From Memories to AGI

Memories of Hardship and Personal Tragedy: a Way to Build Strong AI?

Discussions around AI and strong AI are intensifying after the recent undeniable successes of large language models and diffusion algorithms. The questions often focus on whether the current state-of-the-art generative algorithms can be "truly" creative. People who focus on the algorithms might conclude that those systems only combine a large amount of training data through complex parametric transformations. 

When reading Bold Ventures by Charlotte Van den Broeck, I realized there might be a way for the current AI approach to reach the next level. What if we could add unique memories to AI models? Van den Broeck describes the lives and hardships experienced by successful architects of different periods. While extraordinary, their biographies are remarkably similar in a few critical areas. Creative people and artists in various professions often have complicated lives. They often endured personal tragedies. We all know examples of famous musicians or painters who had unhappy and tragic lives despite their enormous successes. Perhaps adding those memories as a synthetic data set to generative algorithms will give the artificial neural networks a spark of individuality and uniqueness. Many writers explored those ideas. Philip K. Dick, Arthur C. Clark, and many others wrote at length about artificial intelligence and the quest to reach AGI.