Automation and Ice Cream

A guy named Ken Fox, who posts frequently at X, designs automation systems–electrical components, pneumatics, and software–especially for the food processing industry.  Here’s his ice cream cone filler at work: video.

There are a lot more videos at his X feed.

When people talk about manufacturing, they usually seem to think about metalworking in one form or another–but there are other important types of manufacturing, including the process industries…refining, fertilizer manufacturing, plastics processing…pharmaceuticals manufacturing…and food processing.

Also, I notice that a lot of people judge the level of automation in a particular company or across an entire national economy by counting robots.  I don’t think this is a very good metric.  How many humanoid robots would it take to equal the performance of Ken’s ice cream cone filler, or any of the other automation systems in his video collection?  You could in principle make a CNC machine tool by having a humanoid robot turn the wheels on a manual machine tool, but it makes a lot more sense to just mount the servos directly on the machine.  Similarly, elevators could in principle have been automated by having a humanoid robot handle the controls, but it was simpler to just build the logic into the system.

There will be a big role for humanoid robots, certainly, but I suspect that in many cases they will be a temporary bridge to a more comprehensive system.

Anyhow, enjoy the videos!

The Social and Economic Influence of AI and Robotics

…some historical precedents.

There is currently much discussion of the impending effects of artificial intelligence and robotics on employment, the economy, and our society as a whole. (here, for example)   I think it’s useful to look at some historical precedents, always keeping in mind the caution that ‘past results do not guarantee future outcomes.’

Peter Gaskell’s book Artisans and Machinery is about the effects of the industrial revolution, as seen by a contemporary observer.   I reviewed and excerpted it here, along with some much later commentary by the British writer and scientist CP Snow.

My post Attack of the Job-Killing Robots (three-part series) is a 30,000-foot view of the history of automation over the centuries and of some resulting automation panics.

Your thoughts?

Retrotech: Making a Tunic, 1700 Years Ago

The tunic was found in the Norwegian mountains.   Textile historians recreated it using the technologies current when it was made–pulling the wool naturally rather than shearing, spinning it into thread (with no spinning wheel), and weaving it into cloth. The labor required was estimated by having skilled people do a sample amount of each task required and extrapolating to the complete garment.

Total labor requirement was 780 hours.   The linked post estimates the cost at almost $38000, apparently assuming Norwegian labor rates.

I don’t think anyone would produce such garments using such expensive labor, though (unless it was for some very affluent niche market) but would use cheaper Asian or South American or even American labor.   Maybe a reasonable number including overhead and supervision would be something like $5/hour. Which still gives a cost of $3800.

And if someone made it for their own use, or that of someone in their family, that 780 hours would represent a pretty large piece of their work capacity for the entire year.

As Paul Graham noted, clothing was very expensive right up to the Industrial Revolution.

Productivity Problems: Is ‘Shunning Technology’ Really the Main Villain?

Andy Kessler, a very smart and generally insightful guy, has a recent WSJ column titled ‘The is One Puzzling Job Market’ and subtitled ‘Why has productivity lagged for so long? Because huge sectors shunned technology.’

This assertion doesn’t feel right to me.   In the case of the healthcare industry, for example, Kessler says “Medicine is unproductive. It’s a doctor-intensive chronic-disease-treatment business. But with prevention and diagnostics to find disease early, perhaps we’d need fewer oncologists and cardiac surgeons.” Perhaps, but it’s not as if diagnostics–mammograms, for example–have been ignored.   Prevention can involve, for example, better diets and obesity reduction–these things are really more about accurate science, proper statistical analysis, and honest and effective public communication than they are about technology per se.

A major technology initiative in healthcare of the the last decade or two has been the wide use of electronic medical records.   While these do have considerable potential, the current implementation reality is different.   I don’t think I have ever heard or read a physician or other healthcare professionals who had anything good to say about these systems.   The perceived productivity impact is negative.

It is certainly true that telemedicine has great potential for productivity improvements, and also probably for better paytient outcomes, since it makes it far easier to get an appointment than is the case with traditional practice approaches.   But some of the same advantages can also come from local clinics with an emphasis on quick availability and more use of nurse practitioners and other alternatives to the need to see physicians for every visit.

As another example of an industry with poor productivity, Kessler cites education.   I think we can agree on the poor productivity. But is the problem really lack of technology? How about the massive administrative overheads, the insistence on instructional methods that don’t work very well (in teaching reading, for example), and the overweening power of the teachers’ unions?   Indeed, schools have been quite eager to spend money on ‘technology’.    The kind of projects that Michael Schrage referred to as ‘sparkly tools’ will not do much good until these other problems are addressed.

In transportation, there are indeed technology improvements that can be made in air traffic control and, for railroads, in rail car tracking and hot-bearing detection to prevent derailments, for example.   But there are also physical infrastructure issues–no matter how great your air traffic control system is, an airport’s capacity is going to be limited by the number of parallel runways, and, in some wind conditions, the availability of crosswind runways.   There are also management and process issues–in freight rail, for example, is the current vogue employment of very long trains, now under the banner of ‘precision scheduled railroading’, really a good idea from the standpoints of productivity and market growth?

Kessler says:   “Bell Labs invented the transistor in 1948, but its parent, AT&T,   had 10 to 20 years of old vacuum-tube inventory and so delayed using transistors.”   This claim makes no sense to me.   I can’t imagine that any company, even AT&T would have built up a 10-20 year inventory of just about any commodity, let alone inventory of items in a field which was already known for rapid change.   And early transistors weren’t cheap, and did have their limitations.

There is indeed an apparent paradox when you consider all the technological improvements of recent years–and then look at the productivity numbers.   But I suspect that much of the cause for this disconnect will be found in:

Mediocre or outright bad management. There is a tremendous amount of wasted motion and effort in a lot of organizations today. There’s always some of this, of course, but my sense is that it’s been getting worse, rather than better.   See for example this article about Google, written by a guy whose startup was acquired by that company.

Google has 175,000+ capable and well-compensated employees who get very little done quarter over quarter, year over year. Like mice, they are trapped in a maze of approvals, launch processes, legal reviews, performance reviews, exec reviews, documents, meetings, bug reports, triage, OKRs, H1 plans followed by H2 plans, all-hands summits, and inevitable reorgs.  

Unwise mergers and acquisitions.   Although company combinations can be beneficial, too often they are done under sets of assumptions that turn out to be, shall we say, optimistic.   How much productivity is lost as a result of all the legal and finance work done to enable these combinations and in the organizational disruption that often follows?   (And then, in some cases, to unwind them via a spinout?)

Excessive regulation, particularly ideologically-driven regulation.   In Washington, DC, childcare workers will now be required to have associates’ degrees.   There are many other examples of pointless education and training requirements.   And the ‘industrial strategy’ programs favored by the Biden administration are very likely to direct resources into politically-favored…but not particularly productive..companies and entire industries.

Bad technology implementations.   There are a lot of examples of technology implementations that seemed promising, but resulted in either complete failure or marginal…if any…productivity gains.   Often, there problems are a result of failing to systematically think about the overall business process and the potential people problems involved.   See the sad story of Target Canada, and Zeynep Ton’s description of retail inventory systems that carry meaningless balances because the work of the checkers, and the way in which the feedback loop from goods availability to sales numbers worked, is not properly understood.

There are certainly many technologies now available, and becoming available, that can greatly enhance productivity.   But it is difficult for any technology or combination of technologies to improve productivity enough to overcome the drag of the structural problems sketched about..and many others.   As Lewis Carroll said, we must run as fast as we can just to stay in place, and if we want to go anywhere, we must run twice as fast as that.   Unless we do something about the sources of the persistent backward motion.

Your thoughts on productivity and technology?

What’s the Deal with Construction Productivity?

The data summarized here indicate that productivity in the US construction industry–both labor productivity and total factor productivity–has been much weaker than productivity for the US economy in general.   Indeed, productivity seems to have gotten worse:

To be clear, the raw BEA data suggest that the sector has become less productive over time. A lot less productive: value added per worker in the sector was about 40 percent lower in 2020 than it was in 1970.

Full paper here.

Your thoughts…

Do these findings seem correct?

If so, what are the causes of the poor and declining productivity?

What are the paths to improvement?