Applying for a research grant entails an elaborate folie á deux between applicant and funding agency. The applicant describes in meticulous detail a programme of work, half of which has already been completed and the other half which probably never will be, at least not in the way that is proposed. The more original and important the research question, the more likely it is that this future half will never reach completion, because there is no road map into the unknown. There are of course exceptions, such as clinical trials, sequencing projects, epidemiological studies and genetic analyses, where it is reasonable to expect that what is proposed will be carried out. The current peer review process reinforces this approach to the funding of science. Referees and panel members are complicit with applicant and funder in judging the work of fiction that characterises many grant applications. - Duncan Hull