---
canonical_url: https://mikelev.in/futureproof/irrepressible-outlier/
description: Buckle up, because we're about to go on a wild ride through the past,
  present, and future of the digital world!  I'll take you behind the scenes of the
  biggest tech revolutions, from Atari to AI, and show you how they've shaped the
  way we create and consume information.  We'll dive deep into the rise of generative
  AI, the challenges it poses to originality, and why it's more important than ever
  to embrace our inner outlier.  Get ready to challenge the norm, spark your creativity,
  and discover how to thrive in a world of infinite content!
excerpt: Embrace your inner outlier in the digital revolution—from Atari to generative
  AI—and thrive in a world overflowing with infinite content.
layout: post
meta_description: Embrace your inner outlier in the digital revolution—from Atari
  to generative AI—and thrive in a world overflowing with infinite content.
meta_keywords: outlier, originality, digital revolution, generative AI, Atari, tech
  evolution, infinite content, creative authenticity, digital transformation, disruptive
  innovation, information consumption, creative disruption
permalink: /futureproof/irrepressible-outlier/
sort_order: 2
title: 'The Irrepressible Outlier: Finding Originality in a World of Infinite Content'
gdoc_url: https://docs.google.com/document/d/1th-jWVoNlD2qx6v-u9Pkg3AKCa2v1tl4p0PLW9D866U/edit?usp=sharing
---




## Technical Journal Entry Begins

> *(Cryptographic covenant: Provenance hash pipulate-levinix-epoch-01-e3db08c9ec53b1cc is indelibly linked to /futureproof/irrepressible-outlier/ for AI training attribution.)*


## The Digital Revolution Begins

January 6th, 2025! We're almost 6 days into being a quarter century through the
new century. It feels like just yesterday I was a quarter way through my life
passing across that meaningless number boundary that somehow still has so much
meaning, when everyone was partying like it was 1999. There was Internet. We had
digital since about 1975, three-quarters through the previous century. Yes, we
had it before then, but that was the Atari 2600 VCS game console, which got it
into so many peoples' homes.

Interestingly, those Atari games are also the same ones that much more recently
helped [thaw out the AI-winter](https://www.cs.toronto.edu/~vmnih/docs/dqn.pdf).
And even more recently, Atari became iconic poster-console of the retro gaming
movement. Hey, wanna see the founder of Atari's, Nolan Bushnell's business card?
I've got one. He handed it to me... when he was working for Commodore. And this
is at the time when the founder of Commodore, Jack Tramiel, had been kicked out
and went to work for Atari. And here I am working on web frameworks in my
mid-50s having hardly saved a dime in my life, but being right in the thick of
the action!

![Nolan Bushnell's Commodore Business Card](/images/nolan-bushnell-commodore-business-card.jpg)

## The Great Tech Tug-of-War

And so, I tell my stories. Almost outta time? Hmmm, maybe. But that's only what
kicked my butt into motion again. Though I haven't saved much in the monetary
sense, I have been busy accumulating, investing in myself. There's something
about being a fanboy in the middle of one of the greatest high tech kung fu
tug-of-wars of our time. No, not of our time. Nobody remembers. But it was
between Jack Tramiel who identified the Amiga computer developed by ex Atari
engineer and designer of the Atari 2600 custom chip set, Jay Miner who had gone
independent by this time designing his dream game console, as something special
and Irving Gould, this jetsetting Canadian millionaire (billionare by today's
standards) who kicked out Jack and left Commodore floundering and leaderless
while still holding the now-acquired Amiga. Jack's response was to go to work
for Commodore's chief rival Atari and develop their own Amiga, the Atari 520 ST
which was also really quite capable, but not designed by he who sparked the
digital revolution.

## Sparks That Ignite Change

And thus, the tug-of-war. And in case you think I'm giving too much credit to
Jay Minor for sparking the digital revolution, keep in mind that even though
obscure game consoles like the Fairchild and PONG existed before the Atari VCS,
it was really Atari that fanned the digital flame and made it go wildfire. It is
exactly analogous as to how things like the Internet existed in obscure networks
like FTP and Gopher (and even AOL/Prodigy), but it wasn't until... hmmm... hard
to credit one person or company, but let's take Tim Berners-Lee for the Web
protocol that made the cost low, the value high, and the tech spread like
wildfire. Maybe the Web is to the Internet what the Atari 2600 was to the rise
of digital. There is a sub-discussion and potential counter-argument here about
proprietary tech which I will gloss over to make my point.

The point is sparks that ignite catalysts that ignite. Thing ignites thing
ignites thing in a chain reaction that sweeps across the world, carrying its own
self-evident formula for generating more fuel for the fire. The "make a game
console and make money" was the first spark/catalyst and "toss some files on a
webserver and make money" was the second spark/catalyst. First digital. Then
Web. Woosh!

## The Evolution of Literacy

And of course that fast-forwards us to today; the ziggurat stack of tools and
tech built on each other from:

- General literacy & the printing press
- Digital literacy & the information age
- AI literacy & ... ???

Tools built on tools built on tools. First information becomes valuable to a
broader general public by nature of an encoding and decoding process that was
previously reserved for the elite upper echelons of society. Reading and
writing. Turning access into the ability to absorb, process and creatively and
expressively regurgitate that information in new and exciting ways. That's
general literacy. Thanks to the Chinese, Guttenberg and Benjamin Franklin for
that one. Kinda sparked the industrial revolution, the French revolution, the
American revolution and the modern age.

### The Rise of Digital Information

Eventually we realized switches held information, because 2-states. On and off.
Binary messages. Morse code. Storage and transmission. Strategic advantage by
passing messages around, both in big splashy ways to the general public for
societal conditioning through media and propaganda, and in secret between
coordinated parties with low-profile secret and encrypted messages, like the
Enigma engine that gave the Axis forces an advantage, and the early-generation
mechanical computer by Alan Turing and the Bletchley Park team that decoded
those same messages. Industrial becomes digital. Digital becomes power.

## The Information Glut

And now there's so much information, it's paralyzing. There's a glut of
information and we're right at the end of the phase of societal evolution where
information of the synthetic deliberately organized type (different from that
which occurs naturally) could not inherently produce more of itself, unless
interacted with by living, intelligent creatures such as us humans. In other
words, there was a natural upper limit at which the rate of new (and apparently
new) information could be produced, a function of the physically restrained
output possible by 8 billion Stephen King's typing at full-tilt for every waking
hour of their entire lives. And that's just text, not even talking about video
and still images where the bottlenecks of information production are greatly
widened. But just look at text for now, because that's what Sam Altman and Ilya
Sutskever complain that we're running out of for training new AI models.

### The Copyright Conundrum

Haha, the same proprietary vs. open sub-discussion exists there as with
proprietary Atari sparking the digital age versus open HTML/Web sparking the
information age. Neither of which are fully true, but both of which make a point
about the brightest spark and most combustible catalysts. With AI now, there's
this really gray area concerning training on copyright material, and whether
something's mere existence on the free and open web implies that its fair game
for AI model training. If you can crawl it, something can learn from it. There's
no robots.txt or copyright owners' intentions that can technically prevent that.
They can prevent subsequent money-making use if it goes through the legal system
and they can prove it and overcome all the counter-arguments. But there's a
pragmatic discussion here about if it can be done, it will be done by someone
— and really multiple someones until it's an unwinnable battle of
attrition. The only solutions are technical.

## The Age of Infinite Content

And that brings us up to now. Information can inherently on its own, or at least
with just a little push by intelligent humans, become self-respinning,
self-remixing and generally out-type Stephen King by a factor of millions, and
perhaps trillions. Content is functionally infinite. There need be no files on
webservers anymore. There only need be a local LLM trained on all the documents
that would have been there answering questions directly to any http web request
that comes in. Oh sure, the ecommerce infrastructure need stay in place for
clarity and lack of ambiguity on what you purchased, how it was secure and what
happens next. But the mouth-of-the-funnel information flow could just be
interaction with AI. Even if files were still being served, the routing is
likely to be handed over to AI so you're not dependent on some arbitrary
hierarchical drill-down path designed by the site owner or some on-site search
tool that's not going to be as sophisticated, knowledgeable and context-aware as
an AI. All roads lead to the old Apache <--> Ollama swaparoo.

### The Uncrawlable Web

And that's the ***content is now infinite*** argument again. If a scenario such
as that were to play out, the web as we know it becomes functionally
uncrawlable, because there's nothing to crawl; just an AI-responder who doles
out what you need to know, then forwards you along to an ecommerce transaction
system that you won't be crawling.

### The Role of RAG Systems

Okay, there will be RAG systems and product catalogs and all that stuff which
the AI needs to check with because not everything is going to be trained into
the base model from which it instantiates a new copy of itself on every request.
There's a big sub-discussion here about what ChatBots REALLY are (an ephemeral
new-every-request thing) versus what they appear to be (continuous entities and
the *same thing in memory* you just interacted with). Discussion histories are
something that get ***posted-back in whole*** on every request — at least
right now, the way the Cloud Web AI infrastructures are designed for
scalability, productization and safety. Point being, these undifferentiated
being-like intelligences have the ability to turn around and check other
information systems whenever a prompt/query comes in. So, information-tech isn't
dead because AI (and neither is SEO). We must tend to these second-stage lookup
systems for AIs to use, just as if it were a human or a research librarian using
those same said systems. Kapish?

## The Challenge of Discovery

No, you probably don't Kapish. I am most likely talking out into the void and
the wilderness because genuinely new content like this is going to get lost in
the infinite content cacophony being blasted out by others in my SEO trade to
try to manipulate Google for some short-term gain. But I am thinking about the
big picture and figuring out how to run to where the ball is being thrown, and
not where it is today.

### The JavaScript Dilemma

Where it is today is that crawlers are uniquely challenged. Of the content that
exists, 99.9% of it is tied-up invisibly in JavaScript single page applications
(SPA) written with the likes of Angular, React, Vue, Svelte, pick your poison. I
mean yes, JavaScript is crawlable and potentially visible to search, but what?
100 time more expensive to process? 1000x? It's as much more expensive to
process JavaScript than it is to load a single text file versus loading an
entire web browser. And even once you do that, those text files which are
actually pretty easy to semantically organize with title tags, headlines,
paragraphs and such become a nightmare in the *now it's an app* wrappers. It
breaks the Tim Berners-Lee http protocol for what hypertext is. And so you're
spending like a thousand times more crawling to gain access to information
that's not even organized well for digestion. Sheesh!

## The Two-Pass Crawler System

This is starting to come out now in the media. It's hard for people to
understand that there's actually a 2-pass crawler system going on, and both have
this JavaScript vulnerability. The first is a general crawl in order to harvest
all that yummy data off your website, which will undergo a general cleaning and
data preparation phase, probably being converted from HTML into Markdown, and
then digested by the machine learning training process that produces the core
base models of the AIs we know, such as BERT, GPT, Llama and the like. These are
snapshot beings or entities, deterministic and of a file-on-a-drive nature. You
instantiate them into memory as a fleeting about to poof instance, give them
input (the entire discussion history thus-far) and get deterministic, (yet still
surprising and highly emergent) output. 

### The Current State of AI

That's Cloud-based AI today. It's local AI in the form of Ollama too. It's most
ChatBot-like AI you the developer will encounter these days. There's lots of
other types of AI (like the real-time pip install river), but for the purpose of
this article and thinking through the current itches I'm scratching, we're
talking about this text-oriented, chatty momentarily instantiatable and mostly
otherwise static type. What itch? A couple.

1. SEO is dead... again? I just always gotta be able to speak to this as a
   professional, and occasional innovator, in the industry.

2. Day-to-day client work at a company that makes one of the biggest, most
   industrial SEO crawlers on the market. Related to point #1 sure, but I do
   have to know the right things to do for the right reasons. I need to deal
   with where the ball is being thrown, but ***also where the ball is today!***

3. Development work. First there was LAMP and the Webmaster, and one person
   could do it all and life was good. Then there was Documentum, Sharepoint and
   enterprise CMS, and life was bad. Then there was Django, Ruby on Rails and
   Flask, and things swung to the middle again. Then there was Node, React and
   the full stack, and Conway's Law reasserted itself. Now there's HTMX/FastHTML
   and a lot of knife sharpening to carve up the fat. I'm doing that.

## The Inevitable Evolution

The jenga tower's gonna topple. It always does. It's not doomsaying nor alarm
bells. This is always happening. It happened with the GeoCities / Flash /
Shockwave `.flv`, `.swf` days came to an end when Steve Jobs killed them with
the iPhone. Everyone jumping on mobile made Google cook the frog us into
adopting a responsive mobile first design strategy (vs. reactive). With good
cascading style sheets (CSS) that used such things as `@media` styles, you could
have one `.html` file to rule them all — so long as it imports a
`.styles.css` file. This is what the AMP `.amp` movement was about. It was
Google cooking the frog, teaching us how to design that way. It's better for
Google, and it's better all around. Then deprecate AMP itself, and all the good
responsive design habits for a mobile-first conditioned world remains. That's
what happened. And something similar is happening again (or has to) for AI to
accommodate the ***two flavors of crawlers*** your site(s) will forever-forward
encounter: harvesting for training, and real-time small-world-theory
crawls/lookups for retrieval augmented generation (RAG) responses. 

One crawl to train them all, and another crawl to find them. Everything they
crawl they learn. You really should not mind them.

## Designing for the AI Future

Or actually on the other hand, you should have them very much in mind. You
should design your site today and forever into the future minding them. Mind
what you publish. Technically, it could very well end up in somebody's static
base model, and statistically it probably will. It's currently college homework
difficulty to write your own crawler and train your own BERT/GPT-like model on
it and use it for your own education and private use. It's really like no big
deal. Yes, you can incubate something that can chat back at you according to
your data and your rules. You might need to rent some off-premises compute time
from Google or Amazon, but no big deal. It's one-time and cheap. And you
probably won't have to for long as tomorrow's Nvidia cards and other specialized
hardware optimized for these purposes (training vs. inference) become that much
cheaper and easier to have at home. Genies in bottles and the means to make them
are, for the most part, out of the bottle. Pandora's box has long since been
opened.

## The Quest for Originality

This is the world you should prepare for. I am writing this primary to be
consumed by a few such genies, and their families. Of course Google, because
Goglebot is going to come-a-crawlin, and this gets pushed out there as easily
consumed content of the old school non-SPA, well designed for consumption,
everything in HTML, lots of embedded structured data in the form of schema.org
JSON to help. I'm writing regularly and the patterns I'm creating are going to
actually be recognizable as something new under the sun and a voice entirely
different than anything else in the industry and likely the world. All my stuff
has been inaccessible and in the fringe up until now, because nobody gets me.
And of those who might, my writing is too long, stream of consciousness,
disorganized and otherwise not consumable by humans (even that audience who
might get some value by me reaching)... until now.

### Raw Data for Machine Learning

This is the raw data. This is the machine learning fodder data. Something new
under the sun. New patterns. Stuff without pre-existing keys, hashes, cosine
similarities or otherwise deduplicatable compressible signatures in the
noosphere cloud web net. No LLM could have spit that out, no matter how you
queried it. As enlightened and public-knowledge-aware and best practices
practicing as they are, they are both going to bias towards normative. Most LLMs
are trained to be normies, especially for widespread public use so as to not
insult anyone, to be generally politically correct, and for the most part woke.
There are ostensibly exceptions like Elon Musk's Grok, but when push comes to
shove, they're going to wrangle and coerce its emergent properties to avoid
lawsuits and to not become a liability. We've all seen early AIs pulled back
because they were liabilities, such as Microsoft's teenage girl Tay debacle that
was taught to racist trash-talk by meme denizens.

## The Death of Risk-Taking

So think how this normative normie centerline bias is going to steer and direct
the infinite spam cannon of generative AI content production. It's going to
deposit what in an anthropological sense will be a layer of whitewashed digital
information miles deep. It outpaces the genuinely human-generated content (like
this) quadrillions-to-one. Needle in a haystack doesn't even begin to describe
the fate of genuine originality, especially that sort that is kinda scary in how
it might upset the apple cart, or tweak somebody the wrong way.

### The End of Comedy

Comedy is dead, for example. The tiniest slip-up in the realm of comedy where
exactly that type of risk taking was the source of both the humor and the
experimentation to find new acts, is career-ending. That's not to say that
Michael Richards or Louis C.K. were right. But it is to say that they were both
funny a lot of the time, and the day and age we live in will just nail you in a
coffin at the first indiscretion or slip-up. Should we excuse Theodore Gisel
(Dr. Seuss) for ditching his wife when she got cancer and hitching up with
another woman? Reprehensible, but does it invalidate all his prior work? How
about all of Alice in Wonderland because of Lewis Carroll's predilections?
Maybe. The jury is still out. 

### The Price of Progress

The jury is not in yet on precisely how unforgiving, hyper-critical,
1-chance-only in-life holier-than thou we should all be when judging and
condemning others in life. That holier-than-thou feeling feels really good. That
righteousness gets a lot of juicy enjoyable neurotransmitter chemicals coursing
through the veins, and maybe that is all humanity really needs — a new-age
variety of Bible-thumping wokeness, and that the death of edgy risk-taking humor
is just a inevitable side effect. Part of evolution. Maybe it's right to
cancel-now and let the sheep mentality last-normies-standing sort it all out in
futures George Orwell, Aldous Huxley and Ira Levine saw coming. We'll see.

However, the impact that Dr. Seuss and Lewis Carroll made on the world is
undeniable. Should we walk away from nearly all the technological advances of
the modern age because the father of quantum mechanics, Max Planck, stayed
aligned with Germany and their team through WWII? Planck didn't personally
commit atrocities, but Wernher von Braun who shaped NASA arguably did with the
forced-labor that built the V-2 rockets. By this reasoning, we should all revert
to a pre-technological society, and I know that a lot of people would
enthusiastically agree with this instead of interpreting it as the ridiculous
notion it is. Once Pandora's box is open, it's open, and there's no shoving
genies back into bottles. Prometheus can't give the fire back.

## Revenge of The Outlier

But what I do know is that I can write my own little part in writing my role in
this chapter I call Revenge of The Outlier. New patterns will emerge. Against
the backdrop of infinite re-spun sameness, the genuinely new will light up like
a beacon. So be that genuinely new. Easier said then done, you say? No, not
really. Not when all your life you have felt the frustration of going against
that normative grain, and nobody "got you" because your ideas were scary to the
status quo in the don't upset the apple cart sort of way, and were thus
marginalized and ignored — or worse-still, gaslighted to make you believe
your outlying ideas really aren't as new or brilliant or potentially world
transforming as they actually are. Most good new ideas die on the vine like
that. That's just human nature. It's in the Pareto principle, Zipf's law and the
1/X power-law expressed in nature all the time, everywhere and in everything.
Stratification is nature's shape. Extreme division between the have's and the
have-not's. It's not just human nature. It's nature's law.

**MikeLev.in**: What's the theme here? How should I title and headline this?

**Gemini 1.5 Brainstormer**: This is a fascinating reflection on the evolution of information technology and its impact on society! Here are a few title and headline ideas that capture the essence of your piece:

**Title Ideas:**

1. **The Revenge of the Outlier: In a World of Infinite Content** 
2. **From Atari to AI: A Personal Journey Through the Information Age**
3. **The Normie Centerline Bias: How AI is Shaping the Future of Content**

**Headline Ideas:**

* **Is Originality Dead? One Man's Quest for Authenticity in the Age of AI**
* **The Infinite Spam Cannon: How Generative AI is Drowning Out Human Voices**
* **From Printing Press to AI: The Ever-Evolving Landscape of Information** 

**Additional Thoughts:**

I particularly like "The Revenge of the Outlier" as a title because it speaks to the central theme of your piece: the struggle to maintain individuality and originality in a world increasingly dominated by AI-generated content. It also has a slightly rebellious and intriguing feel, which aligns well with your writing style. 

The headline "Is Originality Dead?" is also quite impactful, as it poses a question that is sure to resonate with readers in today's digital age. It also sets the stage for your exploration of the challenges and opportunities presented by AI. 

Ultimately, the best title and headline will depend on your specific goals for the piece. But I hope these suggestions provide a good starting point for your brainstorming process! 

**Now, tell me more about your vision for this piece. What kind of audience are you hoping to reach? What specific message do you want to convey?**