[navpatel]

The same thing that happened to stable boys, bank tellers and farmers. In 1870 53% of the labour force were farmers, by 1970 that was 4.6% (https://www.agclassroom.org/gan/timeline/farmers_land.htm). If you told someone in 1870s that 50% of the labour force will be changing their job, they would have a similar reaction.

Granted the speed with which changes are occurring now is much more rapid; if we consider the largest number of people affected in the smallest period of time, it's probably self-driving cars & trucks that will make the greatest impact.

Driving or making pizza isn't exactly high-skill work (in terms of time taken on training), hopefully the transition happens slow enough for these people to retrain for other work.

[ec109695]

The difference between then and now is that farm jobs weren't replaced by jobs that requires a college education.

[yourapostasy]

> Driving or making pizza isn't exactly high-skill work (in terms of time taken on training), hopefully the transition happens slow enough for these people to retrain for other work.

This doesn't address the joblessness concern. The trend I see taking root in its infancy today is job stability tending to flow increasingly only to highly-skilled and highly-experienced staff, and many times not even that imparts stability. This has downstream impacts upon complexity management, retention, etc. that are out of scope of discussion in a format as limited as HN.

The game changer is that software and computationally-enhanced hardware delivers the ability to re-use pre-existing infrastructure. Earlier transitions at least required significant infrastructure changes and/or outright new infrastructure be built, and that formed the opportunity and basis for disrupted workers to work within the transition. We might be talking about a Zume that requires new infrastructure today, but make no mistake, a solution reusing pre-existing infrastructure as-is will come down the pipeline. This yanks out the traditional "relief valve" of earlier large-scale technological transitions' disruptive patterns.

When we replaced horses with cars, there was massive infrastructure added for roads, roadside entertainment, oil and oil refining distribution-to-retailing, pipelines, eventually shipping, oil discovery, etc.

When we replaced steno pools with word processors, the "infrastructure" was basic typing skills (shorthand skills were dispensed with, to our detriment IMHO) imposed upon a new generation of workers as a new baseline skill (along with many other baseline skill upgrades with no commensurate increase in wages over the past decades, I might add). The physical infrastructure of steno pools however, was repurposed.

When we replaced the extra cashier at gas stations with card-accepting gas pumps, the infrastructure initially rode on the back of the pre-existing telephone network, then upon the new data networks. The major infrastructure replaced was the gas pumps themselves, and usually only as they gradually wore out so the infrastructure draw for new labor was a net negative when the lost cashier jobs' hours were taken into account.

When we eventually replace the plurality of in-house small-business system administrators with cloud-based services, the physical footprint of the servers will be re-purposed to maybe a storage closet.

We're also using metrics in this discussion the wrong way. Instead of talking about number of jobs lost/gained, we should be talking about absolute wages over time and per capita, with an additional factor expressing stability over long periods. Using number of jobs as the metric loses information about the quality of the jobs, stability over working life, and dispersion of the quality and stability throughout the general working population. If we are being honest with ourselves within the software industry, then we are major enablers to devolving the quality and stability of what jobs that do exist. For example, through the delivery of software that enables decision makers to take what used to be highly unprofitable processes of manually scheduling staff and following up attendance to ensure only part-time employment, and making it not only feasible, but highly profitable.