Certainly activating the body's own defenses against cancer cells seems like an effective approach. Side effects are minimal as the therapy is more targeted, but as the article mentions, we just don't understand enough right now about when they will fail or why they fail.
My own results on dara have been underwhelming. The first month raised our hopes of a complete response, but it totally flat-lined by the second month. By month 3 levels were rising again, and now it seems I've bred a more resistant clone. My next stage of treatment may very well include a CAR-T based therapy, or a checkpoint inhibitor.
One thing the article gets wrong is the cost. Insurance pays up to $10k per infusion, but co-pays are capped at < $10k per year. There is a huge market for these novel agents, and insurance companies are on the hook to pay for FDA approved treatments. From what I can tell this is driving massive R&D investment and a large pipeline of new potential treatments.