Tag Archives: review

Brains and machines

Posted on by

The (sub-)title of After Digital: Computation as Done by Brains and Machines by James Anderson intrigued me. It’s a book of a course taught by the author, a pioneer in neural network research, which was foundational for modern AI, and so explains (in perhaps too much detail for some readers and at too basic a level for experts but fine for me!) how computers compute and how brains compute. It starts with the difference between analogue and digital computing – the former with hardware tailored to specific problems, the latter with generic hardware and the software is decisive – and then goes on to describe how neurons and the brain compute – analogue with digital characteristics, much slower than digital computers but massively more energy efficient.

I didn’t take away big messages, but did get lots of interesting snippets. Computers could get smaller as they got faster, for example, because it takes light a nanosecond to travel about a foot (about 30cm). Between the early 1800s and 1860 the time it took to get a message from New York to Boston dropped to instantaneous rather than a week. It made me ponder the relationship between Godel’s Incompleteness Theorem and the eventual capabilities of AI.

Anyway, if you want an introductory course (as a reader or a teacher) to both computation and neuroscience, it’s excellent. But not a general read. (I loved the cover.)

Screenshot 2023-09-13 at 09.26.56

Middling models in the big real world

Posted on by

I’ve been re-reading Mary Morgan’s excellent book The World in the Model: How Economists Work and Think. It’s relevant to the debate that’s been happening on Twitter – sorry, X – about the way economic research has become ever-narrower and more technical. As Richard Baldwin put it: “We are getting more and more precise answers to less and less important questions.” I’m not even sure that the precision is real. But there is a mania for technique, whether econometric, RCTs, or ‘big data’ methods – and above all for ‘identification’. This means being statistically confident that outcomes can be causally attributed to potential drivers. The identification mania is so intense that I’ve kept a gobsmacking email rejecting an article on the grounds that it ‘wasn’t identified’, when it wasn’t trying to do a causal analysis at all.

The book documents the way ‘models’ have become the dominant way economists work, to the exclusion of other modes of analysis, and argues that this means two kinds of work are excluded: big picture and context-specific detail. The professionally high status work is all “middle level stuff”, wrapped in techniques and modes of thought that prove to incumbents that the work passes appropriate professional hurdles. The dangers are obvious: economics is too silent on big issues, too generic on detail, and extraordinarily conformist.

Morgan concludes: “[D]uring the 20th century, modelling became the way to do economics. The term ‘model’ changed from being a noun to being a verb. ….The epistemic genre of creating and reasoning with models requires a craft skill working with highly formal instruments. … [I]t comes to be thought to be the ‘right way’….” What’s more, economists are taking the requirement for modelling as the way of knowing into other domains such as social policy or everyday phenomena. “Now, when economists look at their small mathematical models they see the real word, and when they look at the big real world they see it as a sequence of their small models.”

The Twitter X thread had a few suggestions – write books rather than journal articles, have one-issue editors for journals more often – but such a strong professional way of knowing and doing is hard to shift. I see some signs of change in academic economics, but don’t know if it will amount to a much broader discipline. Here’s hoping more economists will read this book, or indeed the comments by Nancy Cartwright and Angus Deaton on the limitations of fashionable RCTs.

Screenshot 2023-09-11 at 11.27.38

State-sponsored cyber-crime

Posted on by

The Larazus Heist by Geoff White is a fascinating exposition of North Korea’s role in unleashing large-scale hacking and cybercrime on the world, in its efforts to bypass sanctions and bring in money to the benighted (literally) country. The subtitle spells it out: From Hollywood to High Finance: Inside North Korea’s Global Cyber War. The author is an expert cybercrime journalist, which ensures the book is a cracking read. An early chapter describes the 2014 ransomware hack of Sony Studios, which had started trailing a movie portraying Kim Jong Un in unfavourable light. But the crimes reported are mainly about the intersection between hacking and the banking system, and organised crime, because the money has to be laundered and conveyed to North Korea or at least to places the regime can spend it. So as well as an elite computer hacking corps, the book describes the process of laundering cash through Macau casinos, or Sri Lankan charities, withdrawing notes from ATMs in central India, and trucking tonnes of cash around the Philippines. And then there’s crypto, the land where grift meets large-scale crime.

Apart from the book being a terrific read, what conclusions to take away? That too few people have really internalised the advice not to open email attachments or click on links. That the mesh of banking regulation increases the burden on the honest without much hindering the criminals. That economists/finance folks pay far, far too little attention to the criminal economy (one consequence of the profession’s laziness in studying only data that can easily been found online – looking at small questions with cool econometrics where the lamp happens to be shining, rather than the big, important questions). And that everybody should be very worried about cybersecurity. I learned so much from the book about the vulnerability of everyday life to online attacks from a hostile state like North Korea – and no doubt the other obvious potential attackers. The Wannacry impact on the NHS is a sobering example.

Finally, the book is co-published between Penguin and the BBC; the World Service hosted The Lazarus Heist podcast. In this maelstrom of misinformation we live in, the BBC is more important than it has ever been.

9780241554272-jacket-large

Platforms and information

Posted on by

I was quite excited by the prospect of reading Matchmakers and Market: The Revolutionary Role of Information in the Economy by Yi-Cheng Zhang. But it didn’t quite work for me, perhaps because the intended audience isn’t clear. The reason for excitement is clear: what the digital economy does is change the possibilities for the use of information in production and consumption. As the classic Hayek article observed, ‘the use of knowledge in society’ is the fundamental challenge of economic organisation. Yi-Cheng Zhang is a physicist and more relevantly the honorary director of Alibaba’s Complexity Research Centre. I hadn’t heard of this but it sounds like it should be delivering some very interesting insights.

However the book – very short – is written in a non-technical manner about how digital platforms operate. It’s key point is a concept labelled ‘infocap’, a sort of possibility frontier for the economic agent’s knowledge. There is an information asymmetry between individuals (who can’t know all there is to know about a firm’s products) and firms (who do know this but can’t know all about an individual’s range of preferences, although they can try to alter these). This is an interesting lens on the familiar platform economics but I think people who have read about platforms will not find much new, while although there are no equations the writing is a bit dense for non-specialists who haven’t read the platforms literature. (There are also loads of typos, so presumably the book wasn’t copyedited. I’m the kind of reader who feels the need to take up a pencil and correct these as I go.) I wish the book had actually been more formalized to highlight what might be new.

So, in short, I’d direct economists to the Belleflamme and Peltz text, The Economics of Platforms, and non-economists to a general read such as Platform Revolution by Marshall Van Alstyne, Geoffrey Parker and Sangeet Paul Choudary, or Cusamano, Gawer and Yoffe’s The Business of Platforms.

Screenshot 2023-09-01 at 09.19.33

The Industrial Revolution: causes, consequences

Posted on by

I’m a sucker for any book about the Industrial Revolution, having grown up in its heartland in Lancashire cotton country, and have just read two. One was a re-read, of Joel Mokyr’s classic The Lever of Riches, as I wanted to refer to some aspects as I draft my new book. In particular, the many examples of time saving – for instance, “Although the paper industry is not often thought of as a typical industry of the Industrial Revolution, the Fourdrinier machine was in fact a revolutionary device. It reduced the time involved in making a given piece of paper from three weeks to three minutes.” The book focuses on the relationship between technological progress and economic growth to explore the perennial questions: why Britain? Why then?

The other is Maxine Berg and Pat Hudson’s new book Slavery, Capitalism and The Industrial Revolution. It’s a synthesis of a lot of recent research on Britain’s role in the slave trade and the economies of the Caribbean, and the interaction between slavery and empire and the way the Industrial Revolution in Britain took shape. I’ve read enough about the Industrial Revolution that I’d come across quite a lot of the material before – such as the links between compensation paid to slave-owners after abolition and the growth of financial services and funding of investment and consumption in Britain. The Bank of England has a fascinating working paper on compensation payments, and the UCL project is magisterial. The authors repeat the claim that the debt the government issued to pay compensation wasn’t paid off until 2015 – understandably, as the Treasury tweeted to that effect – although it’s more accurate to say that low interest rates then enabled the  redemption of some consols.

But in any case the book does a terrific job of presenting a synoptic view of the role of slavery in answering the Why Britain? Why then? questions. History is over-determined, in retrospect. Searching for ‘a’ cause is doomed to fail. This book does not make that claim, acknowledging other explanations in the mix. It offers, though, an important perspective on an aspect of the Industrial Revolution that has only relatively recently come into focus.

Screenshot 2023-08-30 at 17.42.10