Will the creation of AI models be standardized, outsourced, and specialized?

Widespread adoption of Large Language Models (LLMs) in business faces initial hurdles due to high costs. Research is focused on creating smaller, specialized models. While some major players may build comprehensive models, the trend is towards vendors hyperfocusing on specific industries. The AI landscape anticipates a mix of foundational models and custom model makers, forming a diverse market. Unlike typical software trends, smaller AI companies are expected to fine-tune models for specific needs, creating an AI supply chain with both general and specialized models.

It will take time for wide adoption of LLMs in the corporate world. The “large” in “large language models” makes them expensive to train and run.

- Hundreds of companies seek to modify and adapt a few large, expensive Large Language Models (LLMs).
- The market for foundational models is expected to remain robust without significant consolidation for at least five years.
- Christian Kleinerman notes the rapid development of AI and uncertainties about adoption make it challenging to predict beyond the short term.
- There is an anticipation that the market for foundational models may eventually consolidate in the future.