Scan barcode
chaptersnchatter's review against another edition
4.0
An excellent read for people without a technology background. Hannah elegantly explains technology without bombarding you with jargon.
I have a degree in computer science so was already aware of how all the algorithms work. I was also convinced that we need regulation for tech. But I still don’t regret reading this because
1. The way Hannah words things for a non-tech audience is simply outstanding! I was particularly blown away by her explanation of the Bayes Theorem
2. I always knew algorithms aren’t perfect but her human case studies made me realise the gravity of these seemingly “occasional” errors. We, as a community of (future) creators of some of these technologies, need to be extremely considerate of the cost of them going wrong.
I have a degree in computer science so was already aware of how all the algorithms work. I was also convinced that we need regulation for tech. But I still don’t regret reading this because
1. The way Hannah words things for a non-tech audience is simply outstanding! I was particularly blown away by her explanation of the Bayes Theorem
2. I always knew algorithms aren’t perfect but her human case studies made me realise the gravity of these seemingly “occasional” errors. We, as a community of (future) creators of some of these technologies, need to be extremely considerate of the cost of them going wrong.
gj377's review against another edition
4.0
I really enjoyed this! Fry writes an incredibly accessible account of how algorithms are already affecting our lives and the world we live in.
Split into different section - crime, justice, art, medicine etc - Fry details the ways in which algorithms are used, and the differing consequences of a naive reliance on technology that's often not wholly understood. Along the way, Fry explains the machine learning concepts that lie behind algorithmic construction and design, detailing them in an extremely easy to read, but not too simplistic, manner - Fry's expertise is readily apparent in how accessible she makes it all.
This book really is exactly what I love in non-fiction books. Well-written, without being too technically dense, but still well-referenced and fairly balanced. This book is split into digestible sections that nevertheless flow into each other organically. As someone who works in machine learning without having studied it, I liked learning a bit about the science as well. I could have done with it being a bit more in-depth, but that's a personal preference.
In an age where algorithms are becoming inescapable, you could do a lot worse than reading Fry's account of where they can go right, and where they can do wrong. Read it now, before it's out of date!
Split into different section - crime, justice, art, medicine etc - Fry details the ways in which algorithms are used, and the differing consequences of a naive reliance on technology that's often not wholly understood. Along the way, Fry explains the machine learning concepts that lie behind algorithmic construction and design, detailing them in an extremely easy to read, but not too simplistic, manner - Fry's expertise is readily apparent in how accessible she makes it all.
This book really is exactly what I love in non-fiction books. Well-written, without being too technically dense, but still well-referenced and fairly balanced. This book is split into digestible sections that nevertheless flow into each other organically. As someone who works in machine learning without having studied it, I liked learning a bit about the science as well. I could have done with it being a bit more in-depth, but that's a personal preference.
In an age where algorithms are becoming inescapable, you could do a lot worse than reading Fry's account of where they can go right, and where they can do wrong. Read it now, before it's out of date!
iris_parsons's review against another edition
4.25
I read this book - Hello World: How to be Human in the Age of the Machine by Hannah Fry - from 16th to 18th February 2024. The book focused on the role of algorithms in everyday life, which many people may not be aware of.
The first chapter focused on power - mainly on the basics of algorithms to support the rest of the book. It explained the story of Stanislav Petrov, a man who was from the Soviet Union who got and alert that the USA had released four nuclear missiles, and decided not to respond to it because it seemed a bit suspicious. This story is important when talking about algorithms, as an algorithm may not have had the caution that Petrov had, which ended up preventing a nuclear war. Key quote: "People are less tolerant of an algorithm's mistakes than their own - even if their own mistakes are bigger" - this is an idea that continues to be explored throughout the book.
In the second chapter, data was explored, and how connected data can be sold to data brokers - countries have started to bring in laws to protect people from this, like ensuring anonymity, but often people's names are in web addresses, so there are ways around this. Additionally, it talked about the story of the Tesco clubcard - Tesco now had information about their customers, which they could use to their advantage, eventually helping them to win a battle in the market against their key competitor, Sainsbury's, but also did things like unveiling a teen pregnancy to a father.
Justice was explored in the next chapter, talking about the bias of algorithms, but also how it is difficult to completely eliminate this, due to biases in data itself. Justice is one of the areas where people are particularly touchy about algorithms, because freedom is important, even though algorithms are much more consistent than humans, who struggle to form the same verdict on the same case multiple times.
The next chapter was about medicine - if you give algorithms access to people's medical data, it could be released, which would be a massive breach of privacy, but could also massively improve public health as an algorithm would be significantly better at noticing patterns. This also talked about the links between humans and algorithms - algorithms can point out problems, while humans properly identify them.
Next was a chapter about cars and the future of driverless cars - the generic topic this explored was how having things automated leads to a lack of skill in humans, which can become fatal in the moments they need to take over.
Geoprofiling (logging places e.g. where crimes were committed geographically) can be used in crime, which I thought was interesting - yet again, this was illustrating how well algorithms can recognise patterns.
Finally, there was a chapter about art, and how that is extremely difficult for computers to make because it is essentially a lucky dip on what is decided as "good quality" and what becomes "popular", meaning that algorithms can't have a set of rules to follow.
The first chapter focused on power - mainly on the basics of algorithms to support the rest of the book. It explained the story of Stanislav Petrov, a man who was from the Soviet Union who got and alert that the USA had released four nuclear missiles, and decided not to respond to it because it seemed a bit suspicious. This story is important when talking about algorithms, as an algorithm may not have had the caution that Petrov had, which ended up preventing a nuclear war. Key quote: "People are less tolerant of an algorithm's mistakes than their own - even if their own mistakes are bigger" - this is an idea that continues to be explored throughout the book.
In the second chapter, data was explored, and how connected data can be sold to data brokers - countries have started to bring in laws to protect people from this, like ensuring anonymity, but often people's names are in web addresses, so there are ways around this. Additionally, it talked about the story of the Tesco clubcard - Tesco now had information about their customers, which they could use to their advantage, eventually helping them to win a battle in the market against their key competitor, Sainsbury's, but also did things like unveiling a teen pregnancy to a father.
Justice was explored in the next chapter, talking about the bias of algorithms, but also how it is difficult to completely eliminate this, due to biases in data itself. Justice is one of the areas where people are particularly touchy about algorithms, because freedom is important, even though algorithms are much more consistent than humans, who struggle to form the same verdict on the same case multiple times.
The next chapter was about medicine - if you give algorithms access to people's medical data, it could be released, which would be a massive breach of privacy, but could also massively improve public health as an algorithm would be significantly better at noticing patterns. This also talked about the links between humans and algorithms - algorithms can point out problems, while humans properly identify them.
Next was a chapter about cars and the future of driverless cars - the generic topic this explored was how having things automated leads to a lack of skill in humans, which can become fatal in the moments they need to take over.
Geoprofiling (logging places e.g. where crimes were committed geographically) can be used in crime, which I thought was interesting - yet again, this was illustrating how well algorithms can recognise patterns.
Finally, there was a chapter about art, and how that is extremely difficult for computers to make because it is essentially a lucky dip on what is decided as "good quality" and what becomes "popular", meaning that algorithms can't have a set of rules to follow.
zainsdad's review against another edition
2.0
2.5 - extra long vox article - not necessarily a bad thing, just not deep enough to warrant a whole book i think. engaging read though!
james_moriarty's review against another edition
challenging
funny
hopeful
informative
inspiring
reflective
slow-paced
3.5
a deep dive into the ethics, uses, positives and negetives of using algorithms, or ai, in a wide range of aspects.