One argument I've heard: if evolution could get rid of sleep, it would have a long time ago. There are several survival drawbacks from sleep, as it leaves one vulnerable. It must be important!
1. The vertebrate retina [1] (cf. the cephalopod retina).
2. The recurrent laryngeal nerve [2], especially in the giraffe [3].
[1] https://en.wikipedia.org/wiki/Retina#Inverted_versus_non-inv...
[2] https://en.wikipedia.org/wiki/Recurrent_laryngeal_nerve
[3] http://blog.eternalvigilance.me/wp-content/uploads/2013/04/g...
How about another example: mammalian females giving birth through their pelvises. This caused the deaths of countless women and children throughout human history. It's not as much of a problem for other mammals but it's a huge problem for humans. We've had to advance modern medicine to the point where we can safely perform major surgery (caesarian section) in order to mitigate it.
It's possible that you could have sleepless neurons. But they would be so different than neurons we are stuck with that we might need to wait few billion years for them to show up.
Saying "if it was better, it would've happened already" assumes humans are perfectly evolved right now, and negates any possibility of future evolution. That argument would be just as valid a 5000 years ago as it is now but presumably humans have evolved since then. Genes mutate randomly, selection pressures change, things in our environment might affect the way genes mutate - its much more complex.
5000 years isn't remotely close enough for sleeplessness to evolve, more like a few million. Its not so much why do humans sleep, but why do apes? Why does anything?
Whatever is going on must be important, and not easy to avoid without serious trade-offs.
The majority of our hominid evolution was in Africa, where we likely developed night time sleep patterns as a defense to the big cats developing better night vision. Other benefits like social interactions, heat-conservation, even directly procreation; As well as the diminished usefulness of any work, from hunting to building in darkness would all play a roll.
I find it fascinating that we likely also developed a fear of the dark as a response to the big cats developing better night vision in exactly the same way gazelles developed better hearing and faster legs.
Creatures could have developed the day/night awake/resting cycle independent of sleep just because at night you want to conserve energy to stay warm, and it's difficult to hunt at night.
Once creature developed day/night modality, taking advantage of that time to run repair processes would be an evolutionary advantage.
The question isn't why people sleep, it's why can't creatures run repair processes 24/7? If they could do that, then evolving to not need sleep would be a lot easier.
Some organisms have more advanced repair mechanisms than others too. There are trade-offs involved, e.g. immune system.
Suppose there exists a far better brain design, but all the brain designs that are part way between the current design and that design are far worse. Any mutation that steps toward the better design will be selected against.
It does make sense in the context of a somewhat hostile environment, where constant awareness is required, but it wouldn't benefit humans as much.
Evolution is kind of like Economics; it's not rational, and it's not currently in equilibrium.
It's a very thorough book and certainly convinces you of the tremendous benefits of sleep.
It's why hamsters are nocturnal. Their predators hunt during the day, so they sleep during the day and play at night.
Evolutionary theory would say the hamsters that like to play during the day got eaten so that only those who slept during the day got to mate and pass on their genes, etc.
If you think sleep is a drawback for humans, trying visiting a forest at night without a flashlight.
Given the path evolution takes it seems a stretch to ever consider it best, perfect, good or anything other than "it worked just enough at the time". Just enough being the key. All we know is that sleep didn't disrupt animals enough to entirely fail the evolutionary test. However plenty of garbage features go along with more powerful features. One in isolation is not proof of it's perfection, I would think.
Sleep wasn't a liability to our survival, especially our ancestors. It actually aided in our survival. Human beings who were active at night would have been the vulnerable ones as they would attract predators and were more likely to perish as a result. Especially considering out vision would be limited in the dark.
In short, sleep kept us away from the dangers at night.
Imagine a hypothetical state of pseudo-sleep that causes us to withdraw during the night (the same as normal sleep), except we keep our eyes open/ stay alert.
We are just as hidden from predators, but now even more equipped to defend ourselves.
Therefore it seems difficult to deny that sleep itself has some form of rejuvenative benefit.
Please try reading my comment or following the thread. I'll paste it here : "I was specifically addressing his point about the drawbacks of sleep - "There are several survival drawbacks from sleep, as it leaves one vulnerable."."
In other words, sleep has rejuvenative benefits and it keeps us away from predators. Sleep can have many benefits. You can stay up all night, but you'd miss out on the rejuvenative benefits of sleep. Or you can safely tuck away and sleep and gain benefits of safety and rejuvenation.