Why do we still have mail? Coal firing plants? Land lines? Diesel trains? Cattle farming? Gas powered cars?
At some point, there was a huge effort in R&D to get these technologies to be reliable and predictable. It wasn’t just a technical effort. Yes, technical breakthroughs were needed (ammonia fertilizer for agriculture, steel milling for railways, asphalt production for roads) but the biggest push was by visionaries who wanted to create an easier future and were willing to go through the insurmountable pressure of overcoming stasis.
Even our polluting dystopia is a wonderland compared to a century ago. Everything was difficult. However, it was the political will, organization of labour, and a changes in culture that allowed for focused work to develop the systems and practices for mass infrastructure improvements (oh, and luck).
A few years ago, in Ontario, we had the start of a movement. The government was offering a feed in tariff that would allow for homeowners and businesses to invest in renewable energy generation. The idea was to stabilize prices so that you could see a very clear ROI for investing in renewables. There was a glimmer of hope that this could spark a huge private investment in power generation, reducing the need to add capacity to the system for peak consumption periods.
Someone imagined what it would be like for every home to generate its own power and to never need to build another coal power plant, with its commissioning, maintenance, and decommissioning, and all the maintenance along the way. This is not to mention that peak power loads usually come from fossil fuel sources more than any other.
However, the idea was that environmentally friendly options could become as reliable as coal or other sources and the result would be long term use. It didn’t catch but it was the type of political will needed to move everyone forward.
The same cycle is happening to the mainframe. The AS/400 became such a reliable backbone of businesses that it still runs many business applications for banks and governments. Only slowly is cloud actually gaining traction and larger companies, and governments, willing to move off of their own infrastructure.
It’s hard to know what trends and features will make something into a reliable and durable technology. However, reliable and durable seem to correlate with long lasting. Either because, hey, it’s another cognitive bias or circular reasoning… you wouldn’t know that it’s reliable unless it stuck around long enough for you to tell. Or, it’s more causal… because it’s so reliable at doing what you sought for it to do that it people just leave it alone.