[Book Review] The History of the AI Era Through Semiconductors

3 min read
TL;DR
  • Despite the title, this is not just a book about semiconductor processes or chips.
  • It explains the last 20 to 30 years of IT history through the choices made by major companies.
  • It connects compute chips, memory, packaging, foundry equipment, data centers, and cloud infrastructure into one flow.
  • It is especially useful if you follow AI news but want a stronger foundation in the physical infrastructure behind it.
Cover
"This review was written as part of the Hanbit Media Reviewer Program, where I received a complimentary copy of the book."

TL;DR

Judging by the title, I expected a book focused only on semiconductors. In reality, it was a company-centered history of the IT industry over the last 20 to 30 years. With AI news pouring in every day, this feels like a must-read for building a stronger foundation.

The First Impression the Title Gave Me

When I first picked up the book, I was slightly worried that it would be packed only with semiconductor processes and chip details. But once I started reading, I realized it uses the current AI wave as the center of the story. It is a fairly substantial book, and it uses that length to cover the broad history of IT over roughly the last two or three decades.

IT History From a Hardware Perspective

As a software developer, some parts I thought I already knew felt new again. Cloud history was the clearest example. I had mostly seen that story from a software perspective, but revisiting it from a hardware perspective gave the same history a completely different texture. It was fun to trace what kind of physical infrastructure made those changes possible.

Technology evolution from computers to mobile devices and AI

A Story Told Through Companies

The biggest appeal of this book is that it explains history through companies. As you follow why TSMC is now in such a strong position, and what choices Samsung, Intel, and SK hynix made along the way, the broader industry picture naturally comes into view. Google, Microsoft, Apple, and Palantir also appear, and the book keeps showing how each company built and improved products and services in response to changing conditions. That made the book easy to read until the end.

How AI Physically Runs

The book also explains what kind of hardware AI actually runs on, starting from the history of data centers. Compute chips, memory, packaging, foundry equipment, data centers, and cloud infrastructure all connect into one flow.

Comparison of CPU, GPU, NPU, and ASIC roles and characteristics

The parts about ASML and lithography were especially fresh because they are areas I do not usually encounter. The relationships among fabless companies, foundries, and cloud companies also become much easier to understand.

Diagram explaining the structure of wafers and dies

From Computers to AI

Computers, the internet, mobile, and then AI. The book walks through how each era led into the next. The parts I already knew worked as a useful review, and the parts I did not know gave me a lot of new context.

Closing Thoughts

This feels like exactly the kind of book we need now, when new AI news appears every day. After reading it, I felt my foundation was stronger, and that made newer AI news much more nutritious to absorb. Most of all, the narrative thread of "company choices" kept the book engaging all the way through.

One small thing I noticed while turning the last pages: next to the author name MrTrigger, I saw the publisher name Baekjoon Lim. There you are again, GOAT.