3 ms·
Maybe it’s not a simple problem, in engineering we’re used to cause and effect in a system. With human scale problems it’s more like observing a cat than a mach
by CraigJPerry 5y ago
Maybe it’s not a simple problem, in engineering we’re used to cause and effect in a system. With human scale problems it’s more like observing a cat than a machine responding to inputs. It seems to me you could fill a page of bullet points with the many causes of homelessness in SF.
If that really is true, then it makes sense to have someone familiar Guide you through various aspects of the problem. Especially the non-obvious parts.
- nuerow 5y ago> "It seems to me you could fill a page of bullet points with the many causes of homelessness in SF. That's a far better outcome than mindlessly brushing everything under the rug with baseless knee-jerk reactions of pinning the blame on moral weaknesses and failings of individuals.
- andreskytt 5y agoIt is a machine responding to inputs but it involves feedback to the level we don’t usually experience with computers. This introduces non-linearity which is difficult for a human mind to process in terms of cause and effect. They are still there, just more difficult to see. Consider Jay Forrester in Brooklin in the 70s. He correctly argued (and was apparently almost beat up by saying that), that simply setting up more housing projects increased the problem. The simplified argument went: The word spreads that there’s housing available, people move to the area, housing gets full but the inertia of the word-spread keeps going and people keep coming in. This leads to creation of more projects (as there are still people on the streets) perpetuating the cycle. Meanwhile, the businesses and residents start moving away from the area because of the projects and many homeless, dropping real estate prices and compounding the problem with another feedback loop. Complex stuff. Whatever the solution if SF (that I like very much) is, it better be looking out for that sort of policy resistance.