Although I'm generally pro more nuclear, this argument isn't massively compelling to me. The failed reactors of the past were considered safe when they were built.
The two large scale events involved a) known problematic design with a lot bad process (Chernobyl, the RBMK, and especially the condition into which the overall system was put before start of the event, would have failed a safety inspection under then-current Soviet rules) b) a plant where owners ignored multiple reports about dangers of tsunami capable of overcoming the defenses, and ultimately failed to contain due to loss of power to run the pumps (especially since all reactors scammed in the area). Again, failing requirements to keep running safely.
Or already tested out (including in an accident) SVBR design... unfortunately, other than the geopolitical issues, it is partially locked by low supply of bismuth (it has a variant without bismuth, but it's not without downsides, iirc)
"The failed reactors of the past were considered safe when they were built."
Honestly, this is a terrible argument. It's like saying 'a few people died from eating bad apples, so we should ban oranges for fear that they are the same'. The technologies being discussed are fundamentally different. Not to mention that we don't apply this 'past performance as an indicator of future performance for different systems' paradigm to any other area of life.
The main difference is that they were not considered fundamentally safe. The newer, safer designs utilize the laws of physics for passive safety. The older ones relied on systems that had to function to prevent failure. It was an engineering design assumption that was wrong (that the systems would always function). That's a huge oversight to not run through emergency scenarios to see what would happen.
If you really want to look at historical data, then we can look the precursor to FAST that was tested for the past 60 years at Los Alamos and the numerous emergency scenario testing of the next gen FAST reactors.