This episode delves into the ongoing debate of the competitiveness between open-source and closed-source models and the reasons behind Meta's decision to publish Llama2 with a permissive open-source license
We cover:
- How much bigger can closed-sourced models be, compared to open-source?
- Are new competitor foundation models doomed, if Meta enters the game?
- What does “second place” look like, in open source?
- Can open-source datasets compete?
- Will very large open-source models soon become illegal?
