Our robot overlords?

I’ve chatted to Martin Ford about his new book Rule of the Robots for a Bristol Festival of Ideas event – the recording will be out on 6 October.

It’s a good read and quite a balanced perspective on both the benefits and costs of increasingly widespread use of AI, so a useful intro to the debates for anyone who wants an entry into the subject. There are lots of examples of applications with huge promise such as drug discovery. The book also looks at potential job losses from automation and issues such as data bias.

It doesn’t much address policy questions with the exception of arguing in favour of UBI. Regular readers of this blog will know I’m not a fan, as UBI seems like the ultimate Silicon Valley, individualist, solution to a Silicon Valley problem. I’d advocate policies to tilt the direction of automation, as there’s a pecuniary externality: individual firms don’t factor in the aggregate demand effects of their own cost-reduction investments. And also policies that address collective needs – public services, public transport, as well as a fair and sufficiently generous benefits system. No UBI in practice would ever be set high enough to address poverty and the lack of good jobs: if you want to pay everyone anything like average income, you’d have to collect taxes at a level more than average income.

But that debate is what the Bristol event is all about!



Actually existing AI

For the first time in a year I managed to get abroad for a few days – the Ambrosetti Forum at the Villa D’Este on Lake Como – and apart from the inherent joy of being somewhere beautiful and sunny and foreign, it gave me plenty of reading time. One of the books I polished off is Kate Crawford’s excellent Atlas of AI. It’s a forensic exploration of the unseen structures shaping the way AI is being developed and deployed in the world, and it is fair to say she is prfoundly sceptical about whether ‘actually existing AI’ is serving society broadly as opposed to making a small minority of (mainly) men rich and powerful. “To understand how AI is fundamentally political,” she writes in the introduction, we need to go beyond neural nets and statistical pattern recognition to instead ask what is being optimized, and for whom, and who gets to decide.”

The book starts with the material basis of the industry, in particular the extraction of rare earths and its voracious and growing consumption of energy. We all surely know about the energy appetite of crypto but one point I hadn’t really appreciated is this: “The amount of compute used to train a single AI model has increased by a factor of ten every year.” The next chapter goes on to discuss the extent to which AI depends on low-cost human labour. I think the way Amazon’s Mechanical Turk works is quite well known – a nice book about this was Ghost Work – but this chapter focuses on Amazon warehouses, image-labelling work (“the technical AI research community relies on cheap crowd-sourced labour for many tasks that can’t be done by machine”), and also – nice neologism – ‘fauxtomation’ when tasks are transferred from human workers to human customers: think ‘automated’ checkouts in shops. The chapter has a nice section discussing the role of time in business models: in an evolution of the industrial organisation of time, the continuing automation of economic activity is requiring humans to work ever-faster. The battle is for ‘time sovereignity’.

There is not surprisingly a chapter on data, underlining the point that it is profoundly relational, but also making the point that the reliance on ‘data’ downgrades other forms of knowledge, such as linguistic principles, and how little questioning there is of where it comes from and what it actually measures. I hadn’t realised that many computer science departments have not had ethical review processes on the basis that they do not have human subjects for their research. It’s also a bit of a shocker to realise that some widely used AI databases are palimpsests of older collections of data embedding unpalatable classifications and assumptions.

There’s a similarly shocking chapter on facial recognition – bad enough in the ungoverned way it’s being used but I hadn’t clocked the latest trend of using it to “identify” people’s moods. And the book winds up back at power. As Crawford writes, “We must focus less on ethics and more on power.” I couldn’t agree more. There are a gazillion ethics statements and loads of ethics research but it won’t change anything. I’d add incentives too, but given the concentration of all that compute, understanding the way AI and power structures interact will shape the kind of world we are in 10 years from now. Excellent book.



The business of digital disruption

It has taken me a shamefully long time to read Will Page’s excellent book Tarzan Economics. Will was Spotify’s economist and before his move into the music business had worked for the Scottish Government. So he has the distinct advantage of being able to wear two pairs of spectacles to look at digital disruption, the policy lens and the lens of the disruptor. The term ‘Tarzan Economics’ refers to the leap businesses need to make from one vine in the business jungle as it descends, to another vine, in order not to crash to the ground. The book combines economic analysis of digitalisation with business advice, and carries off the combination. It isn’t in quite the same league as Shapiro and Varian’s now old classic (1998) Information Rules, but on the other hand is far more up to date and gives us a front row seat for what has happened in the music business.

It was always astonishing to me (see Sex, Drugs and Economics) that the music business reacted to the Napster et al threat by deciding it would be a good idea to prosecute their biggest fans (rather than responding to the demand for unbundling). As Will writes, “If piracy of a particular artist was increasing, the lawyers would be getting worried about intellectual property being stolen, whereas the promotion departments would recognise they had a hit on their hands.” The book identifies the real threat to any content business being the limited amount of time and attention we all have. There are nice examples of more sensible ways of responding (other thany trying to put your customers in prison). One is the Atlanta Falcons having made their stadium a pleasant environment with good, reasonably priced food and no rip-offs. Elasticity of demand is such that revenues per attendee have risen substantially, attending the game winning out for enough people over watching on TV from the La-Z-Boy recliner at home.

The book is structured around (as the subtitle has it) ‘eight principles for pivoting through disruption’. These include for instance tactics for working the dynamics of the long tail and how to price discriminate: voluntary donations can raise more than conventional pricing, for instance, the famous Radiohead experiment in voluntary payment for their album (followed later by release of a special boxed set) being one of the examples here. I found the chapter on the use of data, including Spotify examples, fascinating. There’s a lovely anecdote about Will visiting Richard Thaler, who came up with the answer after the company’s data scientists had failed to discover – despite intense data mining – exactly why Discover Weekly was such a success; I won’t spoil the story.

Will is one of the band of enlightened economists who can think more broadly than many. He gives the specific example here of the limitations of the cost-benefit analysis – for example in deciding whether to invest in a museum; the inspiration and education it provides need to be in the mix along with the visitor numbers and entry fees. Not surprisingly given his Spotify role, he argues that the music industry held on to its damaging old vine for a decade too long but things are looking rosy now – I’m not sure all of the music industry buys into this cheerful, new vine outlook. Nevertheless, it was one of the earliest industries to face profound digital disruption and the lessons are well worth pondering. As Will says, the aim of the eight principles and the examples in the book is not to proclaim answers but to help readers ask the right questions about their situation and pick the most appropriate strategies and tactics.

Whether you are involved in a disrupted business or just interested in digital economics, Tarzan Economics is a terrific read, very nicely written, and with the economic analysis sweetened with plenty of examples. Will will be appearing at the Bristol Festival of Economics talking about it in November.



Spreading the thriving

Jan Eeckhout’s The Profit Paradox: how thriving firms threaten the future of work is a very good read. It’s a game of two halves (yes, someone in my household is permanently watching the football at the moment).

The first half (in fact parts 1 and 2) is a nice synopsis of the reasons the economy (mainly the US but others are covered) tend towards a) concentration across many markets and hence b) diverging fortunes of companies and their employees, and the places where people live. This covers the features of digital technology, superstar phenomena, agglomeration economies – these are familiar to anyone who has been keeping up with the literature, but I applaud how well the book is written. Hooray for an economist who can write so engagingly. This section documents the evidence on concentration and mark-ups, the growing divergence between companies in terms of productivity and profits, and the corresponding decline in the labour share as outsourcing of routine occupations (call centres, cleaners, admin) has progressed. Eeckhout argues that there is assortative matching such that pay and conditions are polarising between people with high value jobs in frontier firms and people with low value added jobs in their contractors.

The book’s focus is on the implications of market power for people as workers, rather than as consumers – although it also notes the excess pricing power too. In sum, it reduces wages, both directly through monopsony power in individual labour markets and also because of the the macroeconomic consequences: with so many people in contingent work with low pay, aggregate demand is inadequate. (Some) firms are doing well but the economy isn’t. And this is the heart of Eeckhout’s argument: “The effect of the tide of market power is lowering wages across the economy.” I find this link persuasive. While there are many economists looking at the elements of this story, the way they are combined here is enlightening.

The second half turns to the much harder question of what to do, starting with an affirmation that for all the disruption the new technologies are a good thing (this chunk reminded me a bit of my own Paradoxes of Prosperity, which first came out in September 2001 and not surprisingly was hardly noticed).

The recommendations boil down to: enforce labour standards; mandate more data openness; and beef up anti-trust policies. In particular (under the last heading) stop big tech making more acquisitions, regulate them rather than break them up (so as not to lose beneficial network economies), and assess market impacts in the round rather than firm by firm. (Tricky to implement but I do remember that in my days on the Competition Commission, as it then was, we often had to include a section of the report on ‘Features of the market’ – problems in concentrated markets do often spread beyond an individual transaction).

I’d agree with all these suggestions in the book but they add up to a meta-suggestion: find the political will to change the institutional architecture so that it delivers fairer outcomes. The technological tides won’t retreat but the effects depend on what institutions confront them. Is Lina Khan’s appointment in the US a sign of lasting change?




Finance and the world

One of the highlights of my career as a  financial journalist was a trip to the Chicago Board of Trade in its open outcry days. It must have been around 1993. I can still remember the physical sensation caused by the explosion of the noise of human voices shouting orders in the huge trading pits as soon as the bell rang. It was an extraordinary experience. But one shouldn’t romanticise open outcry. And as Donald Mackenzie records in Trading at the Speed of Light, one CME trader said to him after the closing bell had rung (this was in 2000): “Look at my glasses. They’re all dirty.” Flecked with spittle from the shouting melee.

I loved this book. I’ve been a keen follower of Prof Mackenzie’s research on high frequency trading since his early articles in the London Review of Books as well as a previous book, An Engine, Not A Camera. Prof Mackenzie describes his research endeavour as ‘material political economy’, and it concerns the interaction between the physics and (literal) construction of financial markets, and specifically HFT, and regulatory or political decisions, or power relations.

HFT is a world which is Einsteinian: the speed of light is the constraint on trade, and what traders (human or algorithmic) know about the market is determined by their physical location and physical connectivity: where are they, to the millimetre; how close to a geodesic curve is their communications connection (and is it fibre optic or wireless or microwave); is it raining (which interferes with certain parts of the radio spectrum more than others); to what extent have they programmed trading activities into hardware (chips and C++) for speed?

I’ve always been deeply fascinated by the physical dependencies and consequences of the online world. For example, I was right to predict in The Weightless World that – rather than the death of distance – digitalisation would enhance the agglomeration effects that concentrated people in certain places (we will see if this changes post-pandemic but I’d still reckon not.) Currently I’m somewhat preoccupied with where the internet is, and with the huge physical investments Big Tech companies have made, the weight of digital activity on climate and minerals.

Trading at the Speed of Light is an amazing, detailed account of why material reality matters for virtual outcomes, and conversely, in the financial markets. Everybody with the slightest interest in modern finance should read it (Prof Mackenzie helpfully flags sections that are technical enough to be readily skipped). The book is based on years’-worth of interviews and attending conferences and visiting remote data centres and masts, snapping photos.

The book describes the arms-race of speed and pushing ever-closer to physical limits. A key interview quotation appears mid-way: “I don’t think there’s any other industry than the finance industry that can pay for it ….It’s mind-numbing to look at this whole industry where you have a lot of people with extended training that spend night and day shaving nanoseconds. Where, if you could put that brainpower to something else, maybe somehthing different……”

Indeed. And yet the HFT we have today is the product of decisions taken by people, politicians persuaded by lobbyists. One of the things I learned from the book is that forex trading remains far more artisanal than share trading, albeit still automated. Alas, this was because of the political power of the banks involved rather than anybody’s deep foresight. Human decisions shape markets shape the world, but the consequences are rarely if ever forseen.

Who knows where it will all end (although, presumably, not well). The book ends by pointing out that ‘material political economy’ is the right lens to turn on both crypto (energy-devouring, CO2-spewing monsters) and the world of Big Tech with its datacentres and algorithmic advertising market. At least in these two cases, regulators are perhaps more aware of the societal challenges than their equivalents were in the early days of algorithmic share and futures trading. But it’s a good while since financial markets served their societies rather than predating on them.