A cell /is/ epigenetic state, and the effect of any piece of DNA is profoundly controlled by other pieces of DNA, RNA, proteins, RNAs interacting with RNAs, RNAs interacting with proteins, proteins interacting with proteins, etc. Scrubbing the proteins, RNA, and so forth from a cell without killing it is inconceivable.
That said, I don't want to say that cancer is fundamentally beyond us. I think cheap sequencing and emerging highly targeted DNA editing systems, along with better understanding of the immune response to cancer, will get us most of the way there in a practical sense.
Repairs of all these things are performed by the cell itself without killing itself, so they're surely physically possible. But they're limited by a lack of outside information -- e.g. if the DNA is too corrupt, the repair processes can't check the uncorrupted DNA of other cells. That could be fixed with advanced enough technology.
killing the cell is actually not a big deal, any given cell is generally replaceable.
What you suggest is more or less what I mention; you could, say, try to add back some genetic machinery for the cancer cells to recognize that they're defective, and then they kill themselves. Or help the higher level repair mechanism (the immune system) target cancer cells.
To my understanding, trying to fix anything at the level between whole cell and DNA mutation is much harder.
Yes, I'm not gonna pretend to know what avenues are promising or how long it'll take to invent -- maybe humans will go extinct or obsolete first. I just disagree with the article's pooh-poohing towards "why don't we include in our portfolio of bets some actually trying for a cure someday".
That said, I don't want to say that cancer is fundamentally beyond us. I think cheap sequencing and emerging highly targeted DNA editing systems, along with better understanding of the immune response to cancer, will get us most of the way there in a practical sense.