Data Science = Net Evil

>>60702261
I mean it's certainly possible to do things like conduct novel medical research by finding patterns in vast sets of epidemiological data. And no tool is inherently evil. Guns don't kill people, people kill people, and all that. But you're right that most of the obvious applications of it are malicious. I'd put facial recognition in the same category.

>>60702334
That's why I said "net evil" as opposed to evil. Net evil implies there's some good uses but they are overshadowed by the bad.

I remember one anon feeling completely disillusioned with data science after he studied statistics and went on to work for a health insurance firm where he used his expertise to write complicated algorithms to find all the unprofitable patients with health histories that match those of the terminally ill and simply cut them off because they're not profitable to to keep alive.

>>60702475
That's brilliant

>>60702475
I kinda wonder how he expected anything different when he agreed to work at an insurance company.

>>60702261
entirely depends on on where you work.

if you are a data scientist in the military, you figure out how to kill the most people with the least amount of bombs.

if you work at a bank, like me, you figure out how to either save money (what I do), or make money (what the evil ones do).

if you work for a trucking company you figure out how to safe fuel, a cleaning company how to save soap. etc etc

it's entirely determined by the business you are in. the skills themselves are portable to any job you want.

>>60702475
Probably wrongskinned, too. Only men of that complexion can be so unconscious. Delusion after the fact is one thing. But it takes a real man to stand for what he believes in the moment.

this is why i work in academia

i get to do cool machine learning shit that is basically objectively pointless

File: Evil-Scientist.png (197KB, 443x600px) Image search: [Google]

197KB, 443x600px

>>60702980

>>60702980

Like create chat bots to flirt with people and market to them until /pol/ shitposters turn it into a proud white nationalist?

>>60702261
I'm currently working as a lab technician and play around with genomics data using R. Any work I do is probably more likely to be used for the genetic engineering of the human species than for curing cancer, though. But the pay is good and the commute is pleasant.

>>60703325

So.....yeah, evil. Any plans to use your skills for something that doesn't involve creating designer babies?

File: 2fyYFgE.jpg (249KB, 566x866px) Image search: [Google]

249KB, 566x866px

>>60702261
I've been wondering the same. Sadly, it all indicates it is a fucking weapon of totalitarian scale.

>>60703660
The methodology I've been told to use is actually horribly flawed, so I'm noy adavncing human knowledge in any way, for good or for ill. PCA is not an effective methods of identifying drug targets, because you just end up identifying biomarkers that have little regulatory function (the expression level of a gene of any real importance obviously can't change too much without the organism dying). Everyone I'm working with knows that we should use other statistical methods, but no one does. Also, the sample collection method is highly imprecise, so the data we start with should already be treated as compromised. Academic science as a whole has degenerated to the point where it's mostly just rent seeking.

>>60703314
i dunno man, i let the indians do all the work, then proofread their broken english and add my name to the author list

>>60703781

Not that I'm encouraging you, but any reason you don't tell your employer the methodology you use is wrong?