In 2012, it's 8 TeV by the way. Hopefully 14 TeV in 2014.
It's a little more complicated than looking at the total center-of-mass energy and saying we can discovery any particle up to the max. A single proton is made of multiple constituents, and a proton incoming with 3.5 TeV (or 4 or 7) of energy represents the total energy of that system. When two protons interact, it's actually two constituents which are interacting, and they will have some fraction of the proton's energy. So typically the probability of producing particles drops considerably as you look for more massive particles.
That said, the central sentiment of your message is correct. There is a lot of potential signals that remain to be investigated. There could even be particles found with considerably less mass than the Higgs, but which have an unusual decay signature which we haven't been sensitive to yet.
What happens next is we study this particle. We want to know if it behaves as is predicted by the standard model, or if it's something different from what we expect. This includes measuring its cross section (the probability of it being created in collision) and its branching ratio (the probability of it decaying to each thing its able to decay to).
Matt Strassler (a theoretical physicist) describes the general roadmap in his blog post here.
Particle physics results are necessarily esoteric. What do we do with experimental knowledge? We use this knowledge to disprove plausible theory and to constrain future theory. Theory is similarly used to give direction for new experiments.
The innovation is fine. The problem is that Netflix' leadership has been unable to communicate with its customers in an intelligent way. They need to tuck us into bed, and tell a bed time story which ends "and then you bought our new product and lived happily ever after." This is what Apple does when they innovate.
Instead, you can look at the Quickster announcement. First paragraph: "I messed up. I owe everyone an explanation." Second paragraph: Talks about how they treated their customers like idiots by saying that they were lowering prices when the prices increased. Paragraph ten: Announce Qwikster.
Why even combine these two messages into a single product announcement?
Fundamental science's goal is to understand how the universe we live in behaves. We accumulate evidence about how the universe works (called "measurements"), and use it to rule out incorrect possibilities.
It's the job of others, such as scientists who don't work on fundamental research (called "engineers"), to decide what we can do with the universe we live in.
All nuclear reactors, at least in the west, are insured. Not only so, they are a sought-after risk because of their high engineering and operational standards. Beyond the cover for individual plants there are national and international pooling arrangements for comprehensive cover.
Perhaps the World Nuclear Association has some bias or they're refering to something different than you are. It's hard to evaluate that since you don't include a source, though.
It's pretty disingenuous to link the cost of a detector to its size, as though size indicates scientific merit.
But maybe I'm just bias because I have studied atoms (0.5x10^-10 meter scale) on a graduate student stipend ($20k per year). This cost my university a measly $400,000,000,000,000 per meter per year.
If you mandate all data be immediately made public, the researcher can be "scooped" by anyone. This is bad for science because it removes the incentive to actually gather the data. This is one argument for why data may be kept internal, at least for a while.
According to my professor (who is very involved in the LHC) the first LHC run will be collecting an integrated luminoscity of 1 fb^-1.
Another professor mentioned today that by the end of the Tevatron's life (in a couple years), it will have collected 12 fb^-1. This is over it's 10-ish year life span.
At this point, some may wonder why the LHC is unable to keep pace with the Tevatron, the old toy. These machines are very complicated, and apparently don't work nearly to maximum efficiency out of the box. Check out this plot of the amount of data collected at the tevatron versus year. The slope is rising continuously, as they improve their beam and detectors to handle more collisions: