These Strange New Minds
How AI Learned to Talk and What It Means
-
- $16.99
Publisher Description
‘Engaging, insightful, panoramic’ Mustafa Suleyman, CEO of Microsoft AI & Cofounder of DeepMind
'An eye-opening exploration of a revolution unfolding before our eyes' New York Journal of Books
Stunning advances in digital technology have introduced a new wave of human-like AI systems. Chatbots like ChatGPT, Claude, and Gemini are already reshaping economies, challenge democracies, and reshaping society in unpredictable ways. And soon, these AI systems could make autonomous decisions on their users' behalf, transforming everything we do. Understanding how they work is crucial.
Can AI systems think, know, and understand?
Could they manipulate or deceive you, and if so, what might they make you do?
Whose interests do they represent?
When will they be able to move beyond words and take action in the real world?
Neuroscientist and AI researcher Christopher Summerfield explores these questions, charting AI's evolution from early ideas in the seventeenth century to today's deep neural networks. His book is the most accessible, up-to-date, and authoritative exploration of this radical technology. With an understanding of AI's inner workings, we can address the existential question of our age: can we look forward to a technological utopia, or are we writing ourselves out of history?
‘As a leading authority...Summerfield is perfectly situated to explore the meaning and implications of these machines that are so uncannily like – and unlike – ourselves’ Brian Christian, co-author of Algorithms to Live by
'You might choose to be alarmed, excited, or indifferent to LLMs, but you should read Chris’s book before you decide’ Mike Woolridge, author of The Road to Conscious Machines
PUBLISHERS WEEKLY
This superlative study from Oxford University neuroscientist Summerfield (Natural General Intelligence) explores how large language models work and the thorny questions they raise. He explains that neural networks learn by guessing the relationships between data points and developing "weights" that prioritize the processing pathways most likely to produce correct answers. Wading into debates around whether LLMs possess knowledge or merely proffer predictions, Summerfield makes the provocative argument that human learning is essentially predictive, depending on the same trial-and-error strategy LLMs use. According to the author, this indicates human knowledge is comparable to AI knowledge. Summerfield is remarkably levelheaded in his assessment of AI's capabilities, suggesting that while obstacles to designing AI assistants that can book trips and pay bills may be resolved in the next several years, it's unlikely LLMs will ever become sentient given their inability to experience physical sensation. The lucid analysis also makes clear that technological improvements will never overcome such pitfalls as determining when to provide answers as definitive or up for debate, since such problems depend on subjective judgment. By inquiring into the nature of knowledge and consciousness, Summerfield brings some welcome nuance and clarity to discussions of LLMs. In a crowded field of AI primers, this rises to the top.
