> limiting selection does not make things safer.

True if you are only concerned with the data void of any timeframe. I'm speaking about the public momentum causing the reduction in known poisons. Progress always starts somewhere.