Could someone explain or point me to a resource as to why using multiple buckets in the HyperLogLog algorithm makes it more accurate than a single bucket?
Something I've been thinking about recently: does anyone think Slack followed Thiel's advice? It entered a space that already existed (didn't go from 0 to 1), with several competitors, solving a problem that was widely known (email is a broken way to get work done). Yet it's becoming a monopoly and will likely be a very valuable business. It seems like there can be tremendous value in making something really good when the existing solutions are not good, and that creating an extremely valuable business doesn't require doing something entirely new.
I also originally thought he had an array of arrays, but now I I think the run-times he mentioned for insertion and deletion would only work for a single array, not an array of arrays.
I like pair programming sessions a lot because both the candidate and the employer are investing an equal amount of time. While the intention behind take-home projects is good, the result is often that the candidate invests several hours into a project just to be quickly passed over by the employer.
What do you mean by this "The real people getting screwed by these paper unicorns though are the late stage employees being sold options as compensation"?
I'm currently a student job searching and this seems pretty relevant.
I've been using this one for several years and have been very happy with it. It is a descendant of several other templates: http://www.sudo.ws/todd/resume.html