Hypertelescopes are being worked on that would combine many, many telescopes into very large arrays of tens, hundreds and even space based arrays of one million kilometers.
Hypertelescopes seem very doable with 25 kilometer baselines in thousands of large lunar craters. They could even go up to hundreds of kilometer baseline hypertelescopes on the moon.
Nextbigfuture reader Goat Guy calculated very large interferometer 'resolution' slightly higher than the Raleigh's limit formula. A bit. Using just the numbers BOLD highlighted about the Keck interferometer, 85 m baseline and 2.2×10?? m wavelength:
AR = 1.22λ/D
AR = (1.22 × 2.2×10??) ÷ 85
AR = 3.16×10?? radians
AR = (…) ÷ 2π • ( 360° × 60 min × 60 sec × 1000 milliarcsec/sec )
AR = 6.51 milliarcsec
6.51 being somewhat larger (less resolving) than the quoted 5 mas. Comes more in line if the 85 m baseline is added with the 10 m diameters of each of Keck's telescopes (comes down to 5.61 mas). Even more likely, is that the 'resolution' of interferometers is somewhat better technically than straight optical imaging mirrors. You know, convolution, edge interpolation, all that.