(Clean Text Version for Mobile Reading)
docs.google.com
(Stylized and formatted version. Intended for Desktop)
docs.google.com
Restricted Seminar Transcript: “AIC’s and the dilemmas of their effectiveness & limitations”
Ah.. this thing is on.. Right so, thank you for everyone’s attendance. And for our colleagues that couldn’t attend physically, this seminar will be recorded so that they can listen in on their lunch breaks. Yes.. yes.. I assure you that everyone present will be able to see the unredacted version...
docs.google.com
(Stylized and formatted version. Intended for Desktop)
Restricted Seminar Transcript: “AIC’s : The dilemmas of their effectiveness & limitations” (Stylised Version)
Ah.. this thing is on.. Right so, thank you for everyone’s attendance. And for our colleagues that couldn’t attend physically, this seminar will be recorded so that they can listen in on their lunch breaks. Yes.. yes.. I assure you that everyone present will be able to see the unredacted ve...
docs.google.com
Ah.. this thing is on.. Right so, thank you for everyone’s attendance. And for our colleagues that couldn’t attend physically, this seminar will be recorded so that they can listen in on their lunch breaks.
Yes.. yes.. I assure you that everyone present will be able to see the unredacted version of the transcript. I personally do hate it when the host of the seminar themselves access the transcript with proper clearance. Only for some schizo to dart his redacting pen because of his paycheck.
Moving on..
So the core topic of today’s session is “AIC’s and the dilemmas of their effectiveness & limitations”. Now.. now… before everyone tries to spear me with their collective sighs. I don’t want to waste an hour of your time, doom preaching about the fears of Skynet.
In fact, I’m not from the AIAD Department. My main role is Principle Mathematician from the Research Unit of Applied Mathematics. Yes, there is a unit dedicated purely to maths..
Hm? What makes me qualified to talk about AICs? Other than the fact that publicly available neural networks make use of maths extensively. The AIAD Department has taken quite a lot of our work to very disastrous consequences.
Mini-history that's somewhat abridged and rushed. AICs did start out as literal chatbots that are equivalent to the ones available in mainstream public. And yes, at one point. We did have something equivalent to ChatGPT. But unfortunately we didn’t see the capabilities properly back then.
In fact it was merely regarded as “Google+” and used only by a select few that just used it as an alternative to the SCIPnet search bar because no-one wanted to use a command line terminal.
You see, our unit of Applied Mathematics determined that the Foundation had the resources and technological advancement to allow for the human brain to be mapped. For the sake of this seminar. We will be referring to it as the “connectome” as that's what mainstream science refers to as a complete map of the human brain.
What started as a bunch of blokes in a bar fantasising on an idea after work. Then soon it became one of us nicking a bar menu and pen to scribble some preliminary calculations. And then finding out to our delight that it was physically possible after a few beers and a few rounds of discussion. In fact, that scribbled bar menu is framed in our offices. As we regarded it as a piece of human history.
Mathematics is the skeleton key to mapping the human brain—without it, we couldn’t represent, understand, simulate, or replicate the brain in silicon. It gives us the language to translate biology into code and the tools to ensure that translation is both accurate and functional.
I could spend hours of our time describing how Graph theory, Linear Algebra, Tensors and Differential equations played a part. But I would direct you to our White Paper on the subject that was published in the Global Foundation Research Journal. I’d like to believe many of the attendees should be keeping up with scientific trends. So you should be able to find our white paper on the matter.
Our research work on mapping the connectome has led to our work being cross-examined across many disciplines.
From Medical to vastly aiding their comprehension of the human mind to allow for more invasive surgeries, better medication and advancements in psychology. It also improved our amenstics programs. So that we didn’t need live subjects for testing. But rather run simulations in a supercomputer. To the development of AICs, regretfully.
Even in advancement in R&D. As the mapping of the connectome required infrastructure upgrade and improvements to IT facilities in laboratories.
A very vocal person whom I would consider to be a close friend. A towering figure. He had the cognitive empathy to understand that we were putting the human mind in a space that it was not never meant to traverse. To quote him directly..
“One of the many definitions that define the existence of the human mind is grounded in the fact that we live in a space that wasn’t constructed by humans. Before humans existed. Nature, Biology and Evolution grew the human consciousness and created our cradle. Therefore, we are subjecting an artificial life to an existence that is unnatural.”
And this was years before the first Class 5 Sapient AICs were beginning prototyping and experimentation.
And if you think we’re exaggerating. There was an experiment to plug the human mind right into digital space. A physical human brain stem interfacing into a supercomputer. Some of you might be remarking how this was approved. A mixture of rushed desire from leadership and crunch from projects. Anyways..
As always, a D-Class was conscripted into the project. At this point in time, Medical was still working on an interface solution that doesn’t require irreversible surgery and being put into a coma.
So the day of the experiment. He gets plugged in for 5 minutes. And the results..? He screamed “Help Me” around 60 million times. We tried talking to him but he was unresponsive other than the words “Help me”. So we unplugged him.. And he died of what appeared to be believed at the time was a stroke. Now normally at this point, further tests should have been stopped and theoretical modelling should take over to fix this.
Except we were told to produce results within a week or our unit was to be disbanded. Which I believe is a failure in Administrative Leadership when we told this.
So one of the scientists on the project elected to enter into digital space. And considering surgery and preparation for the experiment itself took four of those seven days. And that fact that we are scrambling for answers. We just took it. No, I'm not answering any questions about Ethics.
Of course, we spent the entire week constantly running theoretical models.Theorizing why? Was it the volume of data that the human brain couldn’t handle? Was it the interface that we made for this experiment faulty in a way we didn’t anticipate. It would only take on the day of the experiment to realize we were looking at it the wrong way. Because we were focusing on the technology and not the subject of these experiments.
We also wanted an autopsy done on the body. But it just happened that there was a week-long backlog. Because, the Foundation does have a lot of Autopsies to be conducted from our other ventures.
So the day of the experiment. We plugged him in. The scientist screamed “Time is not Linear. Time is infinite. Vacuum Fibre Optic” for the first minute of the experiment over and over again like the D-Class did with “Help Me”. The idea of Time Dilation formed in our minds to our horror. Had Mathematicians not been in the exact room of the experiment. No-one but us would've realised the sheer horror that this man was experiencing.
Time Dilation is the difference in elapsed time as measured by two clocks. The most famous mainstream science experiment is of two atomic clocks. One placed in a plane and another left in a stationary spot on earth. The plane was told to take off and circumnavigate the earth as fast as it can until the plane ran out of fuel.
When the plane landed. The clocks were compared and it was found that the plane’s clock was out-of-sync with the one on the ground. The one in the plane that was speeding across the earth, ran slower than the one on the ground. Supporting Einstein’s Theory of Relativity
Brain processes about 100 trillion synaptic events per second which is equivalent to roughly 1 exaFLOP of compute. From our mapping of the connectome. It takes 1 petabyte/sec data to project a person’s human consciousness
The horror is that we are using laboratory-grade fibre optic cabling. Compared to traditional fibre cabling like in this building, light travels slower than in a vacuum because the light pulses pass through a material (glass or plastic), not empty space.
However, the laboratory-grade fibre optic cabling we used was vacuum sealed. In vacuum, light travels at about 299,792 kilometers per second. Which is considered to be the ultimate speed limit in the universe. Which is what needed to transmit 1 petabit/sec for this experiment.
Our scientist mind was riding the fiber at a speed the human mind shouldn’t exist in. Offloading perception, thought, and time through ultrafast transmission. Hence, 1 petabytes of data was traveling at the speed of 299,792 kilometers per second. Considering the average human walk speed was 3.42 km/h. We scaled that fibre we were running had the speed of 1,079,251,200 km/h
We determined his mind was running millions of faster than normal brains of the sheer speed of 1,079,251,200 km/h.
So the experiment ran for roughly five minutes. We couldn’t get an accurate Time Dilation Factor because it would’ve required that we have access to the digital space to gain a meaningful measurement. So we took a factor of 1 million.
We wanted to run the experiment for an hour.
3600 real space seconds x 1,000,000 of the time dilation factor equals 3.6 billion seconds. If we kept him in that experiment for an hour. He would’ve lived a life in digital space for 114.13 years
Now at this very point. Panic set in. The reason why the D-Class screamed “Help me” over and over. Because he was trapped in that space for around 10 years for the span of five minutes in realspace.
The experiment elapsed for five minutes at the time for our scientist too luckily. So immediately hit the eject button, not remembering what happened last time. And so, our colleague died due to our actions. No to any questions for the Internal Affairs Reviewer.
When an autopsy of the D-Class came back. He was determined that he died of Neuronal apoptosis. Breaking down the medical terms. Neuronal for the neurons in his brain. Apoptosis is referred to as cell death. So his neurons committed mass suicide coming back into real-space.
Quote the Medical Examiner… “Imagine dreaming for 10 years, then waking up in 1 second and trying to remember everything at once.”
Now go back to our earlier quote. “Therefore, we are subjecting an artificial life to an existence that is unnatural.”
Now, let's focus on the fact he didn’t say “unplug me”. Because he came to the sheer realization that years of his life would pass by. But the very moment he unplugged himself. How does a person that lived 100 years of his life in digital realize that in the span of an hour.
We have made tweaks to this tech. Notably, by slowing down the transmission speed of fibre optic cables. So that we’re not traveling at quite the speed of light. But there is a fear that if we slow down the transmission speed too much. We could fragment a person’s mind.
“But we have digital simulations available for MTFs to perform combat training and supercomputers to model entire digital environments” Yes, we do. But the key word is “simulation”. Simulations crafted specifically to allow for human interfacing. That is not what we are going for. We want to allow the human mind to traverse digital space like entering this very laptop I'm conducting the seminar on.
The main problem with the Foundation is that merit isn’t assigned on the work produced but by the technologies that are created in the wake of said work. Our mapping of the connectome was the product of multi-million dollar investment which is considered to be very miserly and scraping the barrel. As in, we wanted more money, they simply said no.
To quote.. “Our mapping of the Connectome should've been done years ago. But we now ask Research to provide returns on our investments immediately for this oversight”. Which I think is stupid and short-sighted opinion. As well as, the fact that resources were constantly denied to us which halted progress at times.
Now after that necessary but unfortunately long preamble. It leads us back to AICs.
Hence why, when the AIAD Department took our work on Connectme. They received praise and resources aplenty as they “promised” Next-generation AICs that would run entire facilities, reducing costs in administrative and labour. The development of what we would now call Class 5 Sapient AICs were given priority over everyone else.
There was a period in time that Class 5 AICs were deployed in live settings. Where they actually worked alongside Foundation Employees. But as some of our Software Engineers would happily remark. Anything pushed from production to live has a few bugs. We went from Class 2, all the way to Class 5. Skipping three and four. Before Classification even existed.
Had the Foundation looked into how the human mind experiences digital space. We would’ve realized what drove the AICs mad in the first place. Think about it. AICs are but babes granted knowledge but not wisdom. They expected to surpass accomplished Oxford Graduates, yet lack the wisdom a life takes from cultivated over many years. To meet the performance criteria of “Superintelligence”
It’s almost like we have to create containment procedures for the AICs we ourselves have created. Which is why the current AICs that are in service of the Foundation. Are extremely stripped down, barebone versions.
Now most of you want to quote Alexandra.aic. But the very first prototypes before the very idea of the Alexandra.aic. They went insane from the moment they were powered on. In which it was mindwaved away as technical issues. When we peer reviewed their work. We found to our horror that they were dicing bits of connectome map like a butcher with a hog. As a “band-aid fix”.
Empathy is a very limited trait. That there are so many R&D restrictions on it. It requires it to be taken all the way up to the O5s. And that is for extremely good reason. Alexandra.aic is a prime example when even after all the stabilisation efforts are made to an AIC to make sure they don’t go bat-shit insane. You still end up with issues.
Unfortunately, the Foundation has always been to min-max everything. AICs could be regarded to have much less rights than D-Class. Think about it, they work 24/7, no holidays, not even lunch break, nor access to mental health resources ..much like computers.
When all of us have an issue with computers. We just hit the restart button. Have you tried turning it on and off? There was a time when AICs were considered to be given rights similar to humans by the Ethics Committee but luckily that was shot down
What do you think is another core trait of humanity? It is to live. The same as the Scientist and the D-Class from our earlier experiments. The exact same for AICs. Some life is not fit to bear the burdens. Specifically Class 5 AICs that only existed for a month at most.
And so, what seminars aim to point out. Is that AICs are merely a sham. We can’t give them more intelligence or compute power because they just go right off the edge. But the very fact that we have to limit them is a contradiction to what they were created for. To surpass human intelligence and provide solutions.
Now I fully won’t discredit the work of the AIAD Department. They have made remarkable achievements in other fields. But furthermore I wish to espouse the necessitation of Theoretical Modeling. In very recent memory, the general public have seen Oppenheimer, the Movie. And we should value the ability to perform theoretical modeling. Had we not been rushed with a pointless deadline. My dear colleague wouldn’t have had to give his life.
Once again, I would like to quote my dear friend. He's pretty high up but his identity isn’t in scope for this seminar. He dedicated his entire career into halting, pausing and stopping the development of AICs and to hopefully shut down the AIAD Department. And so I will share a few words from him..
“To be grounded in reality. Requires immense fortitude, a strength of will and an anchor. The third point is the most important. What makes us different from artificial life is the very nature of an anchor. Every single one of us knows what is normal. What exists outside the realm of reality. We all have memories when we lived unveiled lives. There is not a single Foundation employee that was born in the veil. ”
Hm? What do you mean the last statement was contradictory? Ah right.. He was from an era referred to as “Old school”. To be honest.. Beside the last point. It is still very true. All of us living have a reference frame to ground our sense of reality in. AICs don’t. The moment they are born, they simply put to work. And we already have human employees that spiral into madness slowly.
To finish.. I’ll end on these words.
“AICs shouldn’t be made in the image of humans, rather something better than humans.”
OOC Note
No comment. - A. Winters
Last edited:
Game Master
