The "computer revolution" has resulted in two major trends - the networking of all computers, everywhere, via the internet, and the computerisation, and then automation, of many jobs. What is at the end state of these trends, when everyone has broadband, and everything is automated?
Once, in any given process, humans did all the work, except those parts that had been computerised. The emerging paradigm is one where machines do all the work, except those parts that cannot be computerised. In this world, everything that can be automated, has been automated. Instead of relying on a human operator to start and stop processes, in this new world, the machines are running continuously, and only stop and start when they have to - that is, when they reach a point in the process which cannot be computerised. This point cannot be computerised because it relies on some form of human intelligence: creativity, expertise and/or decision-making capability that is not feasible to computerise. Once a human has provided the necessary intelligence, the machines will resume the process that was paused, pending human input.
ubiquitous always-on high-speed internet access
All services, where possible, from both the public and private sector, will be delivered online. Every operation that deals with customers - be they individuals, employees, businesses buying services, or simply members of the public - will be conducted online. This will leave a core of back-end processes that are done in the real world. Those back-end processes will not be done at a high-street location - they can be done whereever there is a secure environment with electricity. Most customers will never see this facility.
White-collar jobs will be transformed, with almost no paper or filing, and, most significantly, the most critical aspect of a white-collar employee will be intelligence. Machines will do all the work, except the parts that require intelligence. The more intelligent the employees, the better the machines will perform.
Blue-collar jobs will be similarly affected. Automation of manual labour (roboticisation) will again leave a core set of tasks which cannot be automated, this being due to variability, which machines are not good at coping with. Blue collar jobs will revolve around smoothing this variability - shaping the raw materials, such that they are suitable for further processing with machines; or, at the other end of the production process, applying machine-generated products to a variable environment. These manual jobs are the physical interface between the machines, and the world.
Thus, automation is causing a shift in the skills required by all workers, away from the ability to execute repetitive tasks, toward intelligence: creativity, expertise, and/or decision-making capability. A symbiosis is emerging, where humans need machines to do the work, but machines need humans to provide the intelligence. Neither party has the capacities of the other; together, both parties can have both capacities.
The shift is easily identified - afterwards, humans' inbound work queue is machine-generated. Before the shift, humans receive work from other humans. After the shift, machines manage the process, and are responsible for delegating work. This is a commerzmachinen - a device for making money, a single system that integrates the efforts of human and machine for the purpose of profit.
This will radically alter the composition of the workforce. Eventually, all jobs will require either skilled manual labour, or reasoned thought. Humans who are disinclined to participate in either of these two spheres will find their employment opportunities are limited.
Where there is no customer interaction, demands for business attire will be relaxed, and working hours made more flexible.
It may be feasible to have the intelligence provided remotely, and continuously, by using teams of humans located at various points around the globe. This may have the effect of permitting, or requiring, people to have two or three of these "follow-the-sun" jobs, which may be paid per minute. This will reduce the role of Head Office to a storage device, for a very large computer.
Central business districts will become less important to businesses, and they will cut costs by relocating away from the CBDs, to regional centres. CBD rents will fall, as will commuter numbers and pressure on transport networks. Cities will remain, but people will work from home whereever possible, and those that do commute will probably do so to an out-of-town location. There will be less demand for centrally located offices; some skyscrapers may be refurbished into residential apartments, hotels, shops, vertical farms, and/or light industrial units.
A firm's website, server and internet connection will become as important as electricity and running water at their premises. Conversely, an employee's web browser, computer and internet connection will become as important as electricity and running water at their home.
People will conduct a vast portion of their lives online; just as we cannot conceive a world without paper, so too will people in the future be unable to conceive a world without the internet, or a world without machines. How did they live, they will ask.
There will be reduced use of energy, due to lower numbers of commuters and onsite staff; this will reduce pollution. The economy will be more efficient due to automated systems and optimised work processes; this will increase the sustainability of our civilisation.
Various other machinens will also emerge - machines for running government, for running energy and transport systems, and so on. Eventually, the various machinens will themselves be subsumed into a planetmachinen: a single, self-regulating, self-aware, self-replicating, bio-electro-mechanical organism.
As machines cannot sense the world for themselves, but must be told about it via models, and have mappings defined between the sensors they use and the variables in the models, there also exists a role for humans in the creation of models and mappings. That is, in the process of creating abstractions that machines can use. Once humans have modelled a given environment, and defined mappings for that model, machines are able to operate the model. Humans will not be needed to operate the model; they will only be needed to create it, and to provide intelligence, where needed (including repairs and upgrades).
Update 2014: Of course, all of the above assumes humans manage to resolve their ongoing long-term survival issues, regarding over-population, loss of biodiversity, resource depletion, and climate change. It seems unlikely that a planetmachinen could evolve in a volatile, greenhoused, resource-starved and ecologically bankrupt dystopia. Indeed, a planetmachinen is likely to be the tool needed to best manage the Earth. Let us hope it will be created before it cannot.