Why AI Has Only a Seven-Year History—and What AI+ Means for the Future

Wang Jian delivered a speech at the 2024 Bund Conference titled “AI, AI+ and AI Infrastructure,” sharing his perspective on the evolution and future of artificial intelligence.

Key insights

Modern AI differs fundamentally from the AI of the early 1980s.

Two main constraints prevent building something better than ChatGPT: the underlying technology (models and infrastructure) and the depth of problem understanding.

When a technology becomes infrastructure, it achieves the deepest level of societal penetration.

Data is the core component of AI infrastructure, not merely an accessory to models or compute.

AI, AI+, and AI infrastructure represent simultaneous revolutions in technology, mechanisms, and foundational platforms.

He emphasized that AI has a long conceptual past but a very short practical history, tracing its roots from Turing’s 1950s “Intelligent Machines” paper, the early term “computing machinery,” and the Dartmouth conference. He highlighted the role of “Cybermetics” and Herbert Simon’s early predictions about AI’s impact.

Wang noted that the 2017 transformer paper marked a turning point, leading to breakthroughs such as Google’s transformer models, OpenAI’s GPT series, ChatGPT, and DeepMind’s AlphaFold. He argued that these advances share a common foundation: the transformer architecture combined with large‑scale data and compute.

He clarified the meaning of “AI+,” warning against merely attaching an industry label to AI and urging a deeper understanding of the underlying mechanisms. He described ChatGPT as an application platform built on a foundational model rather than a simple product.

Discussing OpenAI’s hybrid nonprofit‑for‑profit structure, he illustrated how massive investment from cloud providers (Microsoft, AWS, Alibaba Cloud) underpins today’s AI unicorns, positioning cloud computing as the essential AI infrastructure.

Wang concluded that the scaling of data, models, and compute by orders of magnitude has created a new AI infrastructure, and that the concurrent revolutions of AI, AI+, and AI infrastructure are shaping the future of technology.

He thanked the audience.