Epoch AI prediction: Within the next year, the pace of inference model development will slow down.

The latest report from the non-profit AI research organization Epoch AI pointed out that AI companies are struggling to continuously extract significant performance benefits from inference models, and that progress in inference models will slow down in as quickly as a year. The report, based on public data and assumptions, emphasizes the limitations of computing resources and the increase in research costs. The AI industry has long relied on these models to improve benchmark performance, but this reliance is now facing challenges. Analyst Josh You of the organization pointed out that the rise of inference models is due to their excellent performance in specific tasks. For example, OpenAI's o3 model has recently significantly improved mathematics and programming skills. However, these inference models solve problems by increasing computing resources to improve performance, but as a trade-off, these inference models require more computation to handle complex tasks, thus taking longer than traditional models.