First of all, you cite data about Oakland when I was talking about San Francisco. The cities are different enough that it's worth noting. I will also state that I am well-informed about the matter and have few "misconceptions" in the obvious sense.

> Nearly 8 out of every 10 unhoused people in Oakland were living in Alameda County when they lost their housing.

That's not what the Point in Time count asks or tries to measure. The statistic reported is the location of last known shelter. So, as an example, someone who moves to San Leandro from Fresno to crash on a friend's couch for two weeks and then is asked to find a different place to stay would count as "living in Alameda County" for the purposes of the statistic. Another example: a longtime homeless person who has cycled in and out of shelters in the region for decades counts as "living in Alameda County" even if they first lost their home in Kern County or out of state.

> Primary Cause of Homelessness (Top five responses, Fig. 19)

This is silly data to cite. Drug use is correlated with money issues, interpersonal relationship problems, eviction, foreclosure, inability to keep a job, and more. Maybe if the survey had a multiple response design, the distribution would be relevant.

> And anyway, how could it really be any different?

I can think of dozens.

Greyhound bus stations. If you’ve ever taken a bus across country before, they pick up a lot of people at prisons and they stay on until some west coast city (LA, SF, Portland, or Seattle). That accumulates, and once they stick around for at least a year, they are considered resident (for some definition of housing lost, that means even if they were housed in a hotel once, or couch surfed at the start). Surveys that rely on self reporting are incredibly inaccurate. One was done in Seattle, and found out that 80% of King county’s homeless population was previously housed in pioneer square, an absurdity that put the entire survey in doubt.

It really is in the self interest of homeless oriented agencies and NGOs to present the problem as local as possible. If it isn’t local, then giving out housing will only make the local problem worse (people will start arriving for their free housing from other areas of the country), you can judge your success by how much worse the problem gets, which isn’t popular with local voters.

Without good information, at any rate, it isn’t weird that we are seeing the problem get worse for every billion we throw at it. Eventually the popular cities will just give up trying very hard because they never had the power to fix it in the first place.

> Do you really think people who get evicted for not paying bills automatically get sent to a different city?

When did I say that?

> How do you propose your magic mechanism to send people from SF away the moment they become homeless?

What does this have to do with anything? The vast majority of tenants evicted in San Francisco receive both legal representation and relocation fees starting at ~$7k per person and more if you claim disability, which most do.

There are ~80 nonpayment evictions in the city annually, and there were ~0 from 2020-2023.

> Someone who ends up homeless is going to stay in a place that is familiar to them.

Probably true of the average homeless person, but there are many more homeless people outside San Francisco than in it, so you only need to believe a small percentage of, say, California's homeless population ends up in the city for local homelessness to be dominated by folks who lost their last stable shelter outside the city.

According to https://www.asaints.org/outreach/hotel-de-zink/, that is Palo Alto’s only homeless shelter and it doesn’t mention a residency requirement. It also doesn’t have the bunks you are referring to.