It's not that racing is more or less dangerous, it's that the risks have changed and the sport has to react to that. Holding a poll on this type of subject at a highly emotionally charged time when the sport has just seen a rider die in an accident is not going to beget objective, dispassionate responses either.
Road safety as a general rule of thumb has largely improved and cycling infrastructure as a result has improved across a large number of countries where the sport is popular. However, races do not take place on roads in their usual form, they take place on closed roads or under rolling roadblock conditions, and a lot of that infrastructure that is there to keep pedestrians, cyclists and motorists safe on the other 364 days of the year can become dangerous in a pro cycling context - as seen from accidents created by avoiding road furniture, getting hung up on the wrong side of barriers, having to hop across reservations and verges and so on.
Issues like were involved in Casartelli's death have been legislated against; helmets are now mandatory. We also see fewer (although they haven't been eradicated) of the type of concrete bollards and blocks in town that were seen in Kivilev's case. But those risks will never be eradicated. Hell, motorsport can't take risk out and they can use closed, purpose-built venues with all manner of high-tech safety infrastructure. Professional road cycling is a sport in the same vein as the Isle of Man TT or the Pikes Peak Hillclimb - while we can continue at all times to make it safer we will never truly succeed in making it safe.
Whether this has any bearing on Mäder's crash is a moot point, but I do feel that the ongoing Premier Leagueification of cycling does play a role in the current shape of risk and danger in the sport. I've railed against this numerous times, but that's usually from the point of view of its impact on the spectacle; this has got me thinking about its impact from a safety point of view as well, however. The ever-increasing professionalism in the sport has massively reduced the gap between the best and worst rider in a top level pro race, but also the way the UCI's tour systems, points systems and so on work has incentivised a very top-heavy infrastructure where teams outside of the top tier (plus the top 2 from the second tier who were part of the battle for licences) have significantly dropped away in competitiveness; the concurrent falling out of the bottom line in Spanish cycling post-Puerto and Italian cycling over the last decade has exacerbated this problem as there are very few ProConti teams with strong enough calendars and enough money to offer riders leadership that wouldn't be better served with domestique roles in WorldTour teams. As a result you have a very strong top level with very few makeweights, which further reduces the gap between the best and worst rider in any given race. Higher pressure on maintaining a spot on a top team may also increase the likelihood of riders taking reckless actions in the determination to keep their man at the front, to not lose time, and similar, and if a rider does take actions like that resulting in a crash, the péloton is typically going faster and has more riders in it than historically, so there's a statistically higher chance of injury. Take for example the pile-up caused by Filip Maciejuk in the Ronde van Vlaanderen, or the accident where Nicole Frain speared into the back of a static péloton that had already stopped because of another crash in the Tour de France Femmes last year - entirely avoidable incidents borne out of the pressure on riders to be at the front when there isn't room, or to be in the group because you lost time at a time when it wasn't expected.
This also then reduces the differences between teams and the amount of the course that is decisive, as more teams can protect their leaders for longer and more riders remain in the bunch. This in turn means that we see smaller time gaps, more riders in contention, and closer racing means a greater level of importance being placed on those areas that are decisive, increasing the pressure on riders in those environments when dropped. It also means larger pélotons arriving at crunch time in sprint stages - especially as GC riders are pretty much never dropped off in these anymore unless they crash - and the rules for acceptable sprint finishes and run-ins have been debated ad nauseam on the board as well. And as long as smaller time gaps become the norm, organisers will have to seek out ever steeper climbs and smaller roads to secure the guarantees of time gaps, seeing how the péloton has treated even the likes of Gran Sasso and Grand Colombier in the recent past. And if there is not the room for a finish at the top (and we don't want cycling to just become a hillclimb competition anyway), then that's going to entail including descents. But at least if the climbs are steep enough to break the péloton up sufficiently, riders can pick their own lines; with consistently small time gaps, riders don't have the faith that they can recover the time that they lose on descents anymore so may end up taking undue risks - something which is not related to the safety for racing of the descent itself, but related to the way riders are taking the stages. Also, the larger packs and groups of equivalent-level riders means more potential slipstreaming and drives the speed riders are descending up yet further.
Arguing about the safety of descents and steepness I think is also a moot point in respect of the Gino Mäder crash and death as well - the Albulapass descent has been taken year after year in the Tour de Suisse without incident and there is nothing particularly notable about it that raises its risk above that of any other major pass that we see in the sport regularly; this is just similar to Wouter Weylandt's crash coming down the Passo del Bocco - highly unfortunate. The organisers also had reason to believe that there would not be anything especially dangerous about it as it's coming off a HC mountain climb - certainly I think some points could be made about certain finishes we've seen over the years where you're seeing a rapid descent right to the line in stages which have not been particularly selective - take that Itzulia stage to Leitza for example - and the suitability of such descents to host finishes in stages of that nature should be challenged. But La Punt I don't think is one of those. With two significant crashes at the same corner in a short period of time perhaps questions could be asked of the signposting of the danger on the route, or questions could be asked if there was perhaps a pothole that should have been filled in or some oil on the road or something like that that might have increased the danger above the norm, but otherwise, I'm afraid it's just part of the risk that road cycling entails. Maybe some ski netting or catch fencing could be temporarily erected at particularly dangerous spots, but then the question would be, what constitutes a dangerous spot, and how dangerous is dangerous enough to merit it? It will never be enough until the whole course is enclosed because there will always be scope for a freak accident, similar to Allan Simonsen spearing off into the only unprotected concrete barrier in the entire length of the Circuit de la Sarthe (the motor racing one, not the cycling one) because the organisers thought it to be practically impossible for a car to go off in such a way as to hit that barrier full on - and they were right... until they weren't, and Simonsen was the unfortunate one to be the victim of the freak accident that enabled that crash.