For a book that is heavily publicized and garnered reviews in major business magazines, this book flatters only to deceive. Unless you are a total novice in this space, a reader is unlikely to find any new examples or insights from the author's treatment of algorithms. Most examples have been discussed ( in terms of technical content and impact on business models, society, behaviors) in magazines like Wired, PopSci and NYT technology pages many years ago. The dated references to recommendation engines like dating websites, those focused on music (Pandora, etc) are all superficial and provide no new insights or a critical appraisal of where those technologies are headed.
The author also overly focuses on Wall Street based scenarios to explain algorithms - he does a particularly bad job in representing algorithms as nothing more than fast calculators - that too, with a fundamentally flawed example based on option trading (I sincerely hope that the author never tried the trade he has mentioned in the book). That misguided example reflects poorly on author's understanding of algorithms and inadvertently proves one thing - algorithms are only as good as the thought that went behind its design.
Despite the superficial treatment, the author makes a few important points in the last two chapters on the need for more skill development in "STEM" disciplines and makes an argument that medical diagnostics is the next main area where algorithms are poised to expand. The discussion is very rushed and provides no meaningful action plan. Moreover, the author fails to acknowledge the vast amount of data that an individual is generating on a daily basis - and concepts of "big data" that could shape how new avenues for algorithms can evolve. Even in healthcare, the author's focus on a tiny sliver of possibilities shows a certain laziness to explore the topic more critically. The role of algorithms in personalizing treatment plans, monitoring for adherence, risk stratification etc are all well-understood frontiers in healthcare - and he chose to ignore them completely.
Overall, a very superficial (but fast paced, entertaining read) treatment of a narrow view of 'machine learning' with very few new insights or examples. An OK read for a beginner to the field.
on September 2, 2012
1. Wall Street, The First Domino - this chapter tells the story of Thomas Peterffy, who was apparently the major innovator in the last 40 years in algorithmic trading. The guy is now a billionaire. It's a VERY interesting story.
2. A Brief History of Man and Algorithms - This spends a lot of time discussing mathematicians of the past, and how their innovations led to
3. The Bot Top 40 - Talks about how algorithms can be used to detect which songs are likely to be hits. Some great stories.
4. The Secret Highways of Bots - The main idea of this chapter is that the SPEED of algorithms is what makes them so valuable. The majority of the chapter is spent telling the story of how two guys spent $200 million building a new communications line between Chicago and NYC so that they could shave 4 milliseconds off the amount of time it took to communicate between the two cities, which gave a HUGE advantage in algorithmic trading. The plan worked and the guys made a ton of money off it.
5. Gaming the System - Algorithms in gaming (poker, etc.)
6. Paging Dr. Bot - Gives examples of companies that are using computers to replace a LOT of the work now done by doctors.
7. Categorizing Humankind - Tells the story of how NASA used algorithms to detect which astronauts would work well together during the 1960s/70s missions, and how this same idea is now being used to create algorithms that can detect your personality over the phone and connect you with a customer service representative whose method of communication matches yours. Very interesting.
8. Wall Street Versus Silicon Valley - Talks about how Silicon Valley and Wall St. compete for talent
9. Wall Street's Loss is a Gain for the Rest of Us
10. The Future Belongs to the Algorithms and Their Creators
How I found out about the book: I preordered the book after I read the author's August 2012 piece in the Wall Street Journal (which was just an excerpt from the book).
What I like about it:
- It is written very clearly, and you can finish the book quickly. The author used to write for Forbes, and it definitely felt like I was reading a magazine article while I was reading the book.
- The book isn't very expensive, and so it seems worth the price to have an extended glimpse into this topic. I don't subscribe to magazines and newspapers at the moment because too many of the articles aren't of interest to me, and it takes time to dig through all the stuff I'm not interested in and find stuff I AM interested in. A book like this solves that problem.
Other books to check out if you like this one:
- The Autobiography of Henry Ford - Ford spends most of the book talking about his method of innovation in manufacturing the Model T, which is exactly the same kind of innovation we're seeing now with the use of computers.
- The Singularity Is Near by Ray Kurzweil - talks about how computers are becoming smarter and smarter, to the point where we'll all be biologically immortal
- A Field Guide to Genetic Programming - this is a great intro to a type of computing that is producing better-than-human results by "evolving" programs instead of having people make them by hand.
on October 26, 2012
Steiner's approach to this topic is excellent, taking us through the widely acknowledged but little understood algorithms of Wall Street, to the Mathematical foundations of computer programming, and then to areas more likely to impact the lives of average readers, from commercial uses and finally to algorithms' potential uses in diagnostic medicine for both body and mind.
Unfortunately, the end result is a jumble of hyperbole, gaps in reasoning, outright plugs for certain companies, outdated examples, and just plain inaccuracies. Several readers, for instance, have commented on the confounding explanation of a delta neutral trade. I'm pretty confident that even the meatiest of the "meatheads" (Steiner's term) in the pit were competent enough to lock in a conversion or reversal (something that eludes Mr. Steiner). His explanation isn't just wrong, it entirely misses the concept of delta neutral, and so simultaneously denigrates both the conventional traders and the brilliance of Mr. Peterffy's arbitrage. This sad theme is repeated throughout the book. Steiner's world is one in which a handful of shining pillars of genius wade through a sea of crusty, intransigent morons, which although possessing a kernel of truth, grossly oversimplifies and thus does no justice to the push for and against the expanded use of algorithms.
I was also dismayed that although Steiner acknowledges on a number of occasions the dangers of runaway algorithms, he entirely avoids the far more subtle ethical questions of control. "No willy-nilly tests, no gut feelings, just data in, data out" says Steiner of a rather aggressively imagined Dr. Algorithm. Unfortunately, there is no such thing as neutral data. Which data are being used? How are they being interpreted? Who is making these decisions? Writing an algorithm for wide public use gives someone, whether it be the programmer, the owner of a music distributor, a hospital, or the government, an enormous amount of power. I feel that this book is sorely incomplete without some discussion of this.
We need straightforward, accurate, easily read books on these wonderful and terrifying tools. This just isn't one of them.
on February 2, 2013
I really enjoy the aim of this book -- to explain how algorithms play such an important role in different areas of our lives. Case studies help add context to what might otherwise be an abstract mathematical musing. But I find the average Joe-oriented approach to come with unintended consequences. The writing is simply hyperbolic. It makes each incremental advancement in automation out to be the apocalypse. Options traders are using options -- well let's pack up and call it a day! Euler started mathematics from a young age -- what a genius! What a remarkable young mind!!
The author, lacking a more meaningful approach to this subject matter, decided to dramatize it as if to catch our attention. Duly noted, and poorly received.
on September 21, 2012
The book is a bunch of examples of what algorithms can do. First example is great. Every chapter after has the same thing applied to a new issue. If you read some tech articles on the internet you can see some of these examples and be wowed without having to buy the book. I was expecting some more technical approach to the subject.
on September 10, 2012
This is a gem of a book. The subject matter alone is compelling; it exists in that rare place where hugely important, life effecting forces go largely unexplained to the fray of humanity. Algorithms and the people who wield them are pulling all the strings here folks! This book details how it all happened and what's to come, and I for one appreciate being clued in.
The author's style is the best kind of journalistic prose - informative, technical when needed, and honed in on the humanity behind such a, dare I say, nerdy topic. I don't agree with the reviews knocking the lack of tech talk. This is NOT a textbook, but rather the type of discovery that's ingested by an engineer and spit out by a journalist and lucky for us Mr. Steiner is both. I can't think of many others who could succeed where he has. Read this book.
on December 9, 2013
I was disappointed with this book. The subject material is (or could have been interesting), but the book tries provides too little information...if you are even slightly technically inclined you will want (a lot) more information than the book provides. At the end of the book I don't feel that I learned much.
Despite many colorful characters and interesting stories, the book drags.
Disappointing; I would not recommend.
on November 6, 2012
Christopher Steiner does an excellent job of explaining algorithms, the "bots" that implement them, and the areas they have been applied to.
Based on detailed interviews with people who developed them, he describes the importance they have in areas of:
intelligence gathering and interpretation,
Automation in the financial markets is one of the major topics of the book. To give one example, his description of Thomas Peterffy's progression from knowing nothing about the financial markets, through building a robot to type orders to buy and sell stocks, to his founding of Interactive Brokers is fascinating.
He documents the diversion of engineers, mathematicians, physicists, and computer scientists first into the financial industry, then out of it. He provides clear evidence that computer automation is affecting social behavior, advertising, commerce, and employment. He makes a strong case for improvement in education -- particularly in areas of mathematics and science.
I am a retired university professor of computer science. During my professional career, I have been involved in some of the areas Mr. Steiner describes, and I continue to write about methods ordinary individuals can use to create their own "bots" to manage their own investment accounts. Christopher Steiner has it right. He has written an excellent, entertaining, and highly readable book. If, as it appears to be the case, the future is Automate Everything, then "Automate This" is a good place to learn what is ahead for us all.
on November 27, 2012
Be warned, this is not an academic history or even a particularly strong business models book. It is just a combination of anecdotes from the authors group at Y Combinator. It has minimal technical detail and for anyone aside from a regular consumer of biz-pulp will be unsatisfying. His attempt to make a distinction between Silicon Valley and Wall Street is downright silly. If you are looking for biz-pulp which really gets at the current ideological climate in the Valley, you have it.
In all fairness, this also could be a five star or one star review. It just depends on what you are looking for.
on December 9, 2014
Algorithms are controlling us more and more. That algorithms can do routinized tasks we already know. Now they can produce and play a symphony à la Beethoven. We are discovering that creativity entails more routines than we was expecting. In fact, all dynamics, however complex and non-linear, have linear dimensions – this is the entry for methodic formalizations. We can do that in positivist mood, reducing complexity to invariant formulae. But we can do that as methodological ability to understand a complex phenomenon by its linear approaches. It’s vey impressing that a computer can easily beat chess champions, do good music, standardize human behavior etc. There is, in the background, a hard epistemological question: the mental tendency to approach complex problems by ordering them theoretically (in logical-experimental format) (we only understand what is ordered, logic, measurable), is it a necessity, an ability, or a defect? Do we understand variation only when we discover how variation invariably varies? Algorithms may suggest it. We are more programmable than we think! Very nice book.