You're absolutely right in the first part - a huge proportion of arable land is used to feed animals. However, this doesn't change the optimal strategy for fee

You're absolutely right in the first part - a huge proportion of arable land is used to feed animals. However, this doesn't change the optimal strategy for feeding people in the medium term, for a bunch of reasons: 1. The arable land used for feeding animals would not necessarily be available for feeding people who are hungry now if it were to be freed up. 2. In general the world's poorest people grow their own food on small plots of land and don't produce a lot of meat. They care about the productivity of their plot of land. 3. Changing the world's eating habits is a gargantuan task. Are you suggesting we stop trying to feed people in the medium term using technology, and allow them to starve while we try to convince westerners not to eat meat? I know this is not what you're suggesting - but it serves to show that we should of course be pursuing all the available routes to alleviating hunger. > OK, let's say we can increase yield by 50%. What comes after that? That's enough to meet the entire calorific intake increase for rice-eating countries by the time population stabilises. By around 2050 the world population will have stopped growing. So what comes after that is that everyone is fed, and we can enjoy a world where very few people go hungry while we fix the much more difficult problems of economics and society. Finally, the UNEP report doesn't talk about the potential for GM crops because it's out of scope - the report is about environmental factors in food production. They do say that their projections could be affected by improvements in GM technology.