Book Review: These Strange New Minds: How AI Learned to Talk and What It Means

My journey of evaluating AI-related books continues. I am fascinated by how AI has evolved from a historical perspective. I have always been keen on understanding how combining two or more scientific domains can take new leaps forward. It is not always feasible to focus purely on your own narrow domain, but rather to look at other domains to bring clarity. I will never forget the chills I got when, during my doctoral studies, I saw resemblances between certain philosophical concepts and object-oriented programming. It might sound strange to you, but I really did, and I even showed these to my philosophy professor at the time. To become a doctor in business, one had to study philosophy as part of the doctoral journey.

These Strange New Minds: How AI Learned to Talk and What It Means by Christopher Summerfield is a lucid, accessible, and authoritative exploration of the rise of large language models (LLMs) and their profound implications for society. Summerfield, a neuroscientist at Oxford and a former DeepMind researcher, traces the intellectual and technological journey from early conceptions of thinking machines to today’s advanced AI chatbots like ChatGPT, Claude, and Bard. Blending history, philosophy, neuroscience, and computer science, he investigates whether these systems can truly "think," what their emergence means for human knowledge, and how their growing influence might reshape our future. The book includes interesting perspectives on the following aspects:

1. History: Summerfield traces the development of AI from its early conceptual origins to the present, showing how ideas about thinking machines have evolved over time. This historical perspective helps contextualize current breakthroughs and debates in AI.
2. Philosophy: He engages with deep philosophical questions about the nature of mind, intelligence, and consciousness. For example, he examines whether it’s valid to say AI systems “think” or possess a “mind,” and explores the human tendency to anthropomorphize machines—attributing mental states like knowing or believing to entities that lack subjective experience.
3. Neuroscience: Summerfield draws parallels between how the human brain processes information and how AI models (especially neural networks) operate. Both fields investigate learning, memory, and reasoning, and the comparison helps clarify both the similarities and fundamental differences between biological and artificial cognition.
4. Computer Science: The technical underpinnings of AI—such as the development of neural networks, machine learning, and advanced reasoning models—are explored to explain how machines acquire language, solve problems, and make decisions. This includes a look at recent advances that allow AI to perform complex tasks and even exhibit forms of reasoning that can sometimes rival human abilities.

The book brings several interesting findings of the AI historical evolution, and they are as follows:

1. LLMs Mark a Watershed in Human History

• The ability of machines to communicate fluently has shifted the landscape of knowledge, cooperation, and conceptualization, breaking humanity’s monopoly on language and reasoning.

2. Human and Machine Intelligence Share Surprising Parallels

• Summerfield provocatively argues that both humans and LLMs learn through predictive, trial-and-error processes, blurring the line between human and artificial knowledge while acknowledging crucial differences, such as the absence of physical sensation and motivation in AI.

3. The Transformer Revolution

• The book highlights the 2017 invention of the transformer model as a pivotal breakthrough, enabling large language models (LLMs) to achieve unprecedented language understanding and generation, which underpins their current capabilities.

4. Risks, Biases, and Hallucinations

• Summerfield is candid about the dangers posed by LLMs, from generating convincing misinformation to perpetuating programmer biases. He notes that while AI can organize and share information, its tendency to "hallucinate" (produce false or misleading content) mirrors, but also differs fundamentally from, human memory errors.

5. The Need for Coordinated Oversight and Research

• The book calls for urgent, coordinated research and regulation to address the societal impacts of AI, warning that technological progress is outpacing our understanding and control of these systems.

Target Audience

• General Readers: The book is written for a broad audience, making complex ideas accessible without oversimplifying them. It’s suitable for anyone curious about AI’s impact on society, even those with little technical background.
• Tech Enthusiasts and Professionals: Readers with an interest in AI, neuroscience, or technology will appreciate the depth and clarity of the technical and philosophical discussions.
• Academics and Students: Summerfield’s balanced approach and integration of history, philosophy, and science make it a valuable resource for students and researchers across disciplines.
• Policy Makers and Business Leaders: Those involved in shaping AI policy or deploying AI in business will find the book’s insights into risks, ethics, and future directions particularly relevant.

Conclusion: Why Read This Book?

These Strange New Minds stands out for its clarity, wit, and balanced perspective at a time when AI discourse is often polarized. Summerfield demystifies how LLMs work, explains their philosophical and practical implications, and urges readers to think critically about what it means to share the world with “strange new minds.” Whether you are excited, alarmed, or simply curious about AI, this book offers the tools and context to understand the most radical technology of our era—and to engage thoughtfully with the existential questions it raises.

PS. If you would like to get my business model in the AI Era newsletters to your inbox on a weekly or bi-weekly basis, you can subscribe to them here on LinkedIn https://www.linkedin.com/newsletters/business-models-in-the-ai-era-7165724425013673985/

My book review LinkedIn Newsletter can be subscribed to at https://www.linkedin.com/build-relation/newsletter-follow?entityUrn=7154173997436309505