wtf? We are using a finite based numbering scheme which potentially could run out. The question is will we run out? IPV6 has only just been introduced so based on current uptake of IP addresses we probably won't run out for many, many years. But there there are plans afoot to hook more and more devices to the internet, consumer devices, not just computing based devices such as laptops, mobile phones.
As for being too close to the question, I would suggest that it is precisely because we are close to it, that we understand it, that we understand the industry, that we can see trends, make predictions with some degree of accuracy and make a more accurate prediction than someone that is not an expert in technology. Forest and trees is not an idiom that is relevant in this case.
I doubt Stephen Fry knows enough about IPV4, IPV6 number spaces to be able to make an informed comment on whether we will run out, his remarks are probably not based on that level of detail, but based on general common knowledge that IP numbers are running out and we have to do something about it (that something is the instantiation of the IPV6 scheme).
Given that IPV6 hasn't been in use for that long, anyone can make a prediction now, the reality is, we won't know for many years whether that prediction is accurate or not. That is, anyone's prediction now is likely to be as accurate as anyone elses. What is the value of even making a prediction under these circumstances? It's akin to making a weather forecast for the United Kingdom for the 10th June 2015, it has little value and is most likely to be wrong.
As T0 approaches the certainty increases, there comes a point where you are so far from T0, it's pointless making the prediction.
If Stephen Fry in years to come, turns out to be correct, I would suggest it's not because of a well thoughtout, intelligent analysis of the data at the time he made the prediction, but down to sheer luck.