He was right. In terms of sustained effort and difficulty, writing a novel is comparable to getting a PhD or creating and shipping a major embedded application from the bare metal on up. Maybe be more difficult.
Getting a PhD requires considerable persistence and a modicum of intelligence, but there are supporting and sustaining systems that make it easier. You have an advisor and a committee who in the best of all possible worlds give you guidance and encouragement. And while there are those who will make fun of you for not being in the "real world" they will also envy you a bit and respect what they presume to be your unusually large cranial capacity. That kind of ego-boost helps. It's positive feedback on how you're spending your life.
Building any new application is daunting, but even as the technical lead on a computer-assisted surgery system that went from nothing to giving surgeons real-time image guidance in the OR, I had investors, domain experts, and my team all giving me feedback and support and help in dozens of different ways.
Writing a novel is incredibly lonely. You do it all by yourself, every day, for years on end, with no positive feedback from anyone. I don't ascribe much value to writing workshops, which I think exist solely as an ameliorative to the profound loneliness of the novelist's life. A first reader or two who you can trust to give you small pieces of critical feedback on early drafts is what you really need.
Workshops tend to produce work that feels workshopped: stripped of individuality and weirdness for the sake of responding to the demands of the loudest--and frequently most conventional--voices in the room. If you listen to the quieter voices you'll find that some people love what others hate. You have to write to please yourself, and that is almost pathologically egotistical. Most people simply aren't arrogant enough to say, "My artistic vision is so important to me that I'm willing to spend thousands of hours entirely alone doing enormously difficult work to realize it."
The technical complexity of novels is also unparalleled. My PhD was in an area where the data were a mess and the theories proved the truth of the maxim that "theorists have never had any trouble explaining the results of experiment, even when those results later proved to be incorrect". The phenomena driving the whole field later turned out to be an artifact of borderline-fraudulent statistical analysis by the team who "discovered" it. So finding a coherent narrative through the conflicting claims in the literature was extremely challenging, on top of the usual difficulties of performing a novel and interesting precision measurement to test one particular theory. But moving a handful of characters through a series of events and choices that carry them toward a satisfying climax is at least as difficult, and made more-so because there are no fixed points. You can do anything. Without constraints, the writer's discipline is the only thing keeping the narrative whole.
Furthermore, a novel is "technical complexity all the way down". Every sentence, every word, every punctuation mark is an opportunity to screw up, to make a mistake that will jar the reader out of the continuous dream of the story. I've been writing since I was in my early teens, and thirty years of practice was still barely enough.
Most novelists will tell you that the first novel is the worst, because there's really no way of understanding the process without doing it, and I agree with that. If you think you want to write, you should start now, at whatever age you are. Get that first novel behind you. Write it. Re-write it. Edit it. Polish it. Publish it. It'll be worth it, if that's what you really want to do with your life. And technical people really should write, both fiction and non-fiction: it is a great way of exploring and ordering your ideas.
All that said, was it worth it for me? Definitely. I learned more about writing and the ideas I was exploring than I could have any other way. Whether the end product is worth it for anyone else is left as an exercise for the interested reader: http://www.amazon.com/Darwins-Theorem-TJ-Radcliffe-ebook/dp/...
It might have helped... or not. Although there are lots of things you can learn from short stories, they are not all that similar to short stories in terms of the skills an author needs.
For a short story writer the primary challenge is to introduce your characters, their problem and their actions in as few words as possible. Short fiction is focused. You've got one or two main characters who have one problem.
Novels sprawl. They have many characters, many problems, and are typically spread over a much large time-scale. There are novels that aren't much more than "longer short stories" but they tend to be pretty thin stuff.
So for a short story writer the problem is cutting out anything that is not absolutely necessary to the finished whole. For a novelist the problem is keeping in control of the huge unwieldy structure, because it's just really difficult to hold it all in your head at once.
Software analogy: micro-controller development vs desktop application development. For the first you're mostly worried about resource usage. For the second you're mostly worried about it turning into an e-mail client, operating system, or EMACS over the course of development.
I had written and published a very small number of short stories before writing a novel, but didn't really study them until after. One of the great things about short stories, it turns out, is that there's so little money in them that there are classes taught by really good authors very cheaply, at least in Canada (for anyone in Vancouver I highly recommend Caroline Adderson's weirdly-named but absolutely excellent course "Fiction Series for the Weekend Student" from SFU, which fundamentally focuses on short story structure: http://www.sfu.ca/continuing-studies/courses/cpw/fiction-ser...)
I did write a couple of throw-away novels before writing "Darwin's Theorem", and while that process was useful, it didn't really prepare me for the real thing, which involves far more re-writing than anything else (at least the way I do it.) A book you're writing for practice can have all kinds of flaws... a book you're writing to publish has to be as perfect as you can make it (think of it as the difference between a weekend project for your own amusement vs code you're going to ship.)
Sure, that's fair. I guess my point was that it's hard to produce something of value without some kind of practice before hand, in some form. (Maybe short to novel isn't a perfect transition, yes.)
I think he might mean that writing for human consumption can be harder than most people think.