The same argument could have been made about wildfire deaths and damage through the early 2000s in the US.
It turns out that improved forecasting, warning systems, and well-established emergency response, rescue, and recovery systems, can all tamp down risk.
Looking at your article, the subhead reveals a split narrative: "Natural disasters may cost more money, but they claim far fewer lives than in decades past."
What's begun to occur with wildfires is that events are now exceeding the capability of even advanced modern technological societies to address them. Wildfires now regularly enter the urban core, not merely the urban fringe. That's something unheard of in the US from roughly the 1920s, and there wasn't a notable large-scale urban fire until the Oakland Hills firestorm of 1991. Even that was arguably at the wildland-urban interface (WUI). Contrast the Tea Fire (2008, Montecito & Santa Barbara, CA), the Tubbs Fire (2017, Santa Rosa and Sonoma, CA), Camp Fire (2018, Paradise, CA), Carr Fire (2018, Redding, CA), the August 2020 California lightning fires (2020, multiple locations, 3,500+ structures destroyed), the Marshall fire (2021, Marshall, Superior, and Louisvill, CO), the Mountain Fire (2024, Ventura, CA), the Hawaii wildfires (2023, Lahaina, HI), the 2025 Los Angeles wildfires (2025, Pasadena and Altadena CA amongst other locations), and the 2026 Spokane complex (2025, Spokane, WA, ~850 structures).
That's just within the US (an advanced modern technological society), omits wildfires not significantly impacting urban regions, and probably skips a few. There've also been unprecedented urban fires in Canada, Europe, and Australia within the past few decades after nearly a century of no such instances.
(Europe is an exception: for some reason there were numerous urban conflagrations in the period 1939--1945, e.g., Dresden and Hamburg, though climate doesn't seem directly implicated as I recall. Tokyo, Hiroshima, Nagasaki and other Japanese cities had a similar experience at the time.)
As more facts come out about the Nepal-Tibet glacial flood event of 2026, it's pretty clear that many casualties and damage were unavoidable at all once the event itself was triggered. Avoiding casualties would have required extensive evacuations of a nearly 100km stretch of the Trishuli River, possibly for weeks, with an affected population of over 1 million souls.
I've addressed this in a few recent HN comments: <https://news.ycombinator.com/item?id=49473419> <https://news.ycombinator.com/item?id=49473201> <https://news.ycombinator.com/item?id=49460081>