How a Word Becomes a Buzzword
The crypto boom created a lot of hype around Bitcoin and the technologies that it uses. The most notable of these is the blockchain, an inventive solution to create a public, secure, and immutable ledger under the right conditions. Unfortunately, the excitement came with the same negatives that it always does with a field in computer science. People overestimate its capabilities, put it where it doesn’t belong, and promote its use before they’ve even learned about it. Many different fields have gotten this treatment; this post goes into a few examples, and highlights some things that blockchain technology could actually be used for.
Many topics popularized through computer science experience the hype cycle, and I’m going to look at just a few of them—namely big data and artificial intelligence.
I remember when everyone that was anyone was talking about “big data” as if it were the solution to all of their company’s problems. It seemed so magical: if we had enough data, if we really knew everything, then of course we could solve any problem! But by just taking a look at the trends for the term, we can see that it likely hasn’t solved all of our problems, and our focus is getting to be elsewhere:
But why are people starting to move away from this buzzword? And why did were they drawn to it in the first place? Big data as it’s popularized is certainly an appealing concept. One reason that so many companies picked it up is because its applications were supposedly endless; no matter the field you were in, it would solve your problems. The other driving factor I mentioned above: the magical quality of it. Few people seemed to really understand what was behind “big data” and only knew it as a buzzword. Rarely did people actually think about the process of mining data, storing it, or manipulating it into something that was usable. Instead, it was treated as something that we needed to collect in a very vague way, and something that if we had it would make the company better as a whole.
This one’s a doozie—and it’s still going strong. I’m going to focus on the subset of AI that is machine learning, which is a process of classifying or predicting things based on statistical methods. Essentially, a machine learning program is profiling objects based on certain qualities. Given the negative connotation of profiling in popular media, you may already be able to see where this is going. There have been frequent cases of machine learning algorithms profiling by race, notably in cases where beauty is evaluated and criminal track records are predicted. In the cases of predicting who will commit crimes again, the algorithm has often been proven to be wrong.
My personal guess is that the programs are biased because, at least for these cases, they rely on biased data from humans. Machine learning is completely reliant on the data points it gets (as is pretty much anything that learns), and so needs to have good data to make good predictions. Not only does it need good data, it usually needs a lot of it, at least relative to the bad data. And it’s not always feasible to get that in every field.
AI became a buzzword in a very similar way to big data: it seems magical and applicable to everything. You can see unnecessary uses of machine learning in a lot of these fields that can’t acquire the necessary data because people assume machines will be able to solve our problems. And plenty of people promote this idea because they see machine learning techniques solving other people’s problems, without understanding the meat of the solution.
Blockchain suffers from the same qualities as artificial intelligence and big data: while most people don’t know what it’ll do, they still think it’ll solve their problems. I’m going to look at the pros and cons of a few potential applications of blockchain technology (and let xkcd take care of describing a potentially harmful one…)
Medical facilities often suffer from the arduous process of both maintaining their patient’s privacy while sharing their patient’s information with those that need it to care for that same patient. Often, clinics will have solely paper records that they fax to other clinics when the patient requests it. Though the process of digitizing records is certainly growing and getting significantly easier, even then the records can be difficult to share through some secure server.
Blockchain offers a potential solution: when the patient’s information should be public to the patient, that information could (after being encrypted) be put on a public blockchain. Then the patient could choose who can decrypt this information, and once they do so it can immediately available to this person. This way, clinic should, in theory, be able to immediately access information they need about a patient and don’t already have without any hassle of going through another clinic. You can read more about this solution in the description of an implementation below, MedRec:
This sounds great, but there are certainly some caveats. The biggest one seems to be that the public blockchain is just that: public. While is more or less necessary in some respects, as it means the blockchain can be provably immutable in some scenarios, it also means anyone can see what’s on it—and if they decide they want to track an address, they can. While they might not be able to find out the identification of an address, they can find all medical operations that have happened to that address, linking them together over the course of the blockchain.
The UPS announced a while ago that they were attempting to patent a blockchain based solution for their delivery methods. In doing so, they’ll be joining the Blockchain in Truck Alliance, in a hope to get packages delivered in a more timely and trustworthy manner. The essential idea behind using blockchain in this application is that transactions are able to be recorded and verified through the blockchain in the same way they would be with a cryptocurrency, so it would be easy to show when a package had been ordered, shipped out, and arrived to someone’s doorstep.
But clearly the excitement for this idea, and blockchain technology in general, has overtaken the actual idea. Just in researching for this article, I came across this news site that said:
No where did the author mention that this could be true, and wasn’t always necessarily true. This approach to delivery services suffers from many of the same pitfalls as the health care solution does: someone could easily track which addresses receive which packages, even if they can’t identify the address.
Many companies, such as UPS, have successful stories of using blockchain technologies in their systems. It’s lowered costs, made processes faster, etc. But I’d be very skeptical about these success stories in general: these companies, as far as I’ve seen, never have any real baseline. When they have a success story, it could easily be because they completely revamped their system to work digitally instead of physically, or redesigned their already existing system already to use newer technologies or frameworks. There’s little reason to assume that these successes are due to blockchain—and though it might be useful for many things in the future, it’s still hard to see a fit outside of the currency space right now.
How a Word Becomes a Buzzword
Research & References of How a Word Becomes a Buzzword|A&C Accounting And Tax Services
Source
0 Comments