In 2017, Google researchers introduced the transformer, laying the groundwork for AI chatbots

In 2017, Google released a scientific paper, which spoke of an algorithm that led to the development of what is now called the transformer architecture. More than nine years later, this architectural change has significantly influenced how machines generate and process language. It became a key step in bridging the gap between hard-coded computer programs and modern AI chatbots.

Why Attention Outperformed Earlier Approaches
The algorithms for processing languages before this discovery required step-by-step processing of the information.