Genius Makers by Cade Metz: The People Behind Deep Learning
An in-progress read of Cade Metz's Genius Makers, from 2021. Hinton through the AI winter, the talent war at Google and Facebook, DeepMind's AGI goal. It ends before ChatGPT, and I read it in doses.
The Book You Read in Doses Because It's Too Rich to Binge
The Book I Haven't Finished But Keep Thinking About
Confession time: I started Genius Makers months ago, and I'm still not done. I am enjoying it, and it is excellent. I keep putting it down and picking it back up because it's dense in a specific way. Each chapter has enough ideas and "holy shit" moments that I need time to digest before continuing.
This is the story of the people building AI, more than the technology. The rivalries between labs, Google Brain, DeepMind, OpenAI. The personalities driving it, Geoffrey Hinton, Demis Hassabis, Sam Altman, Ilya Sutskever. The race toward Artificial General Intelligence (AGI) and what that might mean for humanity.
Reading this in 2026, after living through the ChatGPT explosion, the agent mode revolution, and watching AI infiltrate every industry, hits different than it would have a few years ago. The book chronicles events that seemed like academic research or corporate competition at the time but now feel like the opening chapters of a much bigger story we are living through.
I am reviewing this as an in-progress read rather than pretending I finished it. It keeps pulling me back, and it is heavy enough that I need the breaks.
Why it is worth reading, even slowly
The people. You are using AI systems built by the people in this book. It helps to know what drove them and who they were up against.
The competition. Google vs. Facebook vs. OpenAI vs. DeepMind. That race shaped what we have now and what is coming.
The record. It documents the shift as it was happening. Reading it now is like reading about the early internet in the late 90s.
AGI. The quest for Artificial General Intelligence is not science fiction anymore. These people are actually trying to build it, and this book shows how and why.
Personal stakes. I work with these systems every day, and the history tells me where a lot of what I use came from, and maybe what comes next.
If you work with AI, or wonder how we got here, this book is the context. (Buy on Amazon)
Book Details at a Glance
| Feature | Details |
|---|---|
| Title | Genius Makers: The Mavericks Who Brought AI to Google, Facebook, and the World |
| Author | Cade Metz |
| Publication Year | 2021 |
| Genre | Technology, Biography, Business, History |
| Length | ~384 pages (reads denser than page count suggests) |
| Main Themes | AI development, Corporate competition, Deep learning revolution, AGI quest, Technology ethics |
| Key Figures | Geoffrey Hinton, Demis Hassabis, Sam Altman, Ilya Sutskever, Yann LeCun, Andrew Ng, many others |
| Relevance Today | Essential context for understanding where AI stands in 2026 |
| Reading Style | Dense, information-rich, requires active engagement |
| Who Should Read? | AI practitioners, tech investors, anyone trying to understand the AI transformation |
What I've Learned So Far
I am still reading, so no spoilers for myself. This is what has stuck with me from the chapters I've finished.
- Geoffrey Hinton Through the Winter
Early chapters follow Geoffrey Hinton through his decades on neural networks when the field had written them off. Through the "AI winter", when funding dried up and the field was considered dead, Hinton kept pushing.
What got me is that this wasn't stubborn attachment to a failed idea. Hinton had theoretical reasons to believe neural networks would work at scale, even when the results weren't there yet. He needed more compute and more data.
Then AlexNet happened in 2012. Hinton's student Alex Krizhevsky used deep learning to win ImageNet, and suddenly everyone who had dismissed neural networks for decades wanted in. The rest is history we are living through.
The people building today's AI systems spent decades being told they were wrong. They were right, and proving it took decades of persistence. Knowing that changes how I read their predictions about AGI.
Reading this while working with AI agents feels like watching the origin story while living in the sequel. The neural networks Hinton pioneered power the systems I use daily.
- The Talent War, Google vs. Facebook vs. Everyone
Mid-book sections detail the corporate competition for AI researchers. When Google acquired DeepMind in 2014, for a reported $400 to $650 million, it signaled AI was transitioning from academic curiosity to corporate battleground.
The talent war got absurd, researchers commanding million-dollar salaries, companies poaching entire teams, non-compete agreements, secretive projects. Facebook built FAIR (Facebook AI Research) to compete with Google Brain. Elon Musk, Sam Altman and others started OpenAI as a nonprofit counterweight to corporate AI labs.
This corporate competition shaped what AI became. The researchers had academic ideals about open research, but corporate pressures pushed toward secrecy and competitive advantage.
Reading this in 2026, knowing how it played out, OpenAI becoming for-profit despite the name, Google releasing then retracting papers, the whole LLM race, the early moves feel bigger than they could have at the time. The seeds of where AI is now were planted in these corporate decisions.
- DeepMind and the AGI Dream
The DeepMind story is the one that got me. Demis Hassabis was aiming for Artificial General Intelligence from day one, and said so. Actual thinking machines, well past better image recognition or language processing.
The book chronicles the achievements. AlphaGo beating the world Go champion in 2016, AlphaZero learning games from scratch without human data, protein folding breakthroughs with AlphaFold. Each one felt like a step toward AGI.
The questions underneath are big ones. What does it mean for a machine to "understand"? Is AGI inevitable or impossible? Should we be racing toward it or cautious about it?
We are closer to AGI now than when this book was published in 2021. GPT-4, Claude, Gemini feel qualitatively different from earlier AI. Reading about the early steps, I wonder how close we actually are and whether the people building it know.
- The Ethics Problem, Pushed Aside
Scattered throughout the book are warnings about AI ethics, safety, and alignment that were largely ignored at the time. Researchers raising concerns about bias, misuse, or existential risk were often dismissed as fearmongers.
The book documents how ethics kept getting pushed aside for capability. "We'll solve the ethics later" was the implicit motto. Ship the product, win the race, worry about consequences afterward.
With AI bias scandals, disinformation and existential risk all over the headlines in 2026, the early dismissal of ethics feels shortsighted. The warnings were there, and they were ignored.
- The People Are as Important as the Technology
The human story is what makes this book, more than the AI story. These are brilliant, ambitious, sometimes petty, often visionary people with competing motivations and values.
Hinton's academic idealism against corporate pragmatism, Hassabis's AGI dreams against Google's product needs, Altman's open-source rhetoric against OpenAI's commercial trajectory, LeCun's scientific ideals against Facebook's business model.
The personalities explain a lot of why AI developed the way it did. People with their own values, incentives and blind spots built it.
Each chapter brings in new people and new dynamics that change what I thought I understood about how AI developed. Nothing in it reads like "progress marches forward". It is messy, political, human.
Why I Read This in Doses, and You Might Too
Reading it off and on is not a criticism. It is how this book deserves to be read.
Density. Each chapter carries several storylines, technical concepts, corporate maneuvers and ethical questions at once. It's a lot.
Context. I keep stopping to look up a paper, a person, a technology. It makes the reading richer and slower.
Reflection. The ethical and philosophical questions don't have easy answers. I need time with them before the next chapter.
My own work. I am working with AI as I read, so the history keeps connecting to what I am doing now, and I stop to trace it.
Weight. This is not light reading. The stakes are enormous either way. Reading about people racing toward AGI without clear safeguards is exciting and terrifying at the same time.
Modular. Chapters are fairly self-contained. You can read one, put the book down for weeks and pick it back up without losing the thread.
Reading History While Living the Sequel
Publishing in 2021 means the book predates:
- ChatGPT's explosion (Nov 2022)
- The LLM race (GPT-4, Claude, Gemini, etc.)
- Agent mode workflows becoming practical
- AI disruption of creative industries
- Current AI safety and alignment debates
Reading it now feels like watching a historical documentary about events leading up to the present. You know how certain decisions played out, you see the warnings that were ignored, and the inflection points that did not look like much at the time.
The distance helps the book. You are reading how we got to this exact present, and maybe getting hints about where it goes.
Partial Recommendation & Where to Buy
I can't give a full review since I haven't finished, but I can recommend it. It is essential reading if you want to understand the AI transformation we are living through.
Who should read this:
- Anyone working with AI professionally
- Tech investors trying to understand the field
- People curious about how AI actually developed (not the mythology)
- Anyone concerned about AGI and its implications
- Engineers who want to understand the human/political side of technology
Who can skip this:
- People looking for technical detail (this is about people and companies, not algorithms)
- Anyone wanting quick, light reading (this demands engagement)
- People uninterested in corporate competition and personalities
Reading strategy I recommend:
- Don't try to binge it, read a chapter, digest, think
- Look up unfamiliar people and concepts as they appear
- Connect what you read to current AI developments
- Take breaks when it gets heavy (which it will)
📖 Buy Genius Makers: The Mavericks Who Brought AI to Google, Facebook, and the World on Amazon
Related Reading
For more perspectives on AI, technology, and computing:
- CODE: The Hidden Language of Computer Hardware and Software - Understanding the foundations that make AI systems possible
- The DevOps Handbook - Systems thinking for building reliable AI infrastructure
- Antifragile: Things That Gain From Disorder - Building resilient systems in uncertain environments
- Shoe Dog by Phil Knight - The messy human reality of building a transformative company. Like the AI pioneers in Genius Makers, Knight assembled obsessive misfits, fought through corporate competition, and built something world-changing while constantly uncertain
Companion reading:
- Life 3.0 by Max Tegmark (philosophical AI future)
- The Alignment Problem by Brian Christian (AI safety deep dive)
- AI Superpowers by Kai-Fu Lee (China vs. US AI race)
I have not finished it yet.
I keep coming back to it after putting it down, and that says something. It is hard going and worth it. The people and decisions chronicled here shaped the AI systems I use daily and will shape what comes next.
This post contains affiliate links. If you purchase through these links, I may earn a small commission at no extra cost to you. Thank you for supporting this blog!