Hi thread, how about something offbeat?
American Computer Company announces Startling New Revelations
about the Transcapacitor: COMPOUND TRANSCAPACITORS
appear to act like "CONNECTIONIST NETWORKS" leading to their ability
to be Programmed: Could this lead to a New Kind of
CPU capable of Artificial Intelligence
Functions? special to accpc.com written by Asher Idelman and Warran Hite ver. 2.0
A spokesperson for ACC today indicated that a "new tack" had been
adopted for dealing with the Transcapacitor issue, relating to the
possibility that its original "engineers" may have used it in "neural
networking computers", as follows:
For decades, Computer Scientists have been experimenting with
"Connectionist Networks" and "Neural Nets", complex devices which
operate like a network of "nodes" that can be activated by the
interaction with data messages, can be programmed and can also, in
essenceÿ "learn".ÿ The simplest of such "nets" act like arrays of
'adaptive switching networks', so called "neural nets", not completely
unlike the way some parts of the Human Brain are theorized to work, they can receive information, learn about it by storing "patterns" in their switching fabric that can later be used to process other data and derive logical decisions based on what has been "learned".ÿ They are arranged in different "topologies" electronically, for different applications, to interact differently in various applications. Their "learn and store, interact and recall, and apply" process has been adapted, sometimes successfully, to advanced AI Computers, computers with fast, yet limited ability to intelligently interact with data and perform basic decision making applications when encountering similar or dissimilar information.
Various architectures of "neuroconnectionist networks" have been built
by Computer Scientists over the years, varying from ordinary networks of single or multi-level switches, to vast arrays of multi-CPU systems, called "massive parallel computers" like ACC's own Valkyrie,ÿ Intel's Sandia Supercomputer and IBM Deep Blue. In examining the complexities of the Transcap, ACC has been attempting
not to produce a viable Commercial Device, but to learn about the
documents it received that may detail devices that some claim came from an Alien Species.ÿ The Transcap, clearly beyond present human
manufacture (without the original, ACC is having a time getting it to
work stably, and feels it will be several years before the prototype
90GB versions operate on a basis that makes them practical) has yielded a very unusual conclusion by the Human Scientists who have examined it: "It's not that it is unstable, we just didn't realize what it was designed to do!"
While ACC originally had been attempting simply to use it to store large numbers in huge arrays, thousands of times in capacity to highly dense memory, such as SDRAMS or Magnetic or Optical Storage, in house Computer Scientists, testing the device and examining backing documents, have been startled into rethinking the device for a second application.
Whoever engineered it, may have had even denser, faster forms of memory they used as "Storage"-like hard drives, perhaps operating in the hundreds of Exabytes (10 to the 18th) worth of memory, but due to the lack of sufficient understanding by Human researchers in 1947, such devices remain (perhaps hidden in archives) undocumented, and the
Transcapacitor, instead of just being used as "storage" and despite its unbelievable speed, MAY ACTUALLY HAVE BEEN PART OF A COMPUTATIONAL
DEVICE, A LOGIC PROCESSING UNIT, a CPU (rather than a "solid state disk drive") and considered insufficient as a storage device by its original designers, as previously thought, despite the fact that it is hundreds of times more dense and thousands of times faster than Human Storage devices.
This revelation was based on examining how the Transcapacitor is
depicted "connected" to other circuitry, in sketches and drawings of
what appears to be "circuit board debris" allegedly being examined at
Bell Labs by the Department of the Army in 1948 and 1949.ÿ The ACC study took into consideration the unusual "readback" being received in
"modulated signal levels" that Transcaps, arranged per the drawings,
were producing.ÿ At first thought to be "errors" in manufacture, these
"modulated signal levels" are now being studied as possibly being
"logical responses to information" being placed onto the Transcapacitor Arrays.
ACC's scientists thus have derived more than just a basic suspicion that Complex Arrays of Transcapacitors may actually have had the purpose of receiving telemetry, video signals, and logical transaction, and were intended to remember crucial pieces of logic about them, by interaction with what appears to have been SOFTWARE, programs which were embedded in them and which were used to interact with them dynamically, by an array of "adjunct nanoprocessors", logic circuits designed to enhance "subnetworks" of Trancaps within an overall 90GB or larger array. It appears as if the diagrams are telling a story not previously considered: they refer to large numbers of what we might today call "nanoprocessors" - small computational units which were interconnected with the Transcap Arrays, in a way which suggested that the Transcap Networks could be programmed to "learn" from images, telemetry and logical numbers presented them in streams, by storing complex symbolic representations with significance to a given application, and then, applying these stored "neural network memories" dynamically to "nanoprocessors" intended to issue "commands" to external devices.ÿ The Transcaps appear to be able to be controlled within given sub-network frames and caused to respond differently (essentially making "decisions" on a fundamental level) based on controlled circumstances and the interaction with specific data.ÿ In essence, each Transcap acts like a "line" of neural network nodes that can be set to interact in a specific manner, and can be "dynamically interconnected" with others in a way, that can pass information in and among each other, measuring and identifying the information, and categorizing it for future reference, in essence "learning".ÿ Could this be part of an Alien "Learning achine"?
Because the design of the "nanoprocessors" have not yet been located
within the documentation in the Shopkeeper's Notebook, ACC is not
certain exactly certain of all the details as to how the original
engineers intended to advance from just "storage" to "artificially
intelligent active memory" used as Neural Networking, but research is
underway to try to isolate exactly what the structure of such
"nanoprocessors" may have been.
However, the drawings have led them to conclude that the tight coupling of large arrays of intelligent "neuro-connectionist" switching array networks with many microprocessing devices and advanced software, may represent the future of computing not presently in common use by Human computer engineers. Because of its complex structure, and large "granularity" when compared with an ordinary Transistor, the Trancap could represent an enormous breakthrough in development of neural nets.
In the near future, ACC intends to reveal one of the drawings of what
might have been a highly evolved circuit board from an alien craft, if
what some have alleged took place, an actual encounter with such a craft by the US Military in 1947. We may even give the entire design out FOR FREE to semiconductor manufacturers such as INTEL, AMD, SG THOMPSON, TILLAMOOK, ANALOG DEVICES, NATIONAL SEMICONDUCTOR and others, "for the overall good it might do for Humanity".
------------------------------------------------------------------------
(c) 1998 American Computer Company.ÿ All Rights Reserved.
compamerica.com
ÿ
ÿ
ÿ |
