James Webb alone cost NASA $10 billion, with $4.5 billion in overruns. It took 30 years to design and construct. Carefully designing things to work the first time is expensive and slow; blowing up a billion dollars is the more effective use of money here.
Their point is that if you "waste" $1B as many as nine times with exploding rockets and it leads to working, good rockets, then its economically better than spending $10B to get it right the first time with no explosions.
Of course there's also environmental harm from exploded rockets, and the potential to never find success before running out of money, but as long as they succeed in getting it working perfectly before they've spent as much as it would cost to be confident of it working on the first launch, they'll be happy.
They're just completely different projects that are not comparable. The Webb telescope had no choice but to work.
The SpaceX team has a choice, and they choose "fail fast". There's a gradient that SpaceX can sit on for their development, and for a suite of companies owned by Musk, people would like to see less fail fast and often.
