Academics who specialise in improving the teaching of "hard" sciences like chemistry, biology, maths and physics are increasing in number and influence at Australian universities. Those in academia who have channelled their energies into teaching are delighted with this development. It means that many committed tertiary teachers can now look forward to a new level of recognition and security. But the acknowledgement of teaching-focused (TF) staff as "real" academics at research-intensive Australian universities has created two unforseen problems: (1) research higher-degree (RHD) student supervision; and (2) publication. Everyone knows that academics are expected to publish. In the hard sciences, publishing several papers per year per academic is the norm. Academics usually achieve this by heading a research group; the group leader's name goes on all publications. In some Australian universities, TF academics in the hard sciences are held to similar publication standards. This puts them in a bind. How can they generate the research group that is needed to produce this kind of output when they are teaching hard science students but researching science education? There is also a second part of this problem--the fates of the students themselves. Nobody in academia would dispute that supervising students is valuable, both for the students and for the academic. But here is the rub. TF science academics completed a PhD in experimental (laboratory) science, not education. If they supervise students in a scholarship of teaching and learning (SoTL) project, are they serving only themselves as supervisors and neglecting the futures of their students? Perhaps the paradox here is best illustrated by the Mobius strip. There appear to be two sides to the problem, but everyone actually wants the same thing. All academics want better-trained students who are mature, collegial and able to solve problems. All academics want to supervise and publish. Students benefit enormously from close contact with an academic mentor during a research project and they want to graduate with a degree that makes them preferentially employable. Perhaps the best possible way to achieve all these desirable outcomes would be to mandate that all higher degree students round out their research program with a service and engagement component (just like their academic mentors). This could be community outreach, work for their school or department or an education project supervised by a TF academic. This ties in with the call for new ways of thinking about science graduate training and scientists' role in society. Broadening a science education to include an understanding of education itself provides a real avenue into the new economy for graduates with traditional and non-traditional science PhD degrees alike.