Newton had it easy. He was financially independent, didn't have to worry about writing grants, and didn't know what a publication list was. For Darwin, science was a country gentleman's pastime, not a job that he needed to sustain himself.

The Industrial Revolution changed all this. For the first time in human history, it became possible to produce goods and services on a mass scale. Industrialization proved so successful that science also fell victim to it, and academic research now resembles a commercial operation.

That means that science is driven by customer demand, because society funds basic research only to satisfy well-defined interests - for example, the discovery of new therapies. To meet this demand, scientists must function as efficient machines that convert grant money into publications. Scientists therefore must give up academic freedom and work only on projects for which they can obtain grants.

What makes this situation worse is anonymous peer review, a system that I consider to be censorship. This system is kept alive by an unholy alliance between a narrow elite of "important professors" and influential journal editors who ensure that published results correspond to the current mainstream paradigms. Productivity is then evaluated by quantitative key performance indicators (KPIs) such as the number of papers and the impact factors of the journals in which they are published.

Quality has been replaced by quantity. Consequently, an exponentially growing number of mediocre publications are produced by a large army of scientists, but we have not grown exponentially wiser in the process. In this saturated market only the most competitive can survive. Indeed, some say that competition has now grown to such proportions that it is hindering the free exchange of ideas and promotes scientific fraud.

Perhaps paradoxically, industrialization might have saved us from this fate, by doing what it does best: streamlining processes to make them more efficient, which would allow scientists to do higher-quality work. Instead, these efforts have failed to deliver. If the current trend continues, the scientific establishment will collapse under its own weight, leaving behind only millions of irrelevant publications.

What should we do to prevent the triumph of industrialized mediocrity over creativity and innovation?

Free science from the KPIs: The short-sighted view that every investment must bring a quick return can't be applied to science. Society must support curiosity-driven research as a long-term investment, without demanding immediate usefulness. Other human activities are not subject to direct accountability: Wars, which involve huge investments with even larger negative returns, have been readily financed by the taxpayer throughout the ages.

Free the scientists: The exploitation of postdocs through low salaries and lack of job security must be abolished. PhD-level scientists should be offered decently-paid contracts for an indefinite period so that they can work on long-term projects free of existential worries. Deindustrialized science would probably need fewer scientists anyway, since eliminating mass production strategies means fewer assembly-line workers, which will make this proposal feasible also in financial terms.

Break the power of the editorial office: Scientific journals should return to their original role of disseminating knowledge, instead of controlling the scientists by evaluating them. Anonymous peer review should be replaced by a system of peer commentary, in which everyone openly takes responsibility for his or her opinion. Some publishers, such as BioMedCentral, a sister company of The Scientist, are already trying this. Electronic publications could then function similarly to weblogs, with comments from readers added continuously.

Give unbiased financial support: The current grant system should be replaced by a grant lottery in which winning project proposals are selected randomly, free from any undue peer influence. Such a lottery would take place once a year. An appropriate entry fee, say $5,000, would discourage frivolous submissions, and labs would not be allowed to submit multiple applications in a given financing round. A high-quality, uniform random-number generator would then select every tenth proposal without any human intervention at all. This 10% chance of winning would be similar to the current success rate at the major granting agencies such as the National Institutes of Health or the national Science Foundation. In the end, the same amount of money would be distributed, but "unorthodox" projects would not be subjected to discrimination.

These suggestions are intended to start a discussion on the future of academic science. We need proposals that can help science overcome its present difficulties.

András Aszódi works on epigenome informatics at the Institute for Molecular Pathology in Vienna.