I build robotic cells, specifically industrial manufacturing automation. They are definitely designed for the main workflow; they handle the most common exceptions gracefully but when something rare happens they just throw a fault back to the operator.
The response from users (operators, engineers, customers) is overwhelmingly positive. Quality of parts increases, quantity of parts increases, cost of operations decrease, quality of life seems to increase for all involved. My mantra is "People are bad at being robots. Robots are bad at being people."
Just yesterday, I worked on a cell that's automated the placing of small seals in a plastic assembly. For the past 8 years, for 12 shifts per week, there's been a person sitting at a dial table that indexes every 2.5 seconds. They were responsible for placing a tiny part in a magazine with a buffer of about 20 parts, a really fast operator could work ahead and have 30 seconds to stand up and shake their wrists out or take a drink, and then sit back down to place parts. Now, a robot pours the seals out and picks them up automatically; there's still an operator around monitoring stuff like the amount of talc powder on the seals, the quantity of parts in all the hoppers, running pressure tests every couple thousand parts, adjusting speeds and pressures of the multitude of actuators, and generally optimizing the cell. They can use their human analytical and decision-making work because they're not tied down clumsily using their fingers to grab a seal every 2.5 seconds; the robot does that mechanical function without complaint. The robot cannot refill hoppers, it cannot sense when new parts are contaminated with moisture that won't feed through the machine, or when a nipper is getting dull; outfitting it with sensors and algorithms to attempt those more complex processes would not be a productive endeavor.
People use tools because tools are useful. Scope and adjustability are critical; when a tool takes over the entire process and gets in the way, that's a bad thing, but a good tool is far better than no tool at all.
I didn't expect this comment to get this much discussion, I should have been clearer.
In the context of automation, the "robotic revolution" has already happened between the 4 industrial revolutions - basically, robots exist en masse already, we just don't call them robots. The majority of manufacturing and processing is already hyper optimized to only needing a few overseers to make sure the machines are running smoothly.
When talking about the automation revolution, I think it's important to make the distinction between robotic tooling and autonomous robotic systems, even though they evolve from each other. The difference is when humans interface and at what abstraction humans can operate it on. Generally, tooling is constrained to an individual's output. Robotic augmentation is in it's own category as well (human activities being enhanced by machinery).
Tooling is extremely useful, still has a lot of future potential and I don't mean to undermine work on new tooling. Just - usually when we look at robots, especially with the buzz around deep learning, we set the bar for problem solving at the level of autonomous construction because we assume, sometimes incorrectly, that optimizing our current tooling yields little efficiency benefits relative to the former.
An example I know of that resembles the bricklayer problem is synthesis of human singing vocals into MIDI sequence data. There is now a history of products around that check off the boxes - one (Vochlea Dubler) debuted just last year - and every time, it demos well but the potential audience ends up rejecting it, because it does not really add what they thought it would add to their workflow. Even if the results themselves come out usable(already a wicked problem since the DSP has to deal with a multitude of recording scenarios while achieving low latency), users discover that they need to be talented at "singing like an instrument" if they want to play instruments by singing, which is a technical barrier, not an assistance to spontaneous creation. Practically speaking, they're better served creatively by button input tools that work top-down(e.g. pick a scale, then the keyboard only plays notes within that scale) since those define down the medium and therefore perform a creatively assistive function with a legible design paradigm(different scale = different sound).
The response from users (operators, engineers, customers) is overwhelmingly positive. Quality of parts increases, quantity of parts increases, cost of operations decrease, quality of life seems to increase for all involved. My mantra is "People are bad at being robots. Robots are bad at being people."
Just yesterday, I worked on a cell that's automated the placing of small seals in a plastic assembly. For the past 8 years, for 12 shifts per week, there's been a person sitting at a dial table that indexes every 2.5 seconds. They were responsible for placing a tiny part in a magazine with a buffer of about 20 parts, a really fast operator could work ahead and have 30 seconds to stand up and shake their wrists out or take a drink, and then sit back down to place parts. Now, a robot pours the seals out and picks them up automatically; there's still an operator around monitoring stuff like the amount of talc powder on the seals, the quantity of parts in all the hoppers, running pressure tests every couple thousand parts, adjusting speeds and pressures of the multitude of actuators, and generally optimizing the cell. They can use their human analytical and decision-making work because they're not tied down clumsily using their fingers to grab a seal every 2.5 seconds; the robot does that mechanical function without complaint. The robot cannot refill hoppers, it cannot sense when new parts are contaminated with moisture that won't feed through the machine, or when a nipper is getting dull; outfitting it with sensors and algorithms to attempt those more complex processes would not be a productive endeavor.
People use tools because tools are useful. Scope and adjustability are critical; when a tool takes over the entire process and gets in the way, that's a bad thing, but a good tool is far better than no tool at all.