Consider the human interface. There’s only so many ways we can receive information, some of them less subtle than others. And only so many ways we can express ourselves, again, some pretty subtle and other ways more direct (input:pain :: output:punch).
It kinda comes down to this: as an old fart wishing to communicate something memorable to my grandkids prior to my expiration date, I can either:
a) speak a story, direct in person or indirect to a recording; b) write a story with a stylus or keyboard.
There’s no other way I know; speak or write. Vocal cords or fingers. Sounds or scribblings. Oh, I could turn on the video camera and go into mime mode, but let’s stay semi-serious with this idea.
No matter how exotic the device, it still does nothing more but capture our sounds (and maybe reinforces that with video); and our scribblings.
The “stylus” (sharp stick, burnt stick, graphite stick, pressure-screen stick, telegraph key, keyboard, etc.) has always been slow. Because it is a “serial” device, one character at a time. The keyboard was a big improvement, but has pretty dismal limits … a good telegrapher can send 20 to 30 wpm steadily; a good typist, maybe 60-90 wpm consistently.
A good talker? 100-120 wpm? I’m not sure. Speech is also a “serial” technique: phonemes, syllables, strung out to form words, one after another, which are just symbols to stir similar symbols in a listener.
So, how the hell to we “interface” our grunts or scratches into a computer more efficiently than we currently do?
Well, short of the “neural contact” or “brain wave” technique of the sci-fi writers, it’s gotta link into grunts or scratches … there just ain’t any other way, hmm?
Now, I’m gonna make a basic assumption here. The computer is a helluva long way from understanding what we tell it. I can input “I am going to destroy this computer as soon as I finish this sentence” and the computer hasn’t got a clue what doom that sentence forbodes. If it could understand, then it’s safe to say we’ve got a whole different communication input situation on our hands.
So, we’ve established that the very best computer remains as clueless as the old Royal Standard typewriter I’ve got stashed in my closet … “Hal” from “Space Odyssey 2001” notwithstanding.
That leaves us with the concept of what “intelligence” might be deposited in the computer, which is basically the code of whatever software is resident within. The box ain’t no brighter than the nerd who wrote the software, which leaves me kinda cold when I hear that a computer model on a supercomputer in the basement of an Ivy League university says the earth will warm/chill up/down and all life as we know it will be reduced to cockroaches and piss ants. I’ve just got to assume that whatever “intelligence” was writ into that machine is far from all-knowing and omnipotent. A building full of nerds ain’t “intelligence squared.” It’s just a gaggle of nerds and the group is probably less smart as a whole than the single brightest nerd there.
Gray’s Law of Group Intelligence: Group IQ = IQ brightest nerd / number of nerds.
Well, that might be a touch harsh. It’s totally true when applied to politicians, but nerds tend to consult a bit more kindly together so lets just say they would do better to listen to the few brightest among them, but my point is still valid: there ain’t no such thing as intelligence squared when you pile nerds upon nerds, so let’s not get all worshipful of supercomputers and software models. They’re a helpful tool, but not the Oracle of Delphi. Okay?
But I digress. My point is this. When it comes to improving the human squawk or scribble interface to our communicating machines … (no, Phred, I’m not talking about the “heads-up” eyeball monitoring destruction machine interface! You can’t write a letter to Aunt Sadie that way, unless it’s to tell her how you felt about her farting at the Thanksgiving table last year by sending a hellfire missile down her chimney today!!)
Sheesh! but I do digress again!! That’s the problem with the elderly mind. So many thoughts roll around like marbles and my tired old brain can’t pin ’em down fast enough to bag ’em and silence ’em.
So, if the computer ain’t smart enough to understand what we think, or what we intend, then we’re still limited by what the smartest nerd can write in code to make instruction sets that can sort out what we humans can physically input to it. Take yer choice: squawks, or scribbles.
As far as scribbles, we’re stuck on qwerty keyboards, or — gawd forbid — somethin’ like the stylus gestures of the pressure sensitive screen, like the old Palm “graffiti” writing. That’s a huge step backward, unless you’re skilled at Palmer shorthand symbols. Court stenographers use a kind of shorthand keyboard, but we can’t expect the whole human race to spend a couple years learning that method. Or can we? I dunno, but I figure that if we can’t dispense with qwerty, we sure ain’t likely to adopt shorthand. So, dispense with stylus. We’ve hit the limit.
That leaves voice (squawks, screeches and grunts). Star Trek brought us Scotty talking to his computer. The ultimate nerd and his ultimate communication interface. The most overlooked genius of the entire series was the faceless, nameless, never-mentioned nerd who wrote the code that understood “Scotty talk.” Ever consider that? Of all the plots and premises that Mr. Roddenberry presented on Star Trek, that was the one jewel that never got mentioned. Too mundane, I suppose, but consider this:
We as a technology society are about as close to writing the code that will enable Scotty talking to his computer, as we are to building and flying an autonomous robotic seagull. Yesterday I was watching a flight of a dozen seagulls in a tight group, tracking alongside a ferry, their wingtips sometimes brushing the superstructure while catching — in flight — breadcrumbs being dropped to them by a kindly woman passenger. Try programming that into a mechanical bird, Darpa! And the CPU can be no larger than the biological unit’s brain! The next time somebody calls you “seagull brain” you might take it as a compliment!
But, our limitations aside, that is the goal. That is the future. We need a software interface that can function like a human ear linked to a human brain that can sort out the chaos of our squawks, screeches and grunts and make sense of the language sufficiently well to extract some sense of the communication and act on it.
We have software now that will learn to transcribe “Mary had a little lamb” from my voice, or your voice, or President Obama’s voice, and translate it into text on the screen. But I’d bet a bucket of cold beer that if all three of us together sang a chorus of “Mary had a little lamb” into the computer, it would bomb. Or any one of us singly, after another of us had “trained” the program to understand our voice. See my point? That’s why we still have keyboards and severe limits on the interface. We just ain’t smart enough to write the software that can understand our voices, our accents, our languages, singly or collectively, sad, angry or excited … reliably well enough to replace the keyboard.
But you know what? Sometimes, as I feel the end creeping closer, I wish I could sit down in front of my beloved old Mac, with it’s shiny new LCD monitor screen, turn on its word processor, lean back in my chair and begin to recite some memoirs for my grandkids: “Once upon a time … ”
And that’s when I reach in the desk drawer and pull out the old tape recorder.