Many Saturdays, I attend a furniture restoration workshop. Restoring battered pieces to life is inordinately satisfying in itself (as is the interchange with a richly varied, even eccentric cast of characters united only by their craft motivation). But it is also fascinating to reflect on in the context of current concerns over work and automation.
Could this kind of work ever be automated? Unlikely. It’s simply too analogue, too human, the timber (sometimes literally) too crooked.
John, the instructor/boss, has an array of carpentry and upholstery skills that can only be marvelled at. Some of these could be reproduced in machines, although only up to a point. But as impressive as the craft skills are the prodigies of improvisation brought to bear on some of the repairs. Sometimes I laugh aloud with incredulity at the exuberant ingenuity of the resolutely low-tech means used to accomplish a repair that to me looks impossible. ‘Nothing’s impossible,’ John instructs. Thus the contraption rigged up to mend and straighten the leg of a delicate chest brutalised in a previous repair came straight out of Heath Robinson, involving clamps, stray bits of wood, a steel straightedge, sellotape, twine, and a bit of blanket. This in turn illustrates why all material offcuts – everything – is kept, not in the spirit of meanness, but the exact reverse: because of its potential, almost always eventually fulfilled, for imaginative and joyful reuse.
Two contrasting recent stories about industrial automation bring this extraordinary everyday human ingenuity into sharp perspective. One concerns Tesla, Elon Musk’s electric car company, as it attempts to ramp up production of its smaller Model 3 vehicle. Model 3 is crucial for Tesla’s ambition to move out of its high-end niche to compete as a mass-market manufacturer, and it had confidently predicted that by the end of 2017 it would be pumping out 5,000 of them a week at state-of-the-art California and Nevada facilities. Needless to say, automation is a key part of Silicon Valley’s assumption that it can reinvent car manufacturing, and Musk has boasted of his vision of lights-out plants with almost no humans in attendance at all.
That now seems a distant prospect. In its most recent update Tesla conceded that a series of teething problems and production bottlenecks had left workers struggling to produce cars by hand, with the result that just 260 Model 3s were completed in the last quarter. Closer reading suggests that Tesla is waiting to commit to the huge capital outlays necessary to get production up to the planned 10,000-a-week capacity until it knows it can hit the lower target – that is, until it can make the current levels of automation work. In other words Tesla – which is burning money, reporting a larger-than expected loss in 2017 Q3 – is betting the farm on robots doing things better than humans Some investors are beginning to think it is not a forgone conclusion. As Gary Hamel tweeted: ‘Software may eat the world, but hardware is eating Tesla. Turns out making cars is harder than coding. Who knew?’
Well, Toyota for one, which as Fast Company reports is taking a remarkably different approach to advanced automation. Backstory: Toyota vying to be the the largest has always been the most profitable of the major automakers. The Toyota Production System is one of the undisputed wonders of the management world, a living, evolving thing that represents more than half a century of organisational learning.
Nonetheless, in the early 2000s, Toyota went through a bad patch. It ran into quality problems as it chased the global number one spot, culminating in a loss in 2009, followed by hugely embarrassing product recalls and a $1.2bn penalty imposed by the US Justice Department. Out of the debacle emerged a revised production system, now known as the Toyota New Global Architecture (TNGA), and in 2015 a new head of manufacturing, Mitsuro Kawaii.
But in a startling case of back to the future, the new regime represents bold advance not towards more automation but less – a return to the human craftsmanship that the TPS was built on but was neglected in the go-go years. Kawaii worked his way up from the shop floor, and in line with the conviction that automation should ‘grow organically out of human innovation’, he has launched training exercises using string-and-sealing-wax methods to devise small improvements to workplace activity – a bit more sophisticated than the rigs in my furniture workshop, certainly, but recognisably similar in their reliance on craft and human ingenuity.
Kawaii’s view is straightforward and radical: ‘Humans should produce goods manually and make the process as simple as possible. Then when the process is thoroughly simplified, machines can take over. But rather than gigantic multi-function robots, we should use equipment that is adept at single simple purposes’.
These sentences should be stamped on the brow of ministers, civil servants, CEOs – anyone in danger of succumbing to the idea that digital is the automatic answer to a business problem. If they had been we might not have wasted more than £12bn on failed NHS IT and hundreds of millions more on the grotesque inhumanities of Universal Credit, among many other examples. Robots are the apprentice, the servant, not the master; they are used not to cut costs but to free up people to do things better for customers. To rub it in: in tests, Toyota consistently finds that people can assemble cars faster than robots. What’s more, unlike machines they can improve their own efficiency and work quality.
Human by default, supplemented by frugal automation to do the boring bits that humans can't improve on. It's a formula that works for Toyota's vast Georgetown plant in the US. It's one we'd recognise in our weekly workshop in North London, too.