Robotics is still at the stage computing was at when we had just mainframe computers. Robots are used commercially now to automate the dirty or hard jobs or to the repetitive jobs such as cargo handling and agriculture. In medicine, robots are making a big impact. They are used in unmanned air vehicles too. There are stories of automated cars. The technology should move in the direction of mini-computers and PCs being made in a robotic sense so that they can be used at home.

The cost has to come down. It is not a computing problem anymore.It is the physical costs of sensors and hardware holding it back. The breakthrough would be a low-cost 3D sensor which will enable a robot to understand its environment economically. It would enable it to do the household jobs such as serving tea.

Robotics, according to Hugh Durrant-Whyte, a former professor of mechatronic engineering, Sydney University, is the intelligent connection of perception to action.An arm in a factory or a surgical robot do not have intelligence at all. They are just mechanical things, though it is clever mechanics. They are just machines. A robot is different from a machine on account of its ability to perceive its environment, understand it and make decisions on what to do. Each time the decision could be different.A robot has to understand a particular plant and all other plants.It has to do generalization of the perceptual understanding. We humans understand a tree, and all other trees. This is taken for granted. Robotics really asks that fundamental question — what makes us intelligent ?

Though robots of future are feared, the real possibility of being scary comes from human clones, rather than electronic robots.Human clones are biological robots. Some countries think of robotizing their armies in the next 20 years. Robotics should make the world safer and should not harm it. The choice is moral, and not technical.

Robots vs Humans

The new jobs that technology  makes possible more than compensate for those jobs which are lost through substitution. This is the proposition put forward by an MIT economist David Autor. Some pessimists feel that machine intelligence may be revolutionary, rather than evolutionary. Two breakthrough techniques are cloud robotics in which robots learn from one another. It leads to a rapid growth in competence. Another technique is deep learning in which robots process vast amounts of data to expand their capabilities, forming associations which can be generalized.

The tasks which can be affected by technology can be categorized. Tasks could be cognitive and manual, and could also be routine and non-routine. Middle management tasks which are cognitive but routine are vulnerable to automation. If the task is cognitive but not routine, it gains from technological change. It enables faster processing and presentation. Machines take over previously manual jobs and non-routine ones. Many jobs still require a bouquet of skills, flexibility and judgment. They are not so amenable to codification or being performed by robots.

Will robots eat our lunch? It depends.