Predictive text generated conversation

Thread Starter

strantor

Joined Oct 3, 2010
6,875
There seems to exist a "curiosity continuum" where each person has a different threshold of satisfaction. On one end are people who barely care about or are aware of the world around them and on the opposite end are people who are easily obsessed with understanding the most minute details of everything. People at the extremes are often regarded as retarded or crazy so it's fortunate that most people are somewhere near the middle of this bell curve.

I suspect that most engineering types are somewhere on the latter side of the median. This higher level of curiosity makes it possible for them to solve complex problems and understand the intricacies of the natural world but it also means they're prone to obsess over puzzles, even when they know or suspect the solution will yield no real value and that they could do something better with their time.
I think this is right on the money!

I've had this thought too, particularly as it applies to people at the lower end of the spectrum. I never considered my place on the bell curve, I guess because it's my tendency to assume that most people think like me (so the bell curve never occurred to me). I notice it in my coworkers; when I attempt to train them to do something, it's like leading a horse to water and trying to shove a garden hose down it's neck. I'm easy to train; I ask questions, and I use google for those questions I feel embarrassed asking. Other people I've tried to train, don't. They expect everything spoon fed to them, and then expect to hold you accountable for any holes in their education. I wonder what they google; probably just memes and sports scores. I came up with a little catch phrase to describe it: "learning should not be a purely reactionary event." Like when you touch the stove and get burned, you learn not to touch the stove. For some people, it seems this is the ONLY method by which new information is added to their libraries.

It concerns me that my oldest daughter does not ask IMO a "healthy" amount of questions. She doesn't really seem to be curious about many things. I sometimes ask her "don't you wonder why ______..."; her answer: "um, well, uh, ...noooo, ...I guess?" (sheepishly as she's not really sure what answer I want to hear). I try to instill curiosity in her, though I'm pretty sure curiosity is not something that can be instilled. I sure hope she does well in life; I don't know how you go anywhere or do anything important in life without curiosity - for me, in my life, it is the one thing I owe any ounce of success to.
 

tcmtech

Joined Nov 4, 2013
2,867
Good lord that's almost like the difference between me and my wife.

I take bit of pride and joy in learning something new even if it's just bit of filler info to go with something else I already know.

My wife however largely approaches learning new things like it's some sort of great insult that I expected her to put forth any degree of effort to learn anything no matter how trivial it may be and if she did have to learn something don't expect her to take any responsibility for her having not learned everything exactly perfect because if she didn't learn something it's clearly your fault for not having taught it to her in a way that she would remember.:rolleyes:
 

joeyd999

Joined Jun 6, 2011
6,297
Good lord that's almost like the difference between me and my wife.

I take bit of pride and joy in learning something new even if it's just bit of filler info to go with something else I already know.

My wife however largely approaches learning new things like it's some sort of great insult that I expected her to put forth any degree of effort to learn anything no matter how trivial it may be and if she did have to learn something don't expect her to take any responsibility for her having not learned everything exactly perfect because if she didn't learn something it's clearly your fault for not having taught it to her in a way that she would remember.:rolleyes:
What you fail to realize is that she doesn't have to learn anything because she knows it all already. Haven't you figured out that you're the stupid one?
 

cmartinez

Joined Jan 17, 2007
8,763
Hey dude you know that if there's a little late anyway I have the wrong person to get you excited about that now that may have a meeting this morning I don't want your place from my email until the church and I can't finish.

The above was generated by clicking the third suggested (predicted) word in my smartphone's texting field 45 times.
I bet you first had to seed it with "Hey"
 

OBW0549

Joined Mar 2, 2015
3,566
I cannot understand why I am so taken by this phenomenon. I don't know what question I'm trying to ask, or what I'm trying to understand about it, if anything.
I think the reason autobabble text is so fascinating is because it resembles the ravings of a crazy person. And "crazy" both alarms us and fascinates us, kind of like a crash scene.
 

Thread Starter

strantor

Joined Oct 3, 2010
6,875
I bet you first had to seed it with "Hey"
not mine. I begin a new text, before typing anything, with 3 options: "OK", "I", and "Hey" - this is a Samsung S4. I imagine its suggestions are based on my most-used first words of a text. I wonder what other's first suggestions are.
 

Thread Starter

strantor

Joined Oct 3, 2010
6,875
I think the reason autobabble text is so fascinating is because it resembles the ravings of a crazy person. And "crazy" both alarms us and fascinates us, kind of like a crash scene.
Or, kind of like Loosewire's posts. Maybe that's why Loosie is popular here. He's found an audience of Engineering types on the outskirts of the bell curve who would spend time trying to decipher riddles that we all know have no answer.
 

OBW0549

Joined Mar 2, 2015
3,566
Years ago I had a lot of fun playing around with a program I called Autobabble which was really, really dumb but it produced some hilariously life-like text.

As best I can recall, it worked something like this:

Autobabble took two inputs: a block of seed text, and a parameter I called the "lookback level", or LBL for short. It had two data structures: an array of integer counters with one entry for each ASCII character code and a string variable, CurrentString, of length LBL.

The program began by picking a random point in the seed text and copying the next LBL characters into CurrentString, then zeroing all the counters in the array.

It then proceeded to scan the seed text looking for instances of CurrentString. Each time it found that character sequence, it noted the character following that string, and incremented that character's associated counter in the counter array.

Once it was done scanning, it then proceeded to multiply each entry in the counter array by a different random number (this was my way of injecting some randomness into the process). Then it went through the array and picked the largest product.

The character associated with the largest product was then output to the display and was also appended to the right end of CurrentString, and the leftmost character in CurrentString was discarded.

It then zeroed out all the counters again, and started over with the next scan. Rinse and repeat, ad infinitum.

I know it sounds brain-dead, but it produced some amazing results. Given a block of seed text several kilobytes long, it would produce radically different text depending on the value of LBL. With LBL = 0, the output would be nothing more than an endless sequence of random characters and spaces, each having roughly the same frequency of occurance as in the original seed text. With LBL = 1, the output would be about the same, but you'd start to see short words like "a", "is", "at", and "the" appear here and there amongst the gibberish. At higher values of LBL, like 9 or 10 or more, the program would usually just repeat the entire seed text verbatim, over and over.

But in between these extremes it got really crazy and sounded like a raving madman-- hyperactive, frantic, obsessive and with a VERY short attention span.

It was language-independent, too. At work our Lead Design Technician was a fellow up from Mexico, and he tried seeding the program with a block of Spanish text. Same result, except instead of babbling in English it babbled in Spanish.

It was a beast of a program: several KB of hand-assembled code (anyone remember the Programmer's Reference Cards for the 6800 or the 6502?), hand loaded via a hex keypad onto a homebrew Motorola MC6800-based machine (I didn't have any floppy disk, then). The seed text had to be typed in by hand, too. But it sure was fun!
 
Last edited:

tcmtech

Joined Nov 4, 2013
2,867
Years ago I had a lot of fun playing around with a program I called Autobabble which was really, really dumb but it produced some hilariously life-like text.
I have a similar system at home but I call it my wife.

Insert a inert unassuming comment and get back ever increasing and diversified arguments in return until everything spirals out into the realms of absurdity.

Can be damn funny so long as you don't fall for any of the counter argument baiting that goes with the program running its course.

The other day I inserted the comment about having had a dream that I was back at my old job. Nothing more nothing less. Within less than five minutes she had spiraled that out into a discussion that since Pluto has a bluish haze in its atmosphere that it's clearly a earth like planet that is already inhabited by humans and it's a world government cover up hiding the truth from us. o_O

I am thinking about inserting the comment that 'my butt itches' tonight and see what sort of random bable that spirals out into. Stonehenge, Pagan rituals, Aliens, government cover ups maybe? :rolleyes:
 
Top