... When Bennett begins to connect the evolution of the human brain to where we are in the development of artificial intelligence is when the book, for this reader, gets more interesting. Why can’t machines truly learn? Even ChatGPT, which every industry seems to be embracing these days, can’t “learn things sequentially,” writes Bennett. “They learn things all at once and then stop learning.” We’ve trained ChatGPT using the entire contents of the Internet, but the software can’t learn new things because of the risk that it will forget old things, or learn the wrong things. ...
... he does end his book with a challenge. Evolution gave us our magnificent human brain, he writes, and now that we are in a position to play god and create a new form of intelligence, we must first decide on our goal — are we destined to spread out across the cosmos? Or will we fail, victims of pride or climate change or something yet unseen, just another branch on the evolutionary tree, which will grow on without humans and perhaps never add a limb called “Artificial Intelligence?” ...
See the full article here: https://www.sandiegouniontribune.com/entertainment/national/story/2023-10-23/book-review-a-brief-history-of-intelligence-may-help-humans-shape-the-future-of-ai