CEO of Silo AI, Peter Sarlin, believes open-source models like Poro represent the future of AI, offering a transparent and ethical alternative to closed models from major tech companies. The company plans to release regular Poro checkpoints throughout the training process, with the ultimate goal of creating a family of open-source models covering all European languages. The project is a collaboration between Silo AI and the University of Turku, combining applied AI expertise with academic research in multilingual language modeling.
Key takeaways:
- Silo AI, a Finnish artificial intelligence startup, has unveiled Poro, an open source large language model (LLM) designed to advance multilingual AI capabilities for European languages.
- Poro is being trained on Europe’s fastest supercomputer, LUMI, and is designed to address the challenge of training performant natural language models for lower-resourced European languages.
- Poro represents a collaboration between Silo AI and the University of Turku in Finland, combining applied AI expertise and computational resources with academic leadership in multilingual language modeling research.
- The release of Poro suggests a new era of open collaboration and transparency in the field of natural language processing, potentially democratizing access to performant multilingual models and providing a European alternative to systems from US tech companies.