application on top of existing technology. Instead, they’ve created a new foundational architecture focused on developing specialized AI assistants that excel in specific fields like programming and medicine.
Farang sets its sights on the limitations of today’s AI tools, targeting specialized applications where existing solutions fall short. The company’s initial focus is on optimizing React programming, aiming to generate more efficient and iterative code than today’s large language models can offer. Romanus pointed out the advantages of their architecture, which allows them to create highly specialized models for niche applications that would typically be too costly with conventional methods. For instance, they can analyze unstructured medical data or develop programming tools that truly grasp specific frameworks.
Another key feature of Farang’s technology is enabling organizations to deploy these tailored AI models on their own servers, ensuring full control over sensitive data. This is particularly important for sectors like healthcare, finance, and law, where confidentiality is paramount. Instead of relying on external AI services, firms can train and operate Farang’s models in-house, promoting data sovereignty. Romanus envisions law firms and medical institutions using AI tailored to their unique practice areas and patient histories, all while maintaining privacy.
While Farang is starting with specialized applications to validate its technology, the long-term goal is to establish itself as a leader in artificial intelligence, eventually competing with giants like OpenAI in broader AI capabilities. Romanus believes that by showcasing their architecture’s success in niche markets, they are laying the groundwork to challenge existing leaders in all areas of AI.
The funding round, led by Voima Ventures and the Amadeus APEX Technology Fund, alongside contributions from notable angel investors, will help Farang scale its models and invest in the computing power needed for specialized applications. The firm’s approach represents a notable shift in the AI industry, with potential implications for the future of technology sectors across Europe.
“Content generated using AI”
We create intelligent software and AI-driven solutions to automate workflows, modernize legacy systems, and sharpen your competitive edge.