What happens if a sled gets stuck? What would be the procedure and response time for getting the right personnel down there?
What failsafes are in place for disasters? What if the power grid went offline or there was a flood?
How do you prevent human error? What if people left their gas vehicles on and there was a toxic level of carbon monoxide in the tunnels? What if people jumped or accidentally fell onto a sled while it was descending?
The thing is, we're all MUCH more comfortable with the danger we're familiar with, than the danger we're unfamiliar with. Imagine if planes were just being introduced today - what are the edge case scenarios? What happens if there's mechanical failure at 30,000 feet? What happens if there's turbulence, lightning, birds, volcanic ash? What happens if the pilot is suicidal? What happens if... But we really don't care, do we? Air transport is familiar.
I'm not saying the dangers of new tech aren't important - they're very important; but keep in mind that while the idea of suffocating after a crash or malfunction in a hyperloop tube/tunnel will feel significantly more perilous than any misfortune that may befall you on a plane simply because a plane's risks are now familiar.
When we look back at how planes came of age, it seems like people were idiots to get on them (I recently read about the de Havilland Comet... it disintegrated in mid-air like 3 times in a year - don't quote me on that, but it was bad). Did it have to be so bad? No; of course some disasters could have been avoided with better oversight and accountability, but nonetheless, in the face of risk, the new tech provided such great advantages that its risks were accepted. It's fair to be aware of the way new tech can fail, but we shouldn't dismiss a tech simply because it can't "prevent human error" 100% of the time, or be perfectly immune to failure of any sort.
If we can introduce a new tech that has failure properties similar or better to current tech, but that adds significant value to everyday life, we're netting a very positive gain. We shouldn't let fear any failure prevent us from pursuing new tech, just like we shouldn't let fear of stubbing our toes prevent us from walking.
Without understanding edge case scenarios, we cannot make claims like these. Maybe these tunnels really will be more perilous than planes. Or maybe not. The whole point of asking the 'what if' questions is to clarify these issues.
Airplanes have been studied intensely enough that we know the rates and modes of failure. It's worth pondering the same for new technologies as well.
The innovation here is in digging the tunnels, not their operation.
Actually, this claim isn't based on any particulars of any scenario. It's simply a common human cognitive pattern. We will readily accept a familiar risk over an unfamiliar risk, even if the unfamiliar risk is less risky. We all readily jump in our cars every day despite the statistical fact that car-miles are quite deadly when compared with alternatives like bus-miles, train-miles, or plane-miles.
We're also likely to underestimate the magnitude of a risk when we have some perceived sense of control. Most drivers believe they are above average - so the risk of accidents is undervalued (due to ego). On planes, the risk of harm is overvalued because passengers have no sense of control in failure scenarios.
Like I said, we SHOULD understand how new tech can fail. We ALSO need to be careful about the biases with which we approach a new technology and aware of what we're expecting from it. We tend to overvalue the ways in which a new tech can introduce unfamiliar harm, while undervaluing the ways in which a new tech can remove familiar harm that's currently taken for granted (hence, undervalued). This presents policy/planning problems that can inhibit introduction of society-benefiting innovations. This does not mean that conservative approaches to deploying new tech are wrong - far from it: conservative adoption is responsible - it simply points out that our natural human biases tend to maintenance of the status quo.
My point is that, quite simply, we need to be as realistic and comprehensive in our establishment of the benefits a new tech could introduce, as well as the risks, in order to have a balanced perspective of its overall value.
As for point B you can have doors or what not around the sled. But if people really want to commit suicide by sled there's nothing stopping them. What stops someone from jumping off a bridge with a highway below?