Open Sourced SOTA Verticalized LLM's (eg. Healthcare)

Fund large open collaboratives (eg. BigScience Bloom or Human Genome Project) to build SOTA and open sourced (Apache 2.0) verticalized AI. These project should be driven by individuals and organisations that build an open alliance. And while designing these initiatives, never forget that the antidote to open source is bureaucracy!

This is in line with the recommendations of the Draghi report that mentioned open source and verticalized AI approaches as potential strategies for European competitiveness in artificial intelligence:

1. The report suggests developing sector-specific AI models that could be adapted to the needs of EU industries, building on existing pre-trained open-source AI models and complementing them with proprietary industry data

2. This approach focusing on specific AI applications and vertical, sector-specific AI is seen as more feasible with smaller AI infrastructure investments, and is already happening without the need for additional public financin

3. Draghi proposes an "AI Vertical Priorities Plan" to accelerate AI development across ten strategic sectors where EU businesses would benefit most from rapid AI introduction, including automotive, manufacturing, energy, and healthcare

4. The report acknowledges that open source projects have demonstrated how decentralized innovation can compete with large proprietary firms, even in sectors thought to be governed by scale and "winner-takes-all" dynamics