The 'men or machines' alternative is a highly emotive one. There seems to be some justification in tipping the scales in favour of men being responsible where a task involves thinking, but less so where a task involves doing. This paper considers AI in design and robotics in construction as two distinct problems. It suggests that robotics in construction offers the better, more immediate return on research effort. Further, that it needs also to distance itself from the many issues currently being raised in AI which remain of little or no relevance to robotics.