By the year 2025, Generative Artificial Intelligence (GenAI) is expected to reach a critical turning point, where we will have a clearer view of its true potential. As analysis of its applications across various industries deepens, GenAI, once shrouded in noise, will gradually reveal its true nature.
Recently, Hieb suggested that companies investing in GenAI do not necessarily need to opt for expensive and complex solutions. In fact, an increasing number of businesses are finding that Large Language Models (LLMs) are not always suitable for specific objectives.
Image Source: This image was generated by AI, provided by the image licensing service Midjourney.
In contrast, Small Language Models (SLMs), due to their fewer parameters, are typically more economical and efficient, better suited to solving specific problems, and easier to align with rapidly changing global AI regulatory policies. He noted, "Small Language Models have lower training and operational costs, are easier to modify, and are well-suited for compliance needs and many other target applications." He further emphasized that while Large Language Models are widely used in complex decision-making systems, Small Models are often more appropriate for certain specific use cases.
Although Small Language Models have been in use for over a decade, they have received relatively little discussion. Hieb believes that these models are more resource-efficient and suitable for training and deployment. Past technological trends, such as the rise of the Software as a Service (SaaS) business model, might offer some insights, suggesting that the market may tilt towards Small Models in the future. This trend is due to the advantages of Small Models in efficiency, cost, and ease of implementation. Hieb anticipates that as companies focus more on efficiency and cost, there will be increased application of Small Models in areas such as education, basic code development, recommendation systems, and chatbots.
Hieb also mentioned that some companies like Google, Grammarly, HubSpot, Zoom, and Zendesk are already using Small Language Models in various applications, including smart suggestions in Gmail, customer relationship management automation tools, and chatbots. He believes that with the increasing number of applications, Small Language Models will demonstrate greater potential.
Key Points:
🌟 2025 is seen as a significant turning point for Generative Artificial Intelligence, with its potential being showcased through applications across multiple industries.
💡 Small Language Models, due to their cost-effectiveness and flexibility, are increasingly becoming the preferred choice for companies in specific target applications.
📈 Many well-known companies are already using Small Language Models, with expectations of broader applications and developments in the future.