Micro evolution IS macro evolution... Just over a longer period of time.

Everything in life can digest sugars. There's actually rules in bird-watching not to give white bread to eagles and hawks, because it's bad for them... Because they'll eat it. And like it. Because it's simple sugars. On the converse side, deer and chickens will eat mice, even though its out of their usual, it's a good source of protein.

So here's one of many possible hummingbird paths: Standard bird begins: Flies okay-ish, two eyes, two wings, one beak, two legs, you know what "bird" looks like.

"Bird" does not spontaneously transform. As you say, the beak gets slightly longer. Why? Flowers attract bugs. Bugs are good protein. Longer thinner beak means get to eat more bugs hiding inside half-closed flowers. Minor variation cemented over a couple hundred years. Flower needs bugs to pollinate, though. So flower that attracts more bugs, becomes dominant. How? More sugar. Sugar is the basis of all life... Somewhat. "Long-Beak-Bird" that can drink some of the sugar while also eating the bugs gets more energy, more energy = good. ALL life digests sugars. In fact, all life transforms what they eat INTO sugars, so just taking in nectar, which is close to sugar, is great. So Long-Beak-Bird would like to get extra nectar while eating bugs, but the nectar-rich flowers spray it all over the ground when they land on them. Smaller long-beak-bird needed. More agile long-beak-bird needed. Some are fine, and they branch to a different evolutionary branch. Others continue to mate with smaller and more agile mates that live longer since they get more nectar.

Small-agile-long-beak-bird still isn't a hummingbird, but it is closer than "bird". Lets introduce a problem to the environment: Bugs running out. Maybe the world got colder, maybe a new predator arrived, maybe too many small-agile-long-beak-birds bred and ate too many bugs. SALBB now have flowers with excess nectar, since the bugs aren't eating it, but less bugs. The ones able to keep flying with less protein are more likely to survive to mate. And being agile becomes even more important, since with the flowers overflowing with nectar, you get a lot more from them if you don't dispose of them... And being even smaller still is important with less protein to keep your bulk up. The ones able to process the sugar better survive. Others either die off, or branch to a different evolutionary branch. SALBB now primarily eats nectar, minimal bugs. It's also tiny, not small. Tiny-Agile-Long-Beak-Necter-Bird, TALBNB. But it needs one more notable trait before TALBNB becomes Hummindbird. A bird that can slow itself in flight is better at flower sipping than that one that can't. (crash into flower slower, less nectar spillage) A bird that can slow itself MORE is better than one that can only slow a little. A bird that can briefly slow momentum to the point of stopping for a moment mid-air? Even better. Hover? It's the same thing, just better at it. Hover... backwards? One that can mildly go backwards even just a tiny bit and slowly is hugely better at getting to new flowers on a branch, able to progress along an entire tree branch of flowers instead of taking a flight for each one. If they can go backwards faster, though, they can dodge predators with it.

And now you have an omnidirectional, agile, tiny, long-beaked, nectar-drinking bird. A hummingbird. All through micro-evolution. Whether or not that's the path "Bird" took I don't know, but off the top of my head I could see that path existing through only minor changes and a couple thousand generations.