I am always drawn to articles like this which are full of drama about automation doing profound things to the economy and beyond. This article about an article is especially feeble for reasons I don’t even care to go into. But one thing did catch my eye.

Computers pursue drug development - a robot in the UK named Eve may have just found a new compound to treat malaria…

Now what would you suppose that means? C3PO wearing a lab coat doing, uh, sciency things? It looks like in this case some actual manufacturing robots (arms) were added to the scene to try and enhance the verisimilitude to actual robotic applications.

Since my employer is also in the exact business of curing malaria, I have some experience with this kind of thing. The truth is that when people in "drug design" speak of "robots" they are almost always envisioning a machine that has some kind of imager or detector targeted at a position on a 96 well sample plate. The well on the sample plate can be selected so that by loading 1 plate, 96 samples can be sequentially imaged. That’s it.

I propose that such machines be called "automatons". I feel like if a machine could easily have been built with extant technology in, say, 1801, the year the Jacquard loom was first demonstrated, then it is not a "robot".

Sometimes these biotech applications use automatons to collect images which are then processed with very fancy technology such as computer vision analysis. Coincidentally computer vision is often an important component to real robotics systems, i.e. machines that would have been impossible before the 20th century.

The science may be good and the automation may be optimal, but if I could personally build a machine with the same functionality using only brass gears and a hand crank, well, I’m going to consider it as much of a robot as the turntable in my microwave oven.