The quackery-meter. Which Medium posts are not a waste of time?

Introducing the quackery-meter tool and a brand new quackery unit of measure: the Elon.

A Grumpy Computer Scientist
4 min readSep 14, 2016
The miraculous elixir of Dr. Dulcamara : It moves the paralytics, despatches the apoplectics, the asthmatics, the asphyxtics, hysteria, diabetes, it cures tympanitis, tuberculosis and rickets, and even liver trouble,
which used to be fashionable.

Lately, I have become increasingly frustrated by the amount of tosh the Medium recommender system thinks I should read. The suggestions seem to contain more of what Medium would like me to read, than of what I myself prefer. I need something to improve those daily digests, and I think I can do it with as little as some statistical text analysis, with perhaps a bit of simple ML.

I am thinking of developing a little AI application that distills the vocabulary used in an article, and tries to infer if it is going to be a well balanced and fact-based read, or a load of hot air, wrapped in a sales pitch.

I am not sure how well it would work but I suspect it would at least be able to tell Silicon Valley entrepreneurs from everyone else.

The simple starting hypothesis is that the more an article is filled with buzzwords, superlatives, unrealistically positive, feel-good, business-jargon, we-are-so-great, success-is-all-we-know, jargon, the less worthwhile a read it is going to be.

In case it is ludicrously over the top, then it’s probably by some Silicon Valley entrepreneur or salesperson (see 2021 edit*) — which are all too often impossible to tell apart and even more impossible to read with a straight face — and most certainly a waste of your time on someone’s badly-concealed attempt to sell you something.

As a proof of concept, I distilled a recent article manually. Here is the result. Make an effort to push through it. Suppress your dry-heaves, and judge for yourself:

incredible possibilities, truly unique, paths to greatness, euphoria, privilege, journey, inspiring, excited, massive vision, commitment, relentless, being great every day, achieve their dreams, unprecedented increase, breakthroughs, incredible opportunity, transformation, insights, incredible team, massive opportunity, breakthroughs, reveal insights, well-designed experiences, dramatically improving, amazing team, dramatically accelerated, growth, exhilarating, incredible highs, results, remarkable ride, fortunate, incredible, pushed us hard, constantly blown away, tough love, challenged, outcomes, incredible lessons, leadership, fastest growing, we owned our destiny, very lucky, open mind, unique capabilities, valuable insights, passionate about sharing, achieving great things, daring to be great, audacious goal, achieve their dreams.

By contrast, you will hardly ever find a concentration of hyperboles like the one above in articles written by considered authors on respectable publications. Try scouring the Guardian’s or New York Times’ Tech sections or the MIT Technology Review for corroboration. You will find even less of that language in reputable academic journals and magazines.

It looks like it should not be too difficult to extract a meaningful statistical score, based on this little experiment. And even feed a bunch of similar examples into a machine learning algorithm, and train it to recognize the obvious pattern.

I am not trying to be unfair to the article I used as a test. I just randomly chose something that looked the part at first glance. But, admittedly, this turned out to be an easy one. Once something contains the words “daring to be great”, after all, it can only really go one way.

I appreciate the irony that the present post would never pass the test, as it contains that very snake oil language in an unusually high concentration, albeit having the exact opposite spirit. The tool would not need to be perfect to be useful, though. Especially because its main purpose would be to make a point: that some people heard enough rainbow drivel for one life already, and please can we cut the inspirational stuff, the superlatives, and get factual.

*2021 Edit: Naturally, such a useful measure deserves its own unit of measure, the International Standard unit of measure of bullshit. In the tradition of great existing units like the Volt, the Watt, the Joule, the Farad, etc… I have decided that the unit of measure of deluded quackery should be baptized the Elon.
Of course everyday articles, speeches, tweets and utterances will be measured in milliElons, as to reach one Elon would take an
extraordinary amount of bullshit.

And no, the fact that someone somehow became rich and successful does not imply that they have something intelligent to say or any particular talent. If you are unsure about this, you haven’t been following the news closely enough, especially from 2016 onwards.

What do you think? Do you know of any existing tools already doing something similar? Would you find a tool like this useful?



A Grumpy Computer Scientist

UK-based AI professor interested in AI, mind, science, rationality, digital culture and innovation. Hobbies: incessantly fighting nonsense.