👾 Econ 04 | AI in Production: Goods vs Services

In the previous article, while discussing how AI/Robotics could be applied in industrial production, we touched upon the dichotomy of goods and services. Let’s explore this further. 

While we all know the distinction, for completeness’s sake, and in the first-principles spirit of these articles, let’s lay out definitions of these to start with. 

Goods are physical outputs as part of a standard manufacturing process. Within the Production Equation (inputs + processing = output) I’ve mentioned repeatedly in this series, the inputs in this case are mainly physical, and the processing is also essentially physical - using the hand, or tools that enhance manual processes. 

Services, by contrast, are essentially cognitive - it’s part of what’s called the knowledge-based industry. Increasingly, a large number of us, especially in the industrialized world, are likely to be part of this knowledge-based activity - including me, as I type away to bring you this very article!

Manufacturing is often called blue-collared work, whilst the services industry white-collared. 

What’s funny is that the historical expectation of automation was that this would culminate in the blue-collared world, with all manufacturing getting fully automated, and only in the distant future would machines come for white-collared work in any significant manner. And even within the cognitive space, the more mundane work (accountants?) would be exposed to automation first while the more ‘elevated’ types (doctors, artists) would be at the very end of the timeline.

However, this is not how things are looking currently. Many were jolted into this reality with the release of ChatGPT in November 2022 - those who realized that, and as I’ve enthusiastically highlighted before, this is the first time in all history that we have a machine that can convincingly speak like a human. 

So what does this all mean?

Going by the current trends it’s definitely likely that services will see faster automation compared to manufacturing. This is because the cognitive type of AI - such as the large language models (LLMs) and similar that power the field of Generative AI (generating language, and by extension, images, and other modalities), is advancing at a dizzying rapid pace. 

In contrast Robotics, which does have an increasingly significant AI component, appears to be facing some rather big challenges - the biggest being fashioning a mechanical hand that can match the dexterity and versatility of the human hand - touch, grip, lift, etc. But there have been some notable strides in recent days, and it may be the case that using the exponential capabilities afforded by AI-driven research as a whole this could catch up to be on par with the purely cognitive areas. 

Another key development, perhaps still in its infancy and one that could accelerate to change the entire landscape considerably is what may be called convergence. 

This can in fact be seen as a wider technological trend, whereby things in the physical realm are slowly getting moved to the digital realm. The switch from paper letters to SMS and emails is perhaps the most prevalent and, shall we say, by now mundane, example. 

With advancements in the medical field, such as gene-editing, the process of trying and testing new medicines is moving in this direction. With increasingly sophisticated application of traditional AI Machine Learning techniques, we are getting closer to being able to simulate functions of human organs within the digital space, which should open up new doors for medical diagnosis and pharmaceutical testing. 

Similarly, in the field of manufacturing the advent of 3D printing has entailed the digitization of making physical items, by simply specifying their design using a computer program. In fact this has been already successfully applied even in the realm of printing human body organs which have then been implanted in patients. 

I’ll emphasize that some of these developments predate the recent explosion of AI, but the application of some of the latest techniques can be expected to serve as an accelerant in such areas. 

So what is the conclusion, who is going to get ahead - will it be goods, will it be services? Or will they fully converge? We might have to keep exploring this.

—

Subscribe to FF Daily to get the next article in this series delivered straight to your inbox.