Twenty years ago, construction began on the Laser Interferometer Gravitational-wave Observatory (LIGO). Advanced LIGO, with a factor of ten better design sensitivity than Initial LIGO, will begin taking data this year, and should soon make detections a monthly occurrence. While Advanced LIGO promises to make first detections of gravitational waves from the nearby universe, an additional factor of ten increase in sensitivity would put exciting science targets within reach by providing observations of binary black hole inspirals throughout most of the history of star formation, and high signal to noise observations of nearby events. Design studies for future detectors to date rely on significant technological advances that are futuristic and risky. In this paper we propose a different direction. We resurrect the idea of a using longer arm lengths coupled with largely proven technologies. Since the major noise sources that limit gravitational wave detectors do not scale trivially with the length of the detector, we study their impact and find that 40~km arm lengths are nearly optimal, and can incorporate currently available technologies to detect gravitational wave sources at cosmological distances $(z \gtrsim 7)$.