Genius Makers

Genius Makers

The Mavericks Who Brought AI to Google, Facebook, and the World

by

Completed: May 13, 2021
★

I started off thinking this could be a 4 or 5 star book - I find this topic very interesting. Quickly, I realized this would be a 3 star book at best given the author’s seemingly marginal understanding of the field itself (and unclear if the author tried to understand any of the technical details either). Then it became a disappointing 2 star book. Ultimately, a couple of very silly incorrect facts became yet another pet peeve and hence the singular star. Disappointing. I really wanted to like it.

The biggest puzzle is that apparently Walter Isaacson wrote a glowing blurb quote for this book! How?!

Virtually zero technical explanations of anything involved in the field. For example, a terrible explanation of back propagation. And the author doesn’t even attempt to explain Convolutional Neural Networks or even deep learning.

Earlier in the book, the author spends a fair bit of time on the history of how various terms came into existence (e.g., neural networks), declaring them often as marketing gimmicks (e.g., deep learning). However, later in one chapter the author just casually starts referring to all this as machine learning without any context behind that term. Alludes to shoddy editing work if nothing else. There are also too many abrupt ends to stories - an example being the DeepMind project with the Royal Free NHS Trust which a British regulator subsequently deemed to be an illegal sharing of medical data. Again, shoddy editing work if nothing else.

Overall, technical explanations are all too absent from the book. But that isn’t even the biggest issue - even when a technical explanation is attempted, it wouldn’t pass muster as a simple 5th grade explainer and only serves to highlight the author’s poor understanding of the topic (and worse, lack of desire to even attempt to understand it).

Yet another example: when discussing GPUs, the author only says it’s the ‘extra processing power’ of GPUs that made them desirable over CPUs. Not even a simple nod towards parallel processing. Very disappointing.

When discussing Qi Lu (previously at Microsoft, Baidu and Y Combinator), the author writes “he had a way of delivering his ideas in sharp, self-contained, slightly strange technological axioms” with one of the axioms being “deep learning is computation on a new substrate.” I don’t understand what the author is trying to convey here. ‘Slightly strange technological axioms’ is probably what all scientists’s uttering would look like to someone on the outside. There’s no attempt to understand or explain any of this. It’s simply quoted as part of a ‘character sketch’ or something.

And then, exasperated as I was reading this book on a topic I find so interesting, I get to a page where the author talks about “Ed Boyton, a Princeton University professor”. Clearly, he means Ed Boyden, the MIT professor (who did his PhD at Stanford), known for his work in optogenetics. Again, at the very least, a very poor and rushed editing and fact-checking job by the publisher and the editors.

When discussing natural language models, the author does talk at length about Google BERT, but then only includes a passing reference to OpenAI doing some similar work creating their version of BERT. There is no explicit discussion of GPT! GPT-2 was out already in February 2019, and even GPT-3 has bene out since June 2020. No discussion of it for a book published in March 2021 other than simply mentioning in passing that OpenAI did some ‘Google BERT like work’ further shows the author’s understanding of the field.

The author’s character sketches are also bizarre. I wouldn’t claim to be any big fan of Sam Altman, but this description of Altman (stated as a matter-of-fact) is bizarre:

During his rapid ascent, he was motivated first by money, then by the power over the people and the companies in his orbit, and then by the satisfaction that came from building companies that had a real impact on the larger world.

I think ultimately what bothered me is that in totality, I felt duped since the author doesn’t appear to have spent the time and effort doing his own research on the topic he’s writing about, and worse, it doesn’t even appear like he has any desire to understand the topics. The idea appears to have been simply to cash in on a topic that the world is talking about and write a poor version of a ‘thriller’ with some poorly done character sketches.

All book cover images are from Goodreads unless specified otherwise.

Updated: