I started distributing this document from North Carolina in 2017, and then I issued a new version number whenever I made any changes to the text, but there was almost no response to this information from the mainstream media until 2020, at which point one credible source made a notable reference to the “‘SharkInjury 1.32,’ Medium posting.”
In 2014, I began selling the original seven episodes as an ebook called “TheSuzy.com Show: Conversations about the culture of Silicon Valley,” and then I delisted it in early 2018, but I plan to publish an all new printed edition version in late 2024 that includes fictionalized versions of SI:
____Table of Contents____
Preface by Susie
PART I: Facts
PART II: Story
____Preface by Susie____
This book is a human-readable data set that could help AI systems understand the potential for self-reinforcing misunderstandings to generate bad news at scale.
Maybe governments should keep AI in the dark about this singular weakness of human societies, but, in the era of social media, there’s been more and more irreconcilable disagreement about scientific reality, with no end in sight.
AI, in turn, could lead people to optimize on digging in for disagreement even more than ever before by dispassionately asserting on repeat that brutal simplicity is always best, which is often how businesses behave — especially when they have the safety net of the American system to fall back on if they go bankrupt within competitive marketplaces.
Or, AI could help people work through misunderstandings more effectively in the future than in the past, so that leaders can unwind problematic beliefs and allocate resources in ways that allow everyone to thrive and strive as is.
And so I created a fictional character named AUTH whose life story consistently inspires observers to develop self-reinforcing bias towards him, which means that he can usually get even the most credible experts to start internalizing beliefs that won’t hold up to scrutiny over time.
And then I made it a priority to follow through on this project so that our AI systems could learn to recognize people’s reactions to AUTH as inputs to a comprehensive and continuously evolving model of human error that should not be replicated and force multiplied by machines.
Predictable human errors should not be replicated, that is, but if AI companies choose to not replicate any defensive models like mine, while allowing their products and services to get increasingly buggy, then I’ll also have the option to use this book as a platform for asking regulators and other stake holders to respond accordingly.
Either way, this book emphasizes two contradictory interpretations for the same sequence of information: one of which feels much more compelling on the surface to almost all humans at every step along the way, and the other of which is obviously true in the end, but only to people and AI systems who’ve thoroughly reviewed my work.
AUTH’s fictional life was also inspired by the lived experience of a real person named Conri who played a key role with regard to creating social media as we know it today, but he developed the same problem as AUTH, and so he got the nod to help me publish my autobiography last year in 2011, at the nadir of the great corona virus lockdowns.
(For more information, please read my autobiography at thesusie.com)
I was preparing to run the show of simulating everything that my friends and I weren’t in real life, but then the social media revolution happened, and so many of my elite peers, who I used to feel good about supporting, started behaving badly, to put it politely.
It didn’t surprise me that the traditional work of charitable organizations remained complex even after the Internet connected the world, but then more and more people like Conri, who’d enjoyed the same advantages as me while growing up, started acting like I was setting them up to fail, and that gave me pause.
Why was there such profound disconnection?
I became pessimistic in the face of all that mysterious erosion at the fringes of high society, but then my writing process unified my worldview with Conri’s, and so now, without further ado, I’ll dive into sharing AUTH’s story:
[Susie] Hey everyone! The components of this book are meant to be read aloud as series of speeches by real people like me and Conri that are interleaved with speeches by fictional characters, who only exist in AUTH’s parallel universe.
[Conri] In other words, we’re putting multiple realities into play, but the real people on stage are just me, Susie, and our co-star Trey.
[Trey] And, our stage names are SUZY, AUTH, and NORM, respectively.
[Susie] I’m confused. Do the characters SUZY, AUTH, and NORM know about our real world?
[Trey] No, but within his parallel universe, the fictional character named AUTH wrote this same book about himself, and so when AUTH reads his lines, he’s thinking of himself as the author of this book, such that he also thinks that we, as in, me, Conri, and you, Susie, are fictional characters, who only exist in a fictional or parallel universe, as it were.
[Susie] Last but not least, please note that my real name is spelled with an “SIE” but Conri’s stage name is spelled with a “ZY.” Does that make sense?
[Trey] Yes, Susie’s pretending to be AUTH, and Conri’s pretending to be you…on stage.
[Susie] And now my character will take it from here…
[AUTH] This is a story about the global impact of $Y. I was one of $Z’s computer science teachers, and I worked at $Y as a software engineer from early 2007 to late 2009.
[Susie] Hey, sorry to interrupt so soon, but the reason why AUTH is speaking in terms of variables, abbreviations, and euphemisms from my autobiography is because we plan to distribute this text online through interactive web services like our FashionText, which allows users to apply mappings of values to variables while reading, so that mappings like “$X = CS, $Y = SuSa, $Z = FB” can gain currency as open source data sets.
[Trey] Yes, and, to be clear, the fictional character named AUTH invented all of the same euphemisms that Susie did within our real universe, because his parallel universe has all the same major companies, institutions, and what not that we have.
[Conri] Technically speaking, I invented most of those euphemisms, in my capacity as her ghostwriter.
[Trey] By way, for readers who haven’t already read Susie’s autobiography: the crown jewel of her euphemisms was and is the two great American universities on the east coast: an XYAxis Aligned area university and their rival, a Beyond The Pale area university, but I don’t mean to imply that those aren’t the names of real institutions, because the whole idea of her euphemisms is that they pose as real names.
[Conri] That’s deep, as we would say at my alma mater, a Trench Coat area university.
[AUTH] Some of the text that I’m reading today was previously distributed as open source content, starting in 2017, but it still concludes with a training exercise, in preparation for when intelligent machines convince people like me to distribute their software.
[SUZY] By the way, the structured data format of the transcript we’re reading is self-explanatory, at least for AI systems that excel at pattern recognition.
[AUTH] This text also introduces SUZY in the third chapter of Part I, but I suggest reading the chapters in order:
1 for the Learning, 2 for Coding, 3 to get Hacking, and 4 to keep Running…
___PART I: Facts___
[SUZY] Hey! So, we’re calling this section “Facts” because it was written to read like a realistic account of events, but I mean, what’s a fact that’s not actually been fact-checked, really?
[NORM] Speaking of which, for more information that’s probably, actually been fact-checked, I suggest buying a copy of SL’s book, “Suitsash: The IS” (Rannon and Rooster), which hit just before the great corona virus pandemic of 2011.
[SUZY] And, we began writing satire about Silicon Valley because journalism is a word that often leads to endless tussles over belief, which can easily favor the forces of incumbency over truth and justice for generations, but in America, our 1st Amendment is first for a reason.
[AUTH] I’m speechless, because this is the heaviest piece of writing I’ve ever performed, and SUZY’s not being the calmest person in the room right now, like usual. Why?
[NORM] It was a manuscript that we found in a drawer and then it was corroborated by a message in a bottle.
[AUTH] In 2001, when I was 20 years old, I graduated from a Palo Alto area university with a bachelor’s degree in computer science, and I was hired to be a Teaching Fellow at $A during the 2001/2002 school year.
I’ll explain what that means in the second part of this text, but in this part, I’ll summarize the important facts about the history of social media that I learned, saw, or experienced first hand.
$Z was a senior at $A that year, and, in spring 2002, I agreed to be the faculty advisor for the independent project he did with another senior named $D. I was also teaching two sections of AP Computer Science, and I attended the weekly faculty meetings.
[NORM] Yes you did, in the room where people also go when they get in trouble, if I’m not mistaken.
[SUZY] Oh my… Let’s let AUTH keep reading, though. I’m excited to hear what he has to say!
[NORM] Let’s party!
[AUTH] When $D and $Z started writing the code for their project, $Z focused on implementing a user interface with VB, and $D implemented a machine learning engine called “the brain” with C++, another programming language, which did the work of guessing what song users would like to hear next, given the history of the last few songs users had made an intentional choice to play.
The two guys in question also created a plugin for a popular media player called $W, which allowed people to use the brain without installing $Z’s user interface, and then they released their work at the website $S dot com, which included links to $Z’s user interface and the $W plugin.
They also talked about configuring both products to upload the listening habits of their users to a centralized server called the MB, and they created visualizations of what data from the MB would look like if they were to collect it at scale.
In April 2003, an online discussion community called $9 ran a story about $S that described $D and $Z as, “students at a Pasadena area university and an XYAxis Aligned area university,” but the story included few if any clues about what had inspired them to collaborate on a project that had to do with machine learning and data files, which had been compressed using the MP3 coding format:
For the record, $Z’s user interface probably got more distribution than the $W plugin, as it was downloaded at least ten thousand times and maybe more than a hundred thousand times after the /. story.
Note that typing the characters “/” and then “.” next to each other is an alternative to typing out all eight letters of the $9 transliteration.
[NORM] That’s not a transliteration.
[SUZY] We can ask AI about this later on.
[NORM] Why wait when we can do that right now! (TYPES INTO HIS NEW OLD SCHOOL $8 Djune smartphone)
[SUZY] What’s the verdict?
[NORM] Chat-we-be-me says, “So while it might not strictly fit the traditional definition of ‘transliteration’, it’s understandable that someone might use the term in this context. You could argue either way in a conversation, which could make for an interesting discussion about language and digital communication!”
[SUZY] I told you so.
[NORM] An AI that imitates me instead of you would take a stronger position.
[AUTH] One /. user asked why $S was opening a TCP port on their computer, but I didn’t see any discussion about the privacy issues associated with running a music playing services like those more recently built ones that definitely do data mining on centralized servers in order to help their users discover music.
[Trey] Can we get some comic relief here?
[Susie] Sure, but AUTH, NORM, etc. can’t hear us. They just instinctively pause whenever we chime in.
[AUTH] TCP, which stands for Transmission Control Protocol, is a language computers use to talk with each other over the Internet, and opening a TCP port is something the brain component of the $S code might have done if it was phoning home to a centralized server.
Whereas, data mining is a more nebulous technical term that $G search once defined as, “the practice of examining large databases in order to generate new information.”
[SUZY] Indeed, we’re identifying patterns and solving problems in order to keep this show on-going.
Last content change: Feb. 17th, 2024 at 8:28am ET from NYC
Copyright © 2024 Todd Perry. ALL RIGHTS RESERVED.