I point out two possible paradoxical difficulties in the important target of the IODP in subduction zones, i.e., drilling the seismogenic zone. First, an area close to the trench axis, with an apparent slip derived from tsunami data, might not be seismogenic in the conventional sense. It could be an area that has slipped with almost no friction along with a seismic slip at depth. This is inferred from tsunami earthquakes that ruptured the shallowest portion of the plate boundary, which has a stable sliding frictional character. This portion might have entered a no-friction state before tsunami earthquakes, due to the pore fluid pressure rising almost to the lithostatic. The area under the lower trench slope, which is believed to be able to be drilled by the riser ship, therefore might not be seismogenic in the conventional sense. Second, even when drilling is done in the so-called seismogenic zone, it might be difficult to penetrate an asperity, if asperities have a fractal distribution. Careful investigation would, therefore, be necessary to interpret the results of drilling and monitoring at deep borehole observatories in the seismogenic zone. Even if a borehole penetrates a barrier portion, however, monitoring the transience of the frictional properties, including pore fluid pressure, would be useful for understanding earthquake occurrence in subduction zones.