Microsoft's ambitions in the field of general artificial intelligence (AGI) are becoming increasingly explicit. Recently, Microsoft successfully recruited a top research team from the renowned Allen Institute for AI (AI2) and the University of Washington. This team, led by former AI2 CEO Ali Farhadi, will officially join Microsoft's newly established "Superintelligence" department.

Top Minds Gathered: Star Open Source Researchers Move

The list of joining personnel is nothing short of an "all-star" lineup in the AI world:

  • Core Leader: Ali Farhadi, not only a computer vision expert, but also led AI2 to release over 100 models within a year.

  • Open Source Pioneer: Hannah Hajjiri, co-leader of the well-known open-source large language model project OLMo.

  • Multi-modal Expert: Rangachari Krishna, who led the development of key multi-modal models such as Molmo.

These scholars' arrival not only strengthens Microsoft's shortcomings in foundational model development, but also brings AI2's long-standing efficient training experience of "achieving more with less" into Microsoft.

Behind this move lies a deep strategic consideration. Although Microsoft is deeply connected with OpenAI, as their competitive relationship has become public, Microsoft is urgently seeking to strengthen its self-developed capabilities.

By forming the "Superintelligence" team, Microsoft aims to build its own "humanistic superintelligence" system that is safer and more controllable while maintaining the partnership. In the competition for computing power with Google and Amazon, this team with strong open source background will become a crucial factor for Microsoft to achieve "computing power freedom."