News & Announcements

*Key Argument 1 The driving force behind AI development has been scaling up rather than new ideas. The success of the GPT family demonstrates that larger models, more data, and more powerful computational resources have driven performance improvements. *Key Argument 2 The scaling law actually held true for a certain period. As resources increased, capabilities in language understanding, coding, and reasoning improved. *Key Argument 3 There was an expectation that AGI would emerge not from a single breakthrough, but through a continuous process of scaling. *Key Argument 4 However, scaling alone reached its limits, as costs and power consumption skyrocketed while the margin of performance improvement gradually diminished. *Key Argument 5 The AI industry is now shifting its focus from the race for larger models to inference capabilities, algorithmic efficiency, and agent technology. *Implications • The future of AI lies beyond a simple competition over model size; it depends on efficient inference and practical application capabilities. • Future competitiveness is likely to be determined not only by the scale of data centers but also by algorithmic innovation and application services. • Rather than saying the era of scaling has ended, it can be viewed as evolving into a new phase.