So, I've heard it said in another topic on `Topics` that a single list of topics, as far as it is an identifier for a specific person, is far less accurate than a tracking cookie (or similar technology).<p>Well, yes, maybe that is the case _right now_, but these are convenient arguments to push the technology through. There is nothing inherent to the technology that would prevent the `Topics` from being split such that 8 billion (current world population) unique combinations of topics exist.<p>If all topics are completely orthogonal to others, such that the existence of one topic for some user does not give extra information of another topic existing for the same user, we would expect a list of log2(8 billion) ~= 33 topics to suffice as a unique identifier for each individual user, _reported by the browser to every single website that is visited_.<p>So... obviously there is no technological limitation to this list of topics, it is completely feasible to design a list of 33 (or a little bit more) topics in this way.. so what will keep Google from not introducing more topics (as time goes on and people become accustomed to this feature)?<p>My point, in the limit, potentially a few years from now, this feature will be "pareto-dominant" of tracking cookies, or rather in all aspects better at tracking users than tracking cookies ever were.