How to Call B.S. On Big Data: A Practical Guide | Hacker News Reader

How to Call B.S. On Big Data: A Practical Guide

newyorker.com

140 pointssndean24 comments

I love this one:

'Mind the Bullshit Asymmetry Principle, articulated by the Italian software developer Alberto Brandolini in 2013: the amount of energy needed to refute bullshit is an order of magnitude bigger than that needed to produce it. Or, as Jonathan Swift put it in 1710, “Falsehood flies, and truth comes limping after it.”'

It summarizes my workplace experiences (yeah, confirmation bias). I've called bullshit on engineers a few times, but managers seem not to care. I gave up doing it, and instead whenever a engineer makes bullshit claims about a particular process/software/language I ask him "can you show us a working example?" Give them enough rope...

Great article! But when I saw the New Yorker Url my first reation was - let's see how they are going to make this into something negative about Trump this time. They still live up to my expectations.

Oh course, it's en vogue

You mean "relevant" I think.

No I meant, "en vogue". Meaning "fashionable".

"fashionable" is not what call something that's necessary.

That's "in vogue", isn't it?

It was one joke said in passing. And it was directly relevant to the content of the sentence it was in.

I don't think your implied critique holds much water here. This isn't an academic paper anyway... it's an expository article written in the amusing/playful tone for which the publication is known.

Not sure if this was intended as a joke, or another one of those pedantic remarks. If they are serious about that argument they should start telling their advertisers not the total number of Unique Visitors, but the Unique Visitors in relation to the total number of Internet Users. I'd love to see how successful they became over the last 10 years...

That's silly, because every platform then has the same denominator of 'total internet users,' and you're dividing by the same number.

It's not silly because the number changes over time. It could make the difference between an upward trend and a downward trend. (Whether that difference matters is up to you, but it is a difference.)

But everybody faces the same trend in 'total worldwide internet users.' So if an advertiser has its choice of platforms to advertise on, it can compare the same metric, regardless of whether or not you do the division.

Did you really think it was great? The subject matter is critically important, but the article sadly lacking in substance. These are three of the bullet points, "Recognize that bullshitters are different from liars," "Watch out for unfair comparisons," "Remember that correlation doesn’t imply causation" - ok, so are they now passing off middle school science as college learning? This doesn't add much ammo to fight misuse of big data.

agree, the article offers almost no substance.

The title of the course would be just as accurate if they omitted the "on Big Data" part.

It's an applied philosophy course, really, about how to spot misleading claims.

Yeah, I am not sure what's "Big Data" about any of this.

That's very common in Big Data™

The first rule of Big Data is that you probably don't have Big Data.

Instead try to realize the truth. There is no data. Then you may realize it is not the data that is big... it is your head.

Or the portfolio of buzzword targets your investors/MBA frat boys/marketing vets want you to hit.

Well Google Flu Trends was debunked in the article. I guess they are just reiterating and updating some well-loved heuristics to also include some contemporary techniques / fallacies.

My guess about VCs and big data:

My guess is a special case of my guess about VCs and X for any new technology or tech buzz word X, that is, just substitute "big data" for X.

The guess is a special case of the difference between the sizzle and the steak. Or, remember, sell the sizzle, not the steak.

Well, VCs have Web sites and there commonly emphasize the X, the sizzle, they are interested in.

In fact, for a VC, X or sizzle with a dime won't cover even a 10 cent cup of coffee. And a VC won't invest even 10 cents in X or sizzle.

So, why do VCs emphasize their interest in X they don't care about?

Well, VCs want deal flow, that is, entrepreneurs presenting their projects so claim they are interested in current sizzle X. Then entrepreneurs with a project in subject X will eagerly contact the VC.

Now what? Sure, the VC ignores the X, the sizzle, and looks for some steak, usually traction significant and growing rapidly, better still, revenue significant and growing rapidly. They also look at the founders, etc. The one thing they do NOT look at, understand, respect, evaluate, or care about is X or the sizzle.

None of this effort means anything for a VC unless they can see enough steak to start negotiations with the entrepreneurs. Now we see why the VCs said they were interested in X: During the negotiations, they will pretend to evaluate the business based on X, not on the steak. Then the VCs will denigrate, diss, minimize, insult, run down, criticize X as of not very solid business value and lower their evaluation (pre-money) of the business. And THAT'S where the VCs get their payoff for their claims of interest in X.

So, selling/buying the sizzle instead of the steak hides the real value, the steak, and the VCs want to do that.

There must be some explanation for the very common practice of VCs saying they are interested in X. IMHO, no way are the VCs interested in X, big data, AI, ML, etc. -- those topics are just sizzle to hide their interest in what they regard as steak.

If can, then check out my guess: Say, get a VC to explain any of the value of big data, AI, or ML in a reasonably solid and well informed way! Uh, "X is eating the world" is NOT good enough!

There must be an explanation, and my best guess is above!

YMMV!!

Size doesn't matter... :-)

Reply on news.ycombinator.com