So here's today's maths lesson:
Firstly, the population of London is approx. 9.3m with the population of the rest of England (as these estimates relate to England not the UK) being approx. 45.7m. If 17% of people in London and 5% elsewhere in England have already had the disease, that equates to a total of approx. 3.9m people in the whole of England.
There are 55m people who live in England. For 17% of that population to have been infected would mean a total of 9.4m people. Subtracting the 3.9m who are estimated to have already been infected gives us a balance of 5.5m people.
The current infection rate (according to ONS analysis) is 0.25% of the population, which equates to 137,000 in England. To reach your statement that 17% of all people in England will have had the disease in 2 weeks' time, would require those 137,000 to infect (directly or indirectly) 5.5m people within a fortnight.
If we are generous and assume that the infectious period for Covid-19 is just 7 days (the shorter the infectious period the lower the R number), then the R number would be the square root of 5.5m/137k which equals 6.3.
I don't think anyone has ever suggested it being as high as that and, if it was, a far higher number of the population than 3.9m would have already had the virus. I'm afraid I'm going to file this alongside the claim that "we'll all have had it by the end of May" that I read on here a couple of weeks ago.