If I've understood this correctly, the one main flaw this has is that the probability to live to your next birthday is dependent not just on your current age, but also the current year. In other words, if you're 30 right now, the probability you'd need to look up is "the probability to live to my next birthday at age 30 during year 2015". When you're 40, you'll want to look up "the probability to live to my next birthday at age 40 during the year 2025", which could very well be different from "the probability to live to my next birthday at age 40 during the year 2015".
Obviously, though, we don't know how this value will change in future years, so this isn't really a fault of the chart so much as it is an unavoidable limitation. Some sort of model could be introduced for predicting healthcare advances, but I doubt anyone can do so reliably enough to produce a net increase in accuracy here.
During the early 1940s, if you were a young Russian male adult, the chance of living until 1946 was 2 in 3. Pretty strange time period. Each time you walked down the street, you could say "live, live, die, live, die, live, die, live, live" and be statistically correct.
Obviously, though, we don't know how this value will change in future years, so this isn't really a fault of the chart so much as it is an unavoidable limitation. Some sort of model could be introduced for predicting healthcare advances, but I doubt anyone can do so reliably enough to produce a net increase in accuracy here.