Oops, 7 Days. Hey look I don't update on weekends.

A google query for "Itchy Anus Smelly Wind" leads to Chrissy

http://www.mindpixel.com/chris/2005/08/real-human-thoughts-in-realtime-peek.html

I have to say that I'm not terribly suprised by this revelation.
Permalink muppet 
August 23rd, 2005
It'll lead to my boss too if the world is right.
Permalink Cool Dad Needs To Cool Off 
August 23rd, 2005
Would it have been too much to have given an appropriate **Photo Warning**? Another image shattered <sigh>.
Permalink PNII 
August 23rd, 2005
Chris Mckinstry and

saudi womens pussy

hand in hand huh?

this is what people are thinking when they go to Chris's site, let's hope it wasn't a dissappointment. I wonder what the semantic-neural dimensionality of my fat ass is, probably 3.14159.
Permalink Cool Dad Needs To Cool Off 
August 23rd, 2005
from-cat-brains.html

haha, hell yeah.
Permalink Cool Dad Needs To Cool Off 
August 23rd, 2005
"european swallow facts"

Hmm...
Permalink Mat Hall 
August 23rd, 2005
Chris McKinstry should call himself 'Zoltan' and set himself up in a shack in the woods where he can monitor his referer logs using a satellite phone, masturbating furiously.

That's my somewhat offensive take on the situation.
Permalink Cool Dad Needs To Cool Off 
August 23rd, 2005
From his previous entry: "I woke up early and had a light breakfast and a shower and did a couple of other very pleasurable things..." Fnarr fnarr, etc.

I guess one advantage of having pages filled entirely with random crap is that pretty much any search query is going to match something. Exhibit A:

http://www.google.com/search?hl=en&lr=&safe=off&c2coff=1&q=carbon+dioxide+giant+testicles+pizza&btnG=Search

If you follow the link you'll note than only 93% of respondents can confirm having seen the moon. I leave you to draw your own conclusions...

(On an unrelated topic this weed is a lot stronger than I was anticipating, so I'm going to bed before I pass out and wake up lying face first on a drool-soaked keyboard.)
Permalink Mat Hall 
August 23rd, 2005
Does "soles of gas" lead one to the right person? Google is broken!
Permalink foo bar 
August 23rd, 2005
Quick! how many folks can post before Chris notices it on his logs and comes over here to urinate his 7-dimensional trajectory of urine?
Permalink Peter 
August 24th, 2005
>Quick! how many folks can post before Chris notices it on his logs and comes over here to urinate his 7-dimensional trajectory of urine?

I have no idea who this Chris person is, but the idea of getting his bridges burnt over this seems pretty cool. I am on it with you.
Permalink foo bar 
August 24th, 2005
There are some hilarious ones in there...

how long does the average human spend in the bathroom
european swallow facts -- no doubt, contrasted with the African swallow :))
Permalink some guy 
August 24th, 2005
I noticed pretty quickly.
Permalink Chris McKinstry 
August 24th, 2005
Some guy:

You would be laughing your head off if you could see what I can pull from my logs, some people think some really funny thoughts. And because I am modeling funny, I can pull them out automatically.

Other's might find it something other than funny that I can automatically generate pretty good models of the mental and emotional state of large numbers of people in real time and relate that information to an ip address and in turn relate the ip address to a specific geographic location.

It is not hard to imagine a future mental/emotional treatboard somewhere in the bowels of Fort Bragg.
Permalink Chris McKinstry 
August 24th, 2005
Mmm. only 13 posts. could we do better in the future?
Permalink Peter 
August 24th, 2005
Cookie to the first user to push the output from the Dada Engine into the input stream.

Double recursive cookie to the first user to feed the Dada Engine in question its grammar from the mindpix*l results page.
Permalink Snark 
August 24th, 2005
"Other's might find it something other than funny that I can automatically generate pretty good models of the mental and emotional state of large numbers of people in real time and relate that information to an ip address and in turn relate the ip address to a specific geographic location."

And others still (no apostrophe, dude) will find it unlikely that you can do any such thing. Here's a challenge for you -- over the next 24 hours automatically compile a list of all the queries that come from people who are currently mildly annoyed, and send that and the complete list of queries to me when you're done. I don't believe such a thing is remotely possible, and won't be for many many years (my guess is that the strong AI problem won't be solved within my lifetime) -- if you truly have an AI that can deduce emotional state from google queries then you really should be getting a Nobel prize...
Permalink Mat Hall 
August 24th, 2005
No doubt he's just about to get a Nobel prize, it just hasn't been announced publicly yet.
Permalink  
August 24th, 2005
Yeah those Nobel guys are all over his shit.
Permalink muppet 
August 24th, 2005
Why, this stuff is DYNAMITE I tell ya!
Permalink some guy 
August 24th, 2005
Matt, what do you think my seven dimensions are?

They are the invariants across complex brain states. Space, time and affect [emotion] 1+3+3=7. You could call it my Wunt-Einstein invariance estimate. I have an estimate of the location in seven dimensions of nearly two million propositions.

Spatial models of emotion might be new to you, but they are very old. In fact, the oldest in psychology. Read up on Wunt if you want to learn about them.

Below you can find some examples of words in a three-d model of affect - but remember, I have two million propositions, not just words, each with a three-d affective estimate, in addition to the space/time estimate.

The three affective dimensions I am using are: pleasure-displeasure, arousal-nonarousal, and dominance-submissiveness. "Pleasure-displeasure" distinguishes the positive-negative affective quality of emotional states, "arousal-nonarousal" refers to a combination of physical activity and mental alertness, and "dominance-submissiveness" is defined in terms of control versus lack of control.

angry (-.51, .59, .25)
bored (-.65, -.62, -.33)
curious (.22, .62, -.01)
dignified (.55, .22, .61)
elated (.50, .42, .23)
hungry (-.44, .14, -.21)
inhibited (-.54, -.04, -.41)
loved (.87, .54, -.18)
puzzled (-.41, .48, -.33)
sleepy (.20, -.70, -.44)
unconcerned (-.13, -.41, .08)
violent (-.50, .62, .38)

How I estimated the seven dimensions for each proposition is something I am most certainly not going to disclose while I have patents in the works because I have to be careful with publication dates. But, it is not difficult to think of how I could do this. And once you figure out that, you'll figure out how I flagged the gas station bombing query as an extremely hostile query...and how to make a realtime map of the emotional state of the planet.
Permalink Chris McKinstry 
August 24th, 2005
Shields up! Red alert! We need more power, Mr. LaForge!
Permalink muppet 
August 24th, 2005
Surely you can only get an emotional reading from an emotional query, though? If you get the referrer string "calculating internal angles of an n sided regular polygon" or "correct spark plug gap 1976 Chevy Nova" how does that allow you to know anything at all about the emotional state of the person asking?
Permalink Mat Hall 
August 24th, 2005
Every query has emotional content. The queries that you think have no emotional content actually do. It is just that all the dimensions on those queries are close to zero.

In fact, I can estimate the affective coordinates for any text string because I have a massive corpus where every word occurs in many contexts.

You would laugh if I told you how easy it was to estimate the location of every proposition I have. It really is a joke. Ok. No more hints.
Permalink Chris McKinstry 
August 24th, 2005
Fire phasers!
Permalink muppet 
August 24th, 2005
Ok, so let's say I'm trying to fix my 1976 Chevy Nova, and I'm getting really angry -- I keep scraping my knuckles, I'm covered in engine oil, and the damned thing isn't firing properly. I query "correct spark plug gap 1976 Chevy Nova". You claim you can tell I'm angry? Ok, let's assume for the moment you can...

Now imagine I'm so angry that I give up and call my friend to come and do it. He's also not familiar with the car, but he's a calm and patient professional mechanic, and doesn't get angry about the problem. He searches for "correct spark plug gap 1976 Chevy Nova". You can tell that *his* emotional state was calm and collected, even though it's an identical query?

If that *is* your claim then you're obviously either a total fool or a fraud. If that *isn't* your claim then you've pretty much negated your entire body of work.
Permalink Mat Hall 
August 24th, 2005
We can run an inverted tachyon pulse through the main deflector!
Permalink muppet 
August 24th, 2005
Ah, yes. An inverted tachyon emission *might* just destabilise the sub-space field enough to allow the dilithium matrix to reform. Mr LaForge, make it so!
Permalink Mat Hall 
August 24th, 2005
I can place any text string in 3d affective space, is my only claim.

Now granted that, if the strings I chose to place in this space are queries hitting a search engine, I can estimte the emotional state of the person who wrote the query, when they wrote the query. If I collect a large number of queries from a person, I can get an estimate of how the person's emotions vary over time.
Permalink Chris McKinstry 
August 24th, 2005
I'm sensing strong feelings of utter nonsense from you, Chris. Maybe I should advise the Captain.
Permalink muppet 
August 24th, 2005
Invoke Starfleet order 104, section C.

Man, I'm a nerd... :)
Permalink Mat Hall 
August 24th, 2005
"I can automatically generate pretty good models of the mental and emotional state of large numbers of people in real time and relate that information to an ip address and in turn relate the ip address to a specific geographic location."

"I can place any text string in 3d affective space, is my only claim."

So you mean that the first claim isn't actually a claim?

It's like shooting fish in a barrel...
Permalink Mat Hall 
August 24th, 2005
What you quoted is an extrapolation of my base claim, it too a claim. I am simply trying to be clear as to what I can do.

Just imagine a map of the world onto which is projected every google query [I don't have access to every google query, but I get a pretty good and wide slice of them] each as a pixel. The location of the pixel can be estimated from the ip address and its colour can be estimated from my corpus. And because I am tracking queries by ip address over time, I can model large numbers of people over time. Of course there is a problem in associating ip address to a person, but that's another issue.

I can even do forcasting of future global affect maps from past global affect maps.
Permalink Chris McKinstry 
August 24th, 2005
"...its colour can be estimated from my corpus"

And that's the bit I have trouble accepting. Any given phrase can, depending on context, have a range of meanings and emotional loadings, and without knowing anything about the specific individual and the context in which they're making the query you can't really make any sort of value judgement. Does someone searching for "I hate cheese" hate cheese, or are they trying to find out about people who hate cheese?

Now, to my view of your mindpixels, which I've been saving for a rainy day. They have a "yessness" value between 1 (everyone agrees) and 0 (no-one agrees). The way people are supposed to respond is "how do you think most people would answer", and as you're trying to get a picture of a notional "average" person presumably you believe that anyone who answers the questions will have a similar idea of how "most people" would respond.

If this were the case all your results should be clustered around the two extremes; anthying between around 0.15 to 0.85 is too uncertain to be of any value. As far as I can tell, based on the 5% of results publicly available, the results are actually spread fairly evenly between 0 and 1, with a slight increase in responses around the extremes, but insufficient for them to be any use. Without having any frame of reference for the person answering your questions you may as well pick values at random. (I'm assuming that because the number of available results do vary from value to value they're a representative sample, rather than, say, 1,000 from each value. If they're *not* a representative sample, can you give some figures for the actual distribution of scores?)

Plus, of course, you have some obviously nonsense answers. Apparently 7% of people think that most other people would answer the question "have you seen the moon" as "no"; allowing for rounding errors this could be a case of 13 people answering "yes" and one person accidentally answering "no", but if the number of "validations" you're getting is so low then the statistical noise will likely swamp the actual data -- all it takes is for a couple of people to click the wrong button and the whole thing goes horribly wrong...

In other words, I still think you're talking rot, but if you *genuinely* can do even 1% of the things you claim then I'm more than willing to sign an NDA to get a look at your research in its pre-patented state; if it turns out you're on to something then I'll even pitch in for free! (Hopefully I've made it fairly clear that I'm making genuine observations about your work froma position of interest and reasonab.e sense -- I may think you're a crank, but you've put a lot of effort into it so I'm prepared to work at proving it's all rubbish. :)

(Also I'm hoping that you're not using "patent pending" as a validation of anything -- getting a patent granted in no way signifies your idea is valid, otherwise there wouldn't be countless perpetual motion and free energy devices patented...)
Permalink Mat Hall 
August 24th, 2005
My results are not clustered. It has seven peaks in a smooth distribution. Look at the distribution, which is public:

http://www.mindpixel.com/chris/2005/07/shape-of-human-consensus.html

Pay close attention the the symmetry around .5 remembering this is nearly two million forced choice measurements of people trying to create intems that are either 1.0 or -1.0.

Remarkable symmetry that is a strong indication of the underlying continuity of mind.

And for you hate cheese question, yes. If you use the word hate, it is activated in your brain at that moment. If you frequqntly use the word hate, it is a pretty good indication you are a hostile person.

My model of Muppet is pretty interesting as you might guess.
Permalink Chris McKinstry 
August 24th, 2005
Oh, and as for your NDA offer, do you have $2 billion in assets to secure the NDA?
Permalink Chris McKinstry 
August 24th, 2005
"If you frequqntly use the word hate, it is a pretty good indication you are a hostile person."

I'm calling bullshits again. Of all the fat-headed things I've heard people say, this one is very close to taking the prize for requiring the most gigantic hat.

Further to my original point about lack of context introducing ambiguity into random phrases, "hate" has two distinct meanings:

1) the emotion of hate; a feeling of dislike so strong that it demands action
2) dislike intensely; feel antipathy or aversion towards; "I hate Mexican food"; "She detests politicians"

"I hate cheese" falls into the latter category, is nothing to do with hostility, and is merely an expression used to denote that you *really* don't like cheese as opposed to just not being all that keen on it. Assuming it always denotes hostility only goes to prove that you don't actually see the folly of your claims that from content alone you can ascertain the emotional state that accompanied the phrase.
Permalink Mat Hall 
August 24th, 2005
"Oh, and as for your NDA offer, do you have $2 billion in assets to secure the NDA?"

You're just taking the piss now...
Permalink Mat Hall 
August 24th, 2005
Chris-
My mind maps nearly every single one your text strings directly to the BS region. I am not alone in this.

You have a serious, serious credibility and communication problems. I suspect that these are merely symptoms of even more serious issues, but that is beside the point. No matter what the validity of your claims happens to be, I think you should take stock of what you've already *verifiably* accomplished, and congratulate yourself for it. I'm sure that it is worthy of praise on its own merit, even if it doesn't evolve into a strong AI. Fight off every temptation to stretch your accomplishments beyond their real, practical merits.

Once that's done, you will be in good place to begin plotting your future work, either with this system or with another iteration. Involve others, listen to experts, break you're theoties down into the most readable format possible. You'll likely need help with that last bit, but that's okay. Spend some more time writing and exploring your ideas about privacy. If you keep things real, maybe you can come up with a couple cool social experiments and have them performed by some university students.

Mat-
Other than this last bit of condescending advice, I am done indulging our friend here in his flights of fancy. It's lost most of it's humour, and there isn't much other value in it. I think it's getting rather tired and predictable. :)
Permalink Jeff Barton 
August 24th, 2005
> break you're theoties down into the most readable format possible

Please excuse the mistakes, and the irony ;-)
Permalink Jeff Barton 
August 24th, 2005
Jeff, I'd quite like to sit down with him and discuss it in depth based on some deeper knowledge of his body of work, but yes, this is getting a bit tiresome. He's lightened up a bit on the completely inscrutable gibberish, but it's still effectively "I've solved some of the most intractable problems in computing, psychology, and philosophy all at once but I'm not going to show anyone until they recognise my genius based solely on my unsubstantiated claims", and there's very little fun or value in trying to deal with that sort of thing. Still, it does pass the time of day, and makes a nice change from being berated for having no morals, social graces, or intellect in general. :)

Your advice is sound, too. I'm sure the work he's *actually* done has some value, but he's trying to get way ahead of himself, has a great deal of difficulty articulating what he's on about, and would rather make extremely dubious claims about what he's going to be doing Real Soon Now than demonstrate what he can do curently...
Permalink Mat Hall 
August 24th, 2005
You see the thing is guys, this is the new world of the scientific blog, where first to blog gains scientific priority.

It is important for me to have my ideas out there where they can be found with google queries. It links me to like-minded people and it prevents people from getting patents on things I have disclosed when the USPTO hits my site and discovers prior art. And they do - I caught them in my logs just yesterday.
Permalink Chris McKinstry 
August 24th, 2005
Yes because the US patent office was there researching your fine work, and it certainly wasn't a patent office employee merely surfing. Maybe he found your site via google after your 8,000 or so posts about your patentable work.

Your appeals to authority and name dropping are two of the things that make you look like a dunce, Chrissy. Of course, you ARE a dunce, but if you want not to look like one I'd cut those things out.
Permalink muppet 
August 24th, 2005
Oh, stop whining. You sound like my mom. Especially when you call me Chrissy.

You see, I see the USPTO queries. I know why they visited.

Anyway, I just got my first experimental confirmation of my theory.

You'll be reading about it elsewhere.

And Muppet, to paraphrase Douglas Adams: That's for all the hits.

I'm otta here boys.

I'm taking my data and moving on.
Permalink Chris McKinstry 
August 24th, 2005
Thank you for gracing us with your presence! Remember us when the money comes in!
Permalink Jeff Barton 
August 24th, 2005

This topic was orginally posted to the off-topic forum of the
Joel on Software discussion board.

Other topics: August, 2005 Other topics: August, 2005 Recent topics Recent topics