Ancestry's company blog covers genealogy, historical records, DNA-based migration maps and engineering topics like Node.js.
HN commenters link it for surname and name statistics, the popularity of genealogy as a hobby, and how Ancestry transcribed census records.
It may not be so much a great guess as a statistical likelihood! In 2000, there were over 2.3 million Smiths in the United States: https://blogs.ancestry.com/cm/smith-a-short-history-of-americas-most-popular-surname/ (This number has likely grown roughly alongside our population, so let's use 2000 numbers for this.) In 2000, we had 281 million people in the US, which means 0.8% of Americans have the last name Smith.
If we assume every possible SSN could be handed out, a trailing six digit entry could have anywhere up to 999 possible SSNs, so you're firing out a shotgun with a pretty big number of potential targets each time you guess an SSN.
This is really evident when you look at the genetic data from places like Ancestry.com and 23andMe. The states with highest percentage British/Irish are all in the South:
http://www.sciencedirect.com/science/article/pii/S0002929714004765#fig3
http://blogs.ancestry.com/ancestry/2014/04/04/a-genetic-census-of-america/
http://blog.23andme.com/wp-content/uploads/2014/10/Bryc_ASHG2014_textboxes.pdf
IMHO We'll develop learning paths to speed up adoption of new tools, so new junior people will have the skills from school.
Or, we won't need people doing those jobs anymore and they will do something else. There are tons of jobs in history that have been superseded, I sure don't miss some of them https://blogs.ancestry.com/cm/10-of-the-most-disgusting-jobs-in-history/
10+ years ago, Ancestry.com hired people in Asia to transcribe U.S. census returns and vital records. They also used volunteers: http://blogs.ancestry.com/circle/?p=2321
Not sure how it's done now, but I suspect it's very hard to automate owing to inconsistent handwriting styles, unusual names for people and places, footnotes and abbreviations, etc.
Too add a data point, here is a list of the 10 most common <FirstName, LastName> pairs in the United States.
https://blogs.ancestry.com/cm/calling-james-smith-10-most-common-first-and-surname-combinations/
I hesitate to guess ethnicity based on names, but I see mostly Caucasian and Latin names (Smith is represented heavily).
I agree with your sentiment, but a minor nitpick: [Stop using anchors as buttons!][1]
An anchor would be good here if the clicking the icons also triggered a URL change, but as it stands a button would be semantically most appropriate
[1]: (http://blogs.ancestry.com/techroots/buttons-vs-anchors/)