A moodboard of musings on mediums
Interfaces exist to minimize the delay between intent and action.
consider monkey mindPong:
Pager the monkey was taught to play a game.
Given a joystick, an objective, & sweet banana smoothie rewards,
Pager learned to play.
Meanwhile, predictive models were being fine-tuned on her neural data.
Once the joystick was removed, all she had to do was think:
This is the power of deep learning.
And, with this power, we obsoleted Pager’s joystick.
Pager’s intentions simply became her actions.
As frontier AI labs catapult “learning” — as we have come to wield it — into every problem under the sun, I wonder what sorts of joysticks we — humans — will obsolete from our own paws.
With automation comes power.
As our intentions on the computer are met with action faster & faster, our throughput increases, our ability to chunk information accelerates. Each keystroke now commands not just one computer but dozens, moving not mountains but mountain ranges with every click & command.
We are becoming conductors of digital swarms, each gesture multiplied a thousandfold through the software at our command. But as any conductor knows, the greater the ensemble, the more crucial the precision of every movement.
I imagine our future is one where gestures, subvocalization, & command-line interfaces power a day’s worth of work with just a flick of the tongue, a mumble of the throat, & one line in your terminal.
In Neal Stephenson’s Snow Crash (the Sci-Fi novel which gave us the word Metaverse), a character, The Librarian, is a highly advanced, human-like software program, readily accessible to the protagonist Hiro Protagonist, allowing him to navigate through large amounts of data with ease; essentially functioning as a virtual assistant within the virtual world.
A fictional product of 1992, The Librarian always struck me as a charming attempt in crafting a believable, futuristic fiction. At the time, there was a class of problems computers could solve, & there was a class of problems computers could not.
Despite our librarian’s “innate ability to learn from experience” (13.53), he would occasionally take something seriously which Hiro had meant as a joke, leading Hiro to suspect “that the Librarian may be pulling his leg, playing him for a fool. But he knows that the Librarian, however convincingly rendered he may be, is just a piece of software and cannot actually do such things”
The Librarian balks at fuzzy questions such as, “Do you find this information is worth sharing?”
Yet, deterministic commands: “List all Persian rulers in the 1st through 5th centuries, AD”, he tackles with ease.
Reader, you & I might implicitly understand the difference between fuzzy & deterministic commands quite well, but our children will not. Why should they? They’ll never need to learn.
The fuzzy is ours to grasp — like an otter to a rock.
There was once a class of problems computers could not solve, & there is now a class of problems computers can solve much more.
We’ve only just invented the tools which render this class of software possible. When it comes to controlling swarms of them, swarms of complex actions with granular precision, us humans have no solution.
In our soon-to-come world, where every thought ripples out through forests of consequences, how does one keep control?
Some food for thought:
https://www.ri.cmu.edu/pub_files/2015/0/kolling2015human.pdf?utm_source=perplexity
