Skip to main content

The Atomic Human - Neil Lawrence ****

This is a real curate’s egg of a book. Let’s start with the title - it feels totally wrong for what the book’s about. ‘The Atomic Human’ conjures up some second rate superhero. What Neil Lawrence is getting at is the way atoms were originally conceived as what you get when you pare back more and more until what’s left is uncuttable. The idea is that this reflects the way that artificial intelligence has cut into what’s special about being human - but there is still that core left. I think a much better analogy would have been the god of the gaps - the idea that science has taken over lots of what was once attributed to deities, leaving just a collection of gaps.

At the heart of the book is an excellent point: how we as humans have great processing power in our brains but very limited bandwidth with which to communicate. By comparison, AIs have a huge amount of bandwidth to absorb vast amounts of data from the internet but can’t manage our use of understanding and context. This distinction is a crucial one and I’ve never seen it put better.

There are plenty of other nuggets of fascination. For example, from Lawrence’s time working at Amazon it’s interesting to hear how in the time it takes a customer’s web page to load, the system has to work out in the background where the customer is, what the stock is and where it’s located, from this calculating when to stop offering same day or next day delivery. Another random intriguing part is the rift that effectively killed off the predecessor of AI, cybernetics - Lawrence says its demise was caused by a lie that was a ‘fabrication designed to drive a wedge between Wiener (Mr Cybernetics) and McCulloch (cyberneticist turned AI engineer)’. Frustratingly, though, we are not told who told the lie or why they did so.

What gets in the way of this being a great book are its length and (lack of) structure. The content simply doesn’t justify such an endlessly long feeling book. But I could have coped with that if it wasn’t for the way it’s put together. To say it meanders is a huge understatement. It’s quite ironic that at one point Lawrence comments that at Facebook an ex-colleague discovered that ‘instead of a patchwork quilt you needed to weave a tapestry’. This is no tapestry.There’s a sort of greatness to the plethora of scattergun references repeatedly pulling back to central themes of AI vs human intelligence and the twin foci that Lawrence repeatedly visits of the end of the Second World War and his personal experience, particularly when working at Amazon. It is to a popular science book what a James Joyce novel is to a readable one. Some love Joyce… others don’t. I’m afraid I found it hugely irritating - the book cries out for some imposition of order.

One other small moan - you would think from reading this that Cambridge Analytics had been eminently successful in their ability to use post likes to predict psychometric measures. Yet David Sumpter’s Outnumbered tells us that it was a useless predictor of almost all measures, only likely to have succeeded to a degree with one. I have no reason to doubt Sumpter on this.

I am still giving the book four stars because when you get to those nuggets the content is important and interesting. We could have done with a bit more on the practical aspects of controlling AI - I take Lawrence’s point that the essential is preventing AI from being used to make life-changing decisions unchecked by humans (which probably includes not allowing it to drive cars), but it doesn't really suggest how we get practically from here to there. Even so, it’s an interesting book if you can cope with that near stream-of-consciousness storytelling.

Hardback:   
Kindle 
Using these links earns us commission at no cost to you
Review by Brian Clegg - See all Brian's online articles or subscribe to a weekly email free here

Comments

Popular posts from this blog

Math for English Majors - Ben Orlin *****

Ben Orlin makes the interesting observation that the majority of people give up on understanding maths at some point, from fractions or algebra all the way through to tensors. At that stage they either give up entirely or operate the maths mechanically without understanding what they are doing. In this light-hearted take, Orlin does a great job of taking on mathematical processes a step at a time, in part making parallels with the structure of language. Many popular maths books shy away from the actual mathematical representations, going instead for verbal approximations. Orlin doesn't do this, but makes use of those linguistic similes and different ways of looking at the processes involved to help understanding. He also includes self-admittedly awful (but entertaining) drawings and stories from his experience as a long-time maths teacher. To make those parallels, Orlin refers to numbers as nouns, operations as verbs (though he points out that there are some flaws in this simile) a

Everything is Predictable - Tom Chivers *****

There's a stereotype of computer users: Mac users are creative and cool, while PC users are businesslike and unimaginative. Less well-known is that the world of statistics has an equivalent division. Bayesians are the Mac users of the stats world, where frequentists are the PC people. This book sets out to show why Bayesians are not just cool, but also mostly right. Tom Chivers does an excellent job of giving us some historical background, then dives into two key aspects of the use of statistics. These are in science, where the standard approach is frequentist and Bayes only creeps into a few specific applications, such as the accuracy of medical tests, and in decision theory where Bayes is dominant. If this all sounds very dry and unexciting, it's quite the reverse. I admit, I love probability and statistics, and I am something of a closet Bayesian*), but Chivers' light and entertaining style means that what could have been the mathematical equivalent of debating angels on

2040 (SF) - Pedro Domingos ****

This is in many ways an excellent SF satire - Pedro Domingos never forgets that part of his job as a fiction writer is to keep the reader engaged with the plot, and it's a fascinating one. There is one fly in the ointment in the form of a step into heavy-handed humour that takes away its believability - satire should push the boundaries but not become totally ludicrous. But because the rest of it is so good, I can forgive it. The setting is the 2040 US presidential election, where one of the candidates is an AI-powered robot. The AI is the important bit - the robot is just there to give it a more human presence. This is a timely idea in its own right, but it gives Domingos an opportunity not just to include some of the limits and possibilities of generative AI, but also to take a poke at the nature of Silicon Valley startups, and of IT mega-companies and their worryingly powerful (and potentially deranged) leaders. Domingos knows his stuff on AI as a professor of computer science w