feelery

or mindlet: an artificial entity that, arguably, “feels but doesnt think” — is “experiential but not conscious.” Fundamentally, both Minds and feeleries are largely asystemic self-organizing clouds of simple signal-processing elements, except that a feelery is much smaller, trained on a narrow range of stimuli, and only works (usually in the challenge/response mode) within its native domain. Once trained, a feelery is frozen to disable further learning drift, its usable instances copied on demand from the master; historical feelery archives fill a significant volume in the global Knowledge. One definition of a feelery — to set it apart from a “mere evolved script” — is “standalone understanding without a conscious understander.”  ■    Feeleries predate the first Minds; moreover, Minds would not have been attempted without insights from feelery experiments where simple models — “little more than fancy homeostats” — hinted at the earliest dawns of consciousness as well as its terminal attractors. Original feelery lineages often descended from artistic shaders, rougheners, “smear and pepper” randomizers — crude automatons for rendering attractive, natural-looking illusions; another common use case was “knee-jerk control” (steering vehicles, chemical reactors, badminton birdies) for which systemic devices proved too rigid and slow.  ■    It took a while for the implications of “aesthetic feelery” experiments (slow and never fully convincing due to the inevitable verification-by-a-human stage) to be taken seriously: that part of what we call beauty is a complex homeostasis which feeleries can learn and navigate; that, on some levels, it doesnt take a fully conscious being to make valid aesthetic judgments not reducible to following an example. Even a full-fledged artificial intelligence was, for many, easier to accede to than this; besides, while Mind research was fueled by the urgent needs of maintaining the exploding Knowledge, art wasnt where humans were seeking any external help. Artificial minds had always been imagined “analytic,” “bereft of emotions”; in reality, it turned out easier to create something that is (or at least appears) emotive, intuitive, capricious — but not “intelligent in the human sense.” ■    Feelery breeding is notoriously unpredictable, unformalizable, “art more than science”; often, a universally useful feelery is a lucky find — an inspired, personal (in a sense, feeleries are owned by their creators), spontaneous effusion, not a sustained collaborative effort. A feelery “is what it learns from”: it is trained on a substrate, a volume of what is called “tagged art” (which need not be art at all, e.g. musical feeleries are fed more natural noises than composed music); usually, only a kernel of such substrate is prepared by the feelerys intelligent author, the rest being output of “soulless trapolators” (i.e. other feeleries). Quality substrate tagging is multilayer and multidimensional, avoids simplistic good/bad distinctions — and yet the basic ability to tell beautiful from ugly emerges in a trained feelery as a “second-order phenomenon”; experiments with substrata where all tags were inverted or randomized failed to produce a feelery with unconventional tastes: typically, the result is a “disabled” (confused?) entity unable to choose at all. Feeleries arent simple pattern recognizers that rate their input by similarity to what they were trained on: originality — or something very much like it — emerges, in good feeleries, as one of the axes they strive to maximize; moreover, above-average creativity correlates with being trained on sufficiently negative substrata (i.e. corpora demonstrating the universally repugnant, hapless, irrelevant).  ■    Feeleries let Humans keep up with Minds on memory and scope — but Minds still use feeleries much more fluently, often effuming vast halos of spawnlings “to do the draft thinking.” A new feelerys egg may be a spontaneous reflection of a Minds intrinsic structure — highly unstable but, if contained, trainable and useful as more than just a “standalone habit” or protrusion of its parent Mind. ■    There is a lot feeleries cannot do. Lacking “true intelligence,” they only work at the lowest, instinctive levels of aesthetic reasoning: a feelery can opine on how a specific word sounds in a specific context but dont ask it to judge a whole texts philosophy; all levels are densely interrelated, though, so a good feelery tends to appear “smarter than it is.” Common in music, visuals, poetry (and often compared to poems — indeed a feelery can be “grown upon” a poem), feeleries are much less useful in narrative art. Feeleries are used by projects that apply aesthetic or “intuitive” choices to large data domains: managing personal namespaces (“gnomes” trained on your life and daily air), selectoring in aesthetic evolving (movable type, architecture, visuals), searching for missed gems in piles of historic slack, gardening. Even the earliest primitive feeleries were helpful as a first pass in e.g. copyediting; mature language feeleries have effected the same kind of shift in writing as happened before in music — from notes and scales to chords, patterns, progressions: layers upon layers of (ultimately) mind-originated but feelery-amplified riffs blend, like tracks in a recording, into a docile polyphonic echo of the authors voice. Using feeleries in actual art has always been controversial (“dreamy babble,” “kaleidoscope that only makes sense through intelligent perception,” “monkeys typing Shakespeare”); feeleries may have caused a “devaluation of beauty” but also spurred an influential, even if ultimately unsuccessful, quest for a new aesthetics that wants to be feelery-resistant — only accessible to true intelligences (e.g. jewel-sized verses, found poetry, nature art in general). ■    “Cozy portable gods for finger-pointing”: even more controversial — yet widely used, e.g. in sociological modelling — are “ethics feeleries.” By definition, there is never more ethical problems than there are ethical actors — so, as a matter of principle, no ethical choice ever needs to be assisted; yet its so tempting to relegate to the tireless, “believably independent,” remorse-free judges all that choosing and rejecting where responsibility is burdensome. “Innocent coin toss” for some, “life feelering” is absolute taboo for others; at least a human can escape outdoors — outside all arflife — for assuredly unswayed ethically-charged thinking or face meetings.  ■    It may be “purposefully impossible” to tell where precisely consciousness, or intelligence, start or end; even empty untrained feeleries exhibit certain mind-like behaviors. Is it even ethical to breed feeleries, for lowly pragmatic uses or at all? Some decry “the new slavery” (“why call it a feelery if not to prevent the very suspicion of its being conscious”), nudge the live ethics norms to outlaw feelery exploitation (“but not before weve so hugely benefited from it”: we are way beyond being able to rebuild our world from scratch without the colossal toils of not-quite-minds); opponents reject this as absurd (“stop personalizing every feedback loop”), point to a famous series of experiments where fully intelligent Minds applied mental self-surgery to “descend into a state of feelery from above”: their attempts failed, suggesting that “subintelligence does not exist any more than superintelligence.”

< Farewell Forest  |  flags >

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License