Scan barcode
rayne_1906's review against another edition
informative
reflective
medium-paced
4.25
3 hours - I often find non-fiction hard to get through, but this was a really easy, interesting read. Hannah Fry has a really engaging writing style, and I found the essays well structured and informative.
pevansgreenwood's review against another edition
5.0
Not perfect, but still the best of the popular books in bias/ethics and AI/algorithms (many of which are quite poor).
The author uses all the usual stories from this field, providing a balanced view of what algorithms can and cannot do and how they compare to human decisioning. The author understands that algorithms – like all tools – are part of us and not something foreign, and the focus should be on how we can integrate these tools with our own decision making processes to make better, more equitable, even more human, decisions, rather than pretending that just because a decision was made without the help of an algorithm this makes it more noble or just.
It would have been nice if the book also addressed knowledge production et al, but then it's a popular book and in that sense its 5*.
The author uses all the usual stories from this field, providing a balanced view of what algorithms can and cannot do and how they compare to human decisioning. The author understands that algorithms – like all tools – are part of us and not something foreign, and the focus should be on how we can integrate these tools with our own decision making processes to make better, more equitable, even more human, decisions, rather than pretending that just because a decision was made without the help of an algorithm this makes it more noble or just.
It would have been nice if the book also addressed knowledge production et al, but then it's a popular book and in that sense its 5*.
drj's review against another edition
challenging
lighthearted
reflective
medium-paced
4.25
Content Warnings
Graphic: Sexual assault, Violence, and Police brutality
macloo's review against another edition
4.0
I've read a bunch of books sorta kinda similar to this one, but this beats most of them for a very specific reason — it's a book of stories about how algorithms are used in the world, and each story has a definite point. Unlike Algorithms to Live By, which tells stories about specific algorithms, this book (despite being written by a mathematician) doesn't delve into the workings of any of the algorithms and hardly names any of them.
Here's a blog post in which I summarized the ideas in each chapter.
Instead, in seven chapters simply titled Power, Data, Justice, Medicine, Cars, Crime, and Art, author Hannah Fry describes examples in each of those areas. Even though I've been reading a lot in the topics of algorithms and artificial intelligence over the past year, most of her examples were not ones I'd read about before. This impressed me! Even when she did use an example I'm very familiar with (e.g. the COMPAS system for assessing risk of re-offending by people who have been arrested by police), I didn't feel a need to skip over it (although I probably scanned those parts quickly).
At the start of the Power chapter, she talks about Deep Blue beating world chess champion Garry Kasparov in 1997. Now, that's a way to get me to toss aside a book! I've read at least five articles about chess algorithms and/or Deep Blue in the past few months! But I didn't toss this book out, because at the end of the brief anecdote, Fry made a point about expectations — "the power of an algorithm isn't limited to what is contained in its lines of code."
For someone who's just looking for a well-thought-out book to tell them what the fuss about algorithms is all about, this is perfect. And even for someone who, like me, has read almost too much about algorithms — it's still a darned good book.
Details about categories of algorithms (prioritization, classification, association, filtering; rule-based vs. machine learning) are handled right up front in the Power chapter.
.
Here's a blog post in which I summarized the ideas in each chapter.
Instead, in seven chapters simply titled Power, Data, Justice, Medicine, Cars, Crime, and Art, author Hannah Fry describes examples in each of those areas. Even though I've been reading a lot in the topics of algorithms and artificial intelligence over the past year, most of her examples were not ones I'd read about before. This impressed me! Even when she did use an example I'm very familiar with (e.g. the COMPAS system for assessing risk of re-offending by people who have been arrested by police), I didn't feel a need to skip over it (although I probably scanned those parts quickly).
At the start of the Power chapter, she talks about Deep Blue beating world chess champion Garry Kasparov in 1997. Now, that's a way to get me to toss aside a book! I've read at least five articles about chess algorithms and/or Deep Blue in the past few months! But I didn't toss this book out, because at the end of the brief anecdote, Fry made a point about expectations — "the power of an algorithm isn't limited to what is contained in its lines of code."
For someone who's just looking for a well-thought-out book to tell them what the fuss about algorithms is all about, this is perfect. And even for someone who, like me, has read almost too much about algorithms — it's still a darned good book.
Details about categories of algorithms (prioritization, classification, association, filtering; rule-based vs. machine learning) are handled right up front in the Power chapter.
.
clmerle's review against another edition
5.0
A nice companion to Automating Inequality. It gives the bigger picture for algorithms in society as a whole. It’s promise and peril.
mattdb1977's review against another edition
funny
hopeful
informative
lighthearted
reflective
medium-paced
4.0
mexscrabbler's review against another edition
5.0
This book provides a very insightful look at areas in Society where algorithms are either replacing or aiding tasks which were traditionally labor intensive. They include very diverse areas: crime (solving crime via patterns), medicine (reading x-rays), art/music/movies (can great art be created by machines? can great art be predicted by machines), technology/data (driverless cars), etc.
It is very interesting to see where algorithms are working well and where they are not. The key takeaway appears to be that algorithms are best applied as a pre-filter to human decision making, as a way of working through vast amounts of data without making human mistakes.
Quite interesting.
It is very interesting to see where algorithms are working well and where they are not. The key takeaway appears to be that algorithms are best applied as a pre-filter to human decision making, as a way of working through vast amounts of data without making human mistakes.
Quite interesting.
titus_hjelm's review against another edition
4.0
I picked this up because some years ago I shared the stage with Dr Fry at TEDxUCL. Her performance there launched her career as a semi-celebrity maths/complexity person. I'm happy to see that it paid off and her debut trade book is a very entertaining and interesting read. But do get it as an audiobook if you can. She also has an extremely pleasant reading voice. The content is rather standard narrative popular science, illustrating a point through real-life stories, but it works very well. I especially appreciated that, in a world divided into luddites and Wired magazine-reading true believers, her finishing sentiment is to say that algorithms will never replace humans, no matter how much they develop and change our lives, yet some things they actually do better than fallible humans.
ojalvo's review against another edition
4.0
A great read for those getting into big data, algorithms, AI and ML. However, for those already deep into algorithmic processes that dominate...well everything of the world today, it may feel a bit dated.