Comment On Failing the Turing Test

Expert: "X" is the unknown quantity; "spurt" is a drip under pressure.

This give me ideas for candidates interview in the following weeks :D

JohnB:

Expert: "X" is the unknown quantity; "spurt" is a drip under pressure.

No - it's "Ex" as in "Former" or "has-been".
CAPTCHA: Tesla (Signs, signs, everywhere a sign...)

This does not prove the worthlessness of E's papers, if anything it shows that nonsense may be of value as inspiration.

In the sciences it is more common that research is what is called "grounded" and intrinsically empirical, but ground-breaking concepts usually come from what for "normal people" seems like insanity.

That said, most gibberish is just crazy. But do not commit the fallacy of positing a causal (or any kind of) link between E perceiving gibberish as inspiring and E's writings to necessarily be gibberish.

Jeffery should leave a few more papers in his cube. I'd like to know when his Expert gets the clue...

Ozz:

JohnB:

Expert: "X" is the unknown quantity; "spurt" is a drip under pressure.

No - it's "Ex" as in "Former" or "has-been".
CAPTCHA: Tesla (Signs, signs, everywhere a sign...)

<grin> But it fails to account for the "never wuzzers".

I think I need to add a few "white papers" to my resume.

captcha: doom (are you trying to tell me something?)

could there be a subtle psychology going on here, where E trusts his counterpart and does not want to look foolish or even perhaps likes his counterpart and does not wish to make him look foolish? Still, I like the idea in principle, even if in practice it may not prove as telling as it seems on the surface...

captcha: stinky. Yup, I need a shower alright, even the internet thinks I stink...

Just another wanabee:

captcha: stinky. Yup, I need a shower alright, even the internet thinks I stink...

sorry, even the internet stinks i think.

*printing and handing to 'expert' co-worker now*

Similar situation - several years ago, a magazine editor invited poetry submissions from a particular community. "I write sonnets," I said. "Would you consider those?"

"Well, of course," he said. "I'm an EXPERT on poetry. I ESPECIALLY appreciate the obscure and historical forms." So I submitted a few sonnets. Then he sent them back, as editors do, with his suggestions.

1. The poems are too short. Could I make them longer?

2. The meter feels relentless and rigid. Could I try something more lyrical?

3. The rhyme scheme is strange and feels unnatural. Could I make it something more like an ABAB format?

A sonnet, for those who don't know, is exactly fourteen lines long. Every line is in iambic pentameter. The scheme is prescribed under a series of complex rules which dictate not only which lines rhyme, but how the lines relate to one another.

Of course, he's an EXPERT. I just write poetry.

Some of you may be familiar with <a href="http://en.wikipedia.org/wiki/Dissociated_press">Dissociated Press</a>. I pulled out a word-based implementation and ran it across the character descriptions on a horror MUD. The result was a collection of angst-filled gothic gibberish. I grabbed large swaths of it, carefully

arranged
the lines
in such a way
as to seem
like it was
poetry,

and submitted the results.

He was ecstatic. Unfortunately, before he published the issue in which they were to appear, he came across me on a forum talking about what an idiot he was and how I totally bamboozled him with computer-generated garbage.

Mikademus:

This does not prove the worthlessness of E's papers, if anything it shows that nonsense may be of value as inspiration.

Yeah it just proves the worthlessness of E.

Why would cryptographers comment on Moore's Law to begin with? I can't even figure out what the hell the location-identity split is, but it appears to involve middleware (i.e. nothing to do with cryptography as a study and even less to do with the price of transistors).

The paper itself is brilliant. It looks complicated, the figures themselves are impossible to understand (the block size (celsius) is perfect!), and the references are... so fun that I don't even understand how it can't be read without rofling all the day (c'mon! C. Darwin wrote "Operating Systems considered harmfull" together with Sutherland! and what about "Emulating red-black trees and Ipv7"?).

Having a PHd level guy who can't spot this as a fake is quite disturbing but hey, it's so brilliant.

I save it for future use. Thanks everyone :)

CDarklock:

Similar situation - several years ago, a magazine editor invited poetry submissions from a particular community. "I write sonnets," I said. "Would you consider those?"

"Well, of course," he said. "I'm an EXPERT on poetry. I ESPECIALLY appreciate the obscure and historical forms." So I submitted a few sonnets. Then he sent them back, as editors do, with his suggestions.

1. The poems are too short. Could I make them longer?

2. The meter feels relentless and rigid. Could I try something more lyrical?

3. The rhyme scheme is strange and feels unnatural. Could I make it something more like an ABAB format?

A sonnet, for those who don't know, is exactly fourteen lines long. Every line is in iambic pentameter. The scheme is prescribed under a series of complex rules which dictate not only which lines rhyme, but how the lines relate to one another.

Of course, he's an EXPERT. I just write poetry.

Some of you may be familiar with <a href="http://en.wikipedia.org/wiki/Dissociated_press">Dissociated Press</a>. I pulled out a word-based implementation and ran it across the character descriptions on a horror MUD. The result was a collection of angst-filled gothic gibberish. I grabbed large swaths of it, carefully

arranged
the lines
in such a way
as to seem
like it was
poetry,

and submitted the results.

He was ecstatic. Unfortunately, before he published the issue in which they were to appear, he came across me on a forum talking about what an idiot he was and how I totally bamboozled him with computer-generated garbage.

Any repercussions from getting caught?

And so, for the litmus test, anyone care to conjecture where this 'expert' got his ph.d in computer science(?) from? I'd guess "not Stanford". Unless it's stanford-degree-for-a-dollar.com.

this reminds me of my boss

PhD in DB management, MS in Computer engineering

and yet constantly trying to give us advice on how to DESIGN our websites.

I almost wish they knew what they were talking about so I could get some actual insight when I run into roadblocks in my apps. oh well. :)
Reminds me a lot of the MBA mentality -- some kind of undeserved self-importance.

Jeffrey didn't have the heart to tell Ebenezer that his ideas are precisely as insightful as those randomly generated by a computer

Did he have the heart to tell Ebenezer's boss, or the person who signs Ebenezer's paychecks? Or maybe he might want to slip this information into the inbox of the person Ebenezer was so rude to...

There is no WTF here.

Marketing and Sales people use similar products to generate the language they speak at each other. To the rest of us, it sounds like a load of BS.

So, why can't we?

Just another wanabee:

could there be a subtle psychology going on here, where E trusts his counterpart and does not want to look foolish or even perhaps likes his counterpart and does not wish to make him look foolish?

Maybe in some cases, but in this instance, we already know that Ebenezer will scream things like "All of these people are so unprofessional!" when faced with a reasonable request. So it doesn't sound like he's really big on that whole "courtesy" thing.

http://isotropic.org/papers/chicken.pdf

So did E make sure to document the process he used to convert millibars to teaspoons?

There's a vanishingly small but non-zero chance that a generated paper might actually make sense. Maybe he just got really lucky.

I thought the Turing test was to see if a computer could successfully fool a human into thinking it was another human. So, shouldn't this be Passing the Turing Test?

Let us not forget Sokal, the grand-mack-daddy of smacking down self-important pseudo-cognesanti academics with a hoax paper full of gibberish:

http://en.wikipedia.org/wiki/Sokal_Affair

Maybe he's on E!

Seriously, this expert worked on the paper for a half hour, and highlighted notes on the page. The article points this out. He's dumb to have wasted the time, even if he knew better. So, it's hard to believe he realized it wasn't a real research paper.

Jeff:

http://isotropic.org/papers/chicken.pdf

This reminds me of the Parking Lot Is Full strip, "chicken"

http://plif.andkon.com/archive/wc072.gif

snoofle:

Any repercussions from getting caught?

For some reason, he didn't publish the poems. I can't imagine why.

It's not like he could do anything. It was still my poetry.

Andrew:

Seriously, this expert worked on the paper for a half hour, and highlighted notes on the page. The article points this out. He's dumb to have wasted the time, even if he knew better. So, it's hard to believe he realized it wasn't a real research paper.

Dude, you missed the obvious. That's where the big WTF is: an expert, who, yes, DIDN'T realize he was reading gibberish.

Duh. And just so that you know, I wholly believe this story. I've seen such "experts".

Ah, the wonders of computer-generated babble. Useful for so many things, it appears!

The pig in what? Ketchup. No quack.

Poor guy, Maybe he shouldn't be such an asshat all the time.

joe.edwards@imaginuity.com:

I thought the Turing test was to see if a computer could successfully fool a human into thinking it was another human. So, shouldn't this be Passing the Turing Test?

No, for two reasons. First, the turing test assumes a "judge" with reasonable intelligence. E doesn't qualify. More importantly, the turing test is a conversation. It's relatively easy to produce something written that can pass for human creation. It's MUCH more difficult to carry on a conversation.

Mikademus:

...That said, most gibberish is just crazy. But do not posit the writing of commiting an inspiring, or any kind of, link between E perceiving gibberish as causal and E's gibberish to necessarily be fallacies.

It makes almost as much sense either way.

-Harrow.

I once wrote a program which generated random marketing gibberish, using a lot of buzzwords.

Too bad that I lost its source when I did a hard drive cleanup. But it shouldn't be that hard to rewrite.

Jojo:

this reminds me of my boss

PhD in DB management, MS in Computer engineering

and yet constantly trying to give us advice on how to DESIGN our websites.

I almost wish they knew what they were talking about so I could get some actual insight when I run into roadblocks in my apps. oh well. :)
Reminds me a lot of the MBA mentality -- some kind of undeserved self-importance.

Yeah. I have a PhD hanging on my wall. It is from The Internet Institute of Technology.
My final thesis was in two parts:
a. Full name
b. Credit card number

I got them both right!!!


*the joke is in honor of a dear departed friend, Wilson Rogers

H3SO5:

I once wrote a program which generated random marketing gibberish, using a lot of buzzwords.

Too bad that I lost its source when I did a hard drive cleanup. But it shouldn't be that hard to rewrite.

Anything like this one?

ParkinT:

H3SO5:

I once wrote a program which generated random marketing gibberish, using a lot of buzzwords.

Too bad that I lost its source when I did a hard drive cleanup. But it shouldn't be that hard to rewrite.

Anything like this one?

Yes, the concept is similar.

Mikademus:

...do not commit the fallacy of positing a causal (or any kind of) link between E perceiving gibberish as inspiring and E's writings to necessarily be gibberish.

E. didn't "perceiv[e] gibberish as inspiring". He said that the gibberish was "...the natural progression of all the ideas I presented in my incredibly insightful PhD thesis!" -- which, if anything, is a suggestion that the gibberish was inspired by him.

You really ought to think an idea through yourself before you start advising people about how to interpret it.

Those SCIgen articles are really like the texts my groupmates write for an report that we have to write at my university...

The only difference is that the SCIgen articles are grammatically and syntactically correct.

CAPTCHA: pinball (if I only had time for that...)

I, for one, am curious about the legitimacy of E's "PhD" credentials to begin with. Perhaps I'm simply reading too much into things, but his reference to his "PhD thesis" put up a big red flag for me. Masters degrees require a thesis. PhD degrees require a dissertation.

I bookmarked the SCIGen page for when I go for my masters degree.

Should be a breeze.

The references are a bit repetitive, though:

- "Flip-flop gates considered harmful",
- "Operating systems considered harmful",
- "Reinforcement learning considered harmful",
- "Object-oriented languages considered harmful".

GoogleFight says:
"doctoral thesis" - 1.15M
"doctoral dissertation" - 1.09M

Doesn't seem like either is more prevalent.

Saladin:

Ah, the wonders of computer-generated babble. Useful for so many things, it appears!

The pig in what? Ketchup. No quack.

interface show even nurse positive jam

NO QUACK.

Random:

Saladin:

Ah, the wonders of computer-generated babble. Useful for so many things, it appears!

The pig in what? Ketchup. No quack.

interface show even nurse positive jam

Nurse positive jam? We have to many nurses and they are jamming up the system? I'll take a few thank you!

snoofle:

CDarklock:

Similar situation - several years ago, a magazine editor invited poetry submissions from a particular community. "I write sonnets," I said. "Would you consider those?"

"Well, of course," he said. "I'm an EXPERT on poetry. I ESPECIALLY appreciate the obscure and historical forms." So I submitted a few sonnets. Then he sent them back, as editors do, with his suggestions.

1. The poems are too short. Could I make them longer?

2. The meter feels relentless and rigid. Could I try something more lyrical?

3. The rhyme scheme is strange and feels unnatural. Could I make it something more like an ABAB format?

A sonnet, for those who don't know, is exactly fourteen lines long. Every line is in iambic pentameter. The scheme is prescribed under a series of complex rules which dictate not only which lines rhyme, but how the lines relate to one another.

Of course, he's an EXPERT. I just write poetry.

Some of you may be familiar with <a href="http://en.wikipedia.org/wiki/Dissociated_press">Dissociated Press</a>. I pulled out a word-based implementation and ran it across the character descriptions on a horror MUD. The result was a collection of angst-filled gothic gibberish. I grabbed large swaths of it, carefully

arranged
the lines
in such a way
as to seem
like it was
poetry,

and submitted the results.

He was ecstatic. Unfortunately, before he published the issue in which they were to appear, he came across me on a forum talking about what an idiot he was and how I totally bamboozled him with computer-generated garbage.

Any repercussions from getting caught?

Yeah, his license to write poetry was revoked by the American Union of Poetry Writers, Local 345

SCIgen generated thesis:

Figure 2: These results were obtained by Johnson and Miller [33]; we reproduce them here for clarity. Of course, this is not always the case.

So true.

Wow I really wish I had had this utility when I was in grad school!

You know, I have absolutely *no* idea how to reply to the original post. An invisible button, maybe? Alas and alack, I am never going to be "Frist..."

I would, however, like to defend academics, particularly in the field of CompSci, because my father was one from before I was born ... and a damn good one. Taught well, published a quarter of a halfway decent book, and didn't put up with crap.

The saying that "those who can, do; those who can't, teach" was always meant as a sardonic quip. Sadly, with the fax-machine nature of today's PhDs, it has now been outmoded. These days, apparently, "those who think they can do are WTF; those who don't WTF, teach; those who know they can't do, but don't even realise that they are a WTF, go back from academia and piss people around."

And let's not even think about PHBs.

Daedalus:

I, for one, am curious about the legitimacy of E's "PhD" credentials to begin with. Perhaps I'm simply reading too much into things, but his reference to his "PhD thesis" put up a big red flag for me. Masters degrees require a thesis. PhD degrees require a dissertation.

Not all colleges/universities are equal. There are people who have obtained a Masters degree without a thesis, and I know at least 1 who did not have to do a thesis/dissertation to get his PhD.