Only for really big things like Instagram that have a network effect helping them, and as such threaten the incumbent.

For smaller features / ideas, it's much less risky to copy than to buy a company (for, say, 10-100M) and hope the acquisition goes well. Most acquisitions don't go all that well.