The Revolutionary Impact of Large Language Models and AGI on Technology and Society

Recent remarks by Bill Gates and industry leaders highlight AI as the most transformative technology today, surpassing hype around Web3 and the metaverse.

The author reflects on personal experiences with ChatGPT, noting its unprecedented language and code understanding capabilities, and introduces the concept of "emergence" in large language models (LLMs).

LLMs, trained on massive multilingual data via Transformer architectures, exhibit abilities that rival specialized systems, such as translation, and demonstrate a form of reasoning akin to human learning.

The article argues that achieving artificial general intelligence (AGI) will require integrating AI as a universal interface (Natural Language Interface, NLI) across all software, forcing a redesign of both UI and API layers.

Three historical revolutions—graphical user interfaces, Web 2.0, and mobile internet—restructured software; the author contends AGI will be the next, compelling every system to adapt or be rebuilt.

Practical examples illustrate how a voice assistant could broadcast a single health request to multiple apps, showcasing the potential of NLI to enhance interoperability.

From a business perspective, AI-generated content (AIGC) and large models open new commercial avenues, but the distinction between using AI as a suggestion tool versus a decision-maker remains critical due to current reliability concerns.

Successful AI applications are identified as high‑frequency, low‑cost scenarios such as customer service, AI tutoring, and low‑code development, while emotionally nuanced tasks like psychotherapy are less viable.

The piece outlines the four essential pillars for building large models—algorithm, data, engineering expertise, and compute resources—emphasizing that engineering tricks often provide the competitive edge.

It predicts that while major tech giants will dominate general‑purpose models, opportunities exist for smaller firms to develop vertical, domain‑specific models using proprietary data.

Emerging job roles such as Prompt Engineer, UI/UX designer focused on NLI, and AI‑augmented software engineer are discussed, highlighting the shifting skill demands in an AI‑driven era.

Finally, the author speculates on long‑term societal impacts, suggesting that as AI reduces the need for manual labor, human value will shift toward emotional and creative contributions, while a minority of experts continue to advance AI itself.