If there are people who really find this stuff exciting, then the whole "flavor of the month" thing, be it programming languages, databases, operating systems or whatever starts making sense. They actually enjoy the churn!
Legos? They didn't have any staying power. It was amusing at times to snap them together, but that was years and years ago when I was a kid. I sure don't own any now. It was fun to help a friend make some pixel art with it a couple of years ago, but all alone it would have been boring.
It seems we may have stumbled across a fundamental difference here.
I didn't actually find physics in college very interesting, because it seemed more focused on memorization. But I find reading Stephen Hawking books pretty interesting, even though I cannot think of many practical applications of his theories to my daily life, hobbies, or job. I like to understand how things work, and I do find "how things work" to be fun to play with. Materials and chemicals are pretty interesting to play with, as well.
Perhaps it is a fundamental difference. To me, building and understanding things (whether they're computational, mechanical, chemical, whatever) is interesting, but their ends aren't necessarily nearly as exciting, especially if the end is "a business". Business is necessary, sure, and I'd build things for business purposes as a job, but it's not my main interest.
That's probably even my reaction to the average startup on HN; for many, I don't really care about their SaaS or ecommerce solution or whatever, but they often solve interesting problems en route to getting it to work.
Not every programming task is exciting, but some can be fun.
I was making a map of the level on a piece of paper, and bungled it up (the rooms that were supposed to be adjacent appeared on two completely opposite sides of the paper or something like that). The obvious next step was to write a Python script to take a description of the map (room X: exit north leads to room Y, exit south leads to room Z, etc.) and produce a map (in ASCII art; I didn't want to spend too much time on superfluous graphics).
It was more exciting to write that script and see the map it generated than it was to play the game itself.
I may be weird.
Stage 1 was taking screenshots of the stones and shapes. Stage 2 was outlining the shape screenshots with Paint, and using a simple script to look for the outline color, and other colors would change a background color, so I was left with a simple outline. Stage 3 was using an image display program I created and will be selling soon to create a grid with the shape outlines in the top row and the stones with animals in the other rows. Thus I had all the information needed to figure out the code on a single screen. This was so much more fun than doing it the normal way :)
Oh, and then there was the time I solved the last puzzle in Professor Layton, a sliding block puzzle, by programming it in. I used brute force, which of course didn't work (too slow), but then used memoization techniques and such to speed it up. I not only found a solution, but I found interesting stats, such as the minimum number of moves required. It was even more interesting because it was my first program in Python, so I used it to learn the language. Programming to solve video game puzzles is one of the most fun kinds of programming out there.
Of course programming is supposed to be exciting. It's problem solving at it's more basic, and it involves making staff, potentially even stuff with huge impact on the world.
What the fuck, even cooking and truck driving is exciting, why would programming not be?
Making a pizza is boring. Inviting a friend over, teaching her how to make pizza, and then each of us making one and sitting down to watch a movie? That's fun.
Looking forward to a potluck where I'll bring a homemade pizza and homemade apple pie[1] for other people to enjoy? That's when it starts getting exciting. I love sharing that stuff.
[1] Sagan, universe, etc. Yeah, I know.