I come at you from a base libertarian, market friendly ideology.
This discussion is no more about ideology than creationism is about ideology: people irrationally believe in creationism for ideological reasons, but that doesn't make opposition to creationism ideological. We're talking about simple, empirically verified economic facts.
The current, very imperfect system of agriculture has a very good track record in the "not starving" department.
Yes, and the system demonstrably works because of futures markets, global trade, efficiency of scale, and other free market mechanisms; it works despite of government subsidies and price controls (which, while costly, are still pretty limited relative to the market).
You can directly measure and demonstrate how using market mechanisms mechanisms reduces risks, reduces volatility, and improves food availability. The more countries utilize these mechanisms, the more secure their food supply is. We aren't talking about a hypothetical, idealized free market solution vs some working real-world compromise, we are talking about quid-pro-quo incrementally-reduce-market-interference-incrementally-gain-security kind of choices.
Conversely, there is not a shred of evidence that government subsidies of agricultural producers achieve the effects you claim they achieve, and there is no theory by which they would accomplish that. You are waving your hands for ideological reasons, I'm not. You're like a creationist arguing "irreducible complexity" despite two centuries of evidence to the contrary. (But there is tons of evidence that agricultural subsidies have numerous negative effects and are mostly a wealth transfer to the rich.)
Supporters of farm subsidies have argued that such programs stabilize agricultural commodity markets, aid low-income farmers, raise unduly low returns to farm investments, aid rural development, compensate for monopoly in farm input supply and farm marketing industries, help ensure national food security, offset farm subsidies provided by other countries, and provide various other services. However, economists who have tried to substantiate any of these benefits have been unable to do so (Gardner 1992; Johnson 1991; Wright 1995).