Nepal is still searching for thousands of people after the catastrophic flood that tore through Rasuwa and the Trishuli corridor on August 26. More than 1,250 people have been confirmed dead, while more than 5,000 remain missing. This is beyond replacing houses, roads and bridges. The question is what kind of Nepal we build next. The Rasuwa disaster cannot be treated as an isolated event. In July 2025, a glacial lake outburst at the same mountain system sent a destructive surge through the Lhende valley, destroying Miteri Pul and killing at least eight people, with dozens missing. Fourteen months later, the same valley and settlements were struck again. When the same geography repeatedly produces the same destruction, rebuilding in the same places and manner is not recovery. It is preparation for the next disaster. Nepal faced this after the 2015 Gorkha earthquake, which killed nearly 8,800 people, injured more than 22,000 and damaged or destroyed hundreds of thousands of homes across 32 districts. The National Reconstruction Authority was tasked with rebuilding 817,000 private houses and retrofitting 70,000. That reconstruction improved seismic safety, but exposed a deeper problem: compliance alone does not create a successful settlement. A review in Dhading found that about four percent of 53,109 reconstructed houses did not meet minimum seismic requirements, with limited technical expertise contributing. More revealing was what happened to compliant houses. A survey found ten abandoned houses among 94 visited because their floor plans did not fit family structures or everyday needs. Research from Dolakha found satisfaction with housing design falling from 4.67 to 2.09 and thermal comfort from 3.30 to 1.87. Standardized designs restricted farming and livestock activities and ignored local climate and materials. We made safer boxes, but failed to make better homes. This should change how Nepal interprets 'build back better.' Better does not mean replacing mud, stone and timber with concrete everywhere. It means using engineering to improve what communities already understand about their land, climate, culture and way of living. A house can have a tested seismic structure while retaining a Tamang roof form, courtyard, timber detailing and spaces suited to an extended family. The standard should be safety; architecture can remain. Nepal must also protect the culture and traditions embedded in its settlements. Development should not force communities to abandon the knowledge, craftsmanship, festivals, social spaces and building practices that have shaped their identity for generations. Many Western countries lost much of their traditional architecture, local craftsmanship and community life during the Industrial Revolution, when rapid industrialization replaced older ways of living with standardized urban environments. Nepal should learn from that history rather than repeat it and later regret what it has lost. Modernization should strengthen Nepali culture, not erase it. The same principle must now apply to roads. For decades, many hill highways have followed river valleys because those corridors offer gentler gradients and easier construction. But a river corridor is not necessarily a safe corridor. Rivers migrate, banks erode, debris flows occupy channels and extreme floods can turn apparently stable road alignments into traps. The 2026 Rasuwa disaster has demonstrated the cost of depending on roads and bridges concentrated beside active rivers. Rebuilding every washed-out section on its previous alignment would simply restore the vulnerability that the flood exposed. There is another problem we must confront: the way we cut the hills to build those roads. Recent research on Nepal's road network identifies road-cut slope failures as widespread and points to rapid, haphazard and poorly engineered excavation as an important human contribution to landslide risk. One study notes that around 70 percent of slope failures on mountain roads can be cut-slope failures. Current research also highlights inadequate geotechnical investigation, inconsistent use of design guidance and the frequent excavation of local roads by bulldozers without prior slope design. (nhess.copernicus.org) Therefore, the answer is not simply wider roads. In some hill settlements, we may actually need narrower roads that require less excavation, provided they are properly engineered for their intended traffic. A road carrying a few hundred local vehicles a day does not necessarily need the same geometry as a national trade corridor. Instead of cutting a large scar through a mountain to create a wide carriageway, engineers can follow the natural contours, minimize cut heights, stabilize unavoidable slopes, manage drainage and use carefully located retaining structures. Recent studies show why this matters: rainfall increases pore-water pressure and reduces the factor of safety of vulnerable cut slopes. (sciencedirect.com) This also opens the possibility of redesigning the hill road network itself. Rather than every settlement depending on one road entering and leaving through the same vulnerable valley, Nepal could develop connected loop or circular networks wherever geography permits. A low-traffic local road could operate as a one-way or controlled-width loop, providing connectivity while creating an alternative route when one section is blocked. The purpose would not be to make every road wider, but to make the network more redundant. A slightly narrower road with two possible directions of access can be more resilient than a large road that becomes a single point of failure. This is particularly important for Rasuwa and other mountain districts where roads, rivers, settlements and hydropower infrastructure occupy the same narrow corridors. Critical highways should be reassessed using updated flood levels, river morphology, debris-flow paths, landslide susceptibility and future climate scenarios. Where feasible, strategic alignments should move toward more stable terrain rather than remain trapped beside active rivers. Bridges should be located according to river behaviour rather than simply the shortest crossing, and settlements should have alternative evacuation and access routes. The Dutch 'Room for the River' principle offers a useful lesson: infrastructure must sometimes give water more space instead of endlessly trying to constrain it. Bhuj in Gujarat offers an important lesson. After the 2001 earthquake, reconstruction relied heavily on owner-driven rebuilding, supported by government, NGOs and communities. More than 200,000 permanent houses were constructed within three years, while traditional forms were strengthened through engineering rather than discarded. Kobe, after the 1995 earthquake, demonstrated another principle: recovery works better when communities participate in deciding what their neighborhoods become. Noto has shown how damaged traditional roof tiles and old timber can be incorporated into public housing. Their common lesson is governance: people must help decide what is rebuilt. This requires a national framework for resilient settlements and infrastructure. Nepal already has the Building Act, building regulations and seismic codes such as NBC 105 and NBC 205. The problem is enforcement. High-risk districts need permanent teams of engineers, architects, geologists and technicians with authority to reject unsafe plans. A national catalogue could offer several tested structural systems and locally adaptable designs, standardizing safety without standardizing appearance. The same catalogue approach could guide roads: define safe design principles but allow alignment and width to respond to terrain and traffic. Reconstruction should also be economically local. Public money should circulate through local masons, carpenters, engineers and craftspeople. Local materials should be improved where appropriate, and maintenance should be treated as part of resilience. Climate finance and Loss and Damage support should strengthen this domestic capacity rather than create fragmented temporary projects. Supporting local craftspeople also helps preserve the cultural knowledge that gives Nepali settlements their character. Most importantly, reconstruction must answer: whose plan, who decides, and who benefits? Disasters can create an opening for powerful institutions to impose solutions that look modern but disconnect people from their land and culture. Nepal should resist both extremes: returning blindly to the past or copying standardized concrete settlements. We should modernize carefully, preserving the traditions and cultural practices that future generations may otherwise wish they had not lost. The flood has taken homes, roads and lives. We cannot bring many of them back, but we can decide what their loss teaches us. Nepal should not rebuild exactly what the river destroyed. We should rebuild what Nepal has learned it needs: safer settlements, intelligently redesigned hill roads, less destructive excavation, connected and redundant road networks, stronger local institutions, resilient architecture, living cultural traditions and communities with a voice. Recovery should not be a return to yesterday. It should be Nepal's deliberate design for tomorrow - modern enough to be safe and rooted enough to remain Nepali. Phuyal is in the Board of Directors of Nepal Advanced Vision in Innovation and Nano Science Center