Open source models currently account for only 20 to 30% of inference tokens used, indicating that they are not yet winning against closed-source models.

This year has been positive for open-source AI, with many innovative models emerging, and there's a prediction that open-source will surpass closed-source models in usage.

The future of AI relies on open-source models to ensure freedom and prevent centralization of power in AI development.

I believe that as AI models become more efficient and less expensive to train, there will be more reasons for companies to open source their technology.

Open source AI development is progressing faster than proprietary models, as it allows for a diverse range of ideas and contributions from around the world.

The transparency and high leverage of open-source models will lead future generations to prefer creating open-source solutions over closed-source ones, as they grow up learning about state-of-the-art machine learning.

The future will likely see more choices of AI models rather than a single dominant model.

Open source models are essential for ensuring that AI achieves its full economic potential, as they allow for broader access and innovation in AI technologies.

As AI technology evolves, the availability of powerful models may increase, potentially making them accessible for free in the future.