I received Hello World for Christmas along with a few other books, and it was my favourite of the set … I actually abandoned one of the other books part-way through as this one was so much more appealing.
Hello World opens with the famous chess battle between grandmaster Kasparov and Deep Blue. Fry shows how Kasparov’s shock defeat wasn’t caused by Deep Blue’s mechanical power, but by how Kasparov interpreted and reacted to the computer’s actions. The engineers who programmed Deep Blue’s algorithm tactically let it appear hesitant. Deep Blue couldn’t play mind games, but the programmers understood human reactions well enough to make Deep Blue seem like it could, which threw Kasparov off his game.
This example lays out one of the three principles which run through the book – that we often blindly trust algorithms because we see them as infallible machines rather than instructions written by other humans. The other two principles are first that all systems, whether human-led or machine-led, are flawed, and second that algorithms and humans working together creates a better future than either rejecting algorithms or replacing humans with algorithms.
Fry splits the book into societal themes, such as “Medicine” and “Justice”, rather than basing chapters on technical details, which reinforces her aim of focusing on how these technologies affect us rather than just what they are. Following the themes allows Fry to draw separate examples together, building from clear-cut failures and successes to ambiguous situations that challenge us to weigh up the consequences of building algorithms to answer questions which don’t have perfect answers.
“Designing an algorithm for use in the criminal justice system demands that we sit down and think hard about exactly what the justice system is for. Rather than closing our eyes and hoping for the best, algorithms require a clear, unambiguous idea of exactly what we want them to achieve and a solid understanding of the human failings they’re replacing. It forces a difficult debate about precisely how a decision in a courtroom should be made.”Part of the conclusion of the “Justice” section.
Throughout Hello World, it’s clear that Fry is aiming for as many people to understand the book as possible. Her writing style is accessible and friendly, and she chooses examples and analogies that anchor ideas well. The explanation of how “random forest” algorithms work is the clearest effective description I’ve seen, while the section on autonomous cars adeptly shows off both their simplicity and their complexity. Simplicity comes from their ingredients: the brain of an autonomous car is merely an assemblage of simpler sensors that we’ve already seen in other items, including GPS, pressure sensors and motion detectors, rather than a unique technology. Complexity comes from aligning those multiple inputs so that the data from each can synchronise into an overall “understanding” of the situation, and from questions like how to define what a “road” actually is to that bundle of sensors. Her explanation made me feel like I understood the difficulty of creating autonomous cars, and why they might not actually be possible in the way we expect them to be, in a new way.
As a necessary result of Fry’s emphasis on widespread accessibility, Hello World is not a full deep dive into any particular subtopic and doesn’t convey as much technical information as the similarly-themed Life 3.0, for example. Yet the book does discuss some complex and technical ideas that are essential to consider when talking about any systems that try to answer questions. For example, the fact that there are multiple types of fairness, which mathematically can’t all be true at the same time, was incredibly interesting. Her explanation of how almost every problem has an “anti-problem”, so changes that reduce some types of error will guarantee more of other types of error, made the usually equation-heavy topic of false positives and false negatives enjoyable to read.
One minor negative about this book is that while Fry passionately argues that we need increased legal definitions and regulations about algorithms and about our personal data, she doesn’t give any details about those regulations could involve. This is understandable, as Fry is a mathematician rather than a lawmaker. Also, and unfortunately, there are many areas where individuals currently cannot do anything beyond making an individual choice to use or question specific services that rely on algorithms, as many of these systems are opaque “black boxes” or hidden “trade secrets” by design.
Overall, I loved Hello World. It was enjoyable to read as a curiosity trip, but thought-provoking enough to revisit and investigate. Fry uses very accessible language to make these concepts as broadly understandable as possible and covers a wide range of ideas and situations which all link back to her core principles well. As a result, readers who already have a strong knowledge of algorithms and programming may find the explanations too simple. However, the questions provoked by these examples, and the encouragement to question what a good approach to prison sentencing, medical diagnosis, or prioritising safety in a car accident should be, can apply to everyone.
Finally, Hello World gives a satisfying and optimistic answer to its title question. Fry argues that the answer to “how to be human in an age of algorithms” is to more clearly dive into our humanity – both our strengths and our flaws. If we understand when human brains are flawed or biased, then we can work together with algorithms that enhance our strengths or provide a safety net for our lapses.