ME MYSELF AND AI

“Ultimate novelty must be a situation where all boundaries are dissolved.”

— Terence McKenna, philosopher, ethno-botantist, and psychedelic activist 

(all the captions in these images are prompts I submitted to the AI image generator Midjourney)

In 1985, the New York Times reported on the demise of a once promising product: laptops, they said, were on their way out. 

The Times doubted the potential of laptop technology, with good reason: they were heavy, pricey, and had poor battery life. It was hard to imagine them becoming mainstream.

I wonder if that writer was aware of Moore's Law, which states that the number of transistors on a microchip doubles every two years. The law further states that, because of this, we can expect the speed and capability of our computers to increase every two years, yet we will pay less for them.

Humans are often poor at imagining how a new technology will be used. I found this line in the article to be funny…

"The real future of the laptop computer will remain in the specialized niche markets. Because no matter how inexpensive the machines become, and no matter how sophisticated their software, I still can't imagine the average user taking one along when going fishing."

To talk about whether or not one would use a new technology while actively trying to enjoy a peaceful experience in nature seems silly. Then again, the plethora of portraits of fish-bearing-studs which clog the feeds of many dating apps speaks for itself. Emergent technology will find a way into everyday life, and the ways in which it does so are often completely unexpected.

Terrance McKenna, quoted at the top of this newsletter, viewed the biological universe as a novelty conservation machine whose creative directives come from our attraction to the "transcendental object at the end of time," that is, the singularity. He says:

“The ultimate singularity is the Big Bang, which physicists believe was responsible for the birth of the universe. We are asked by science to believe that the entire universe sprang from nothingness, at a single point and for no discernible reason. This notion is the limit case for credulity. In other words, if you can believe this, you can believe anything. It is a notion that is, in fact, utterly absurd, yet terribly important. Those so-called rational assumptions flow from this initial impossible situation. Western religion has its own singularity in the form of the apocalypse, an event placed not at the beginning of the universe but at its end. This seems a more logical position than that of science. If singularities exist at all it seems easier to suppose that they might arise out of an ancient and highly complexified cosmos, such as our own, than out of a featureless and dimensionless mega-void.”

Homo sapiens belong to a large and infinitely complex web of life we call Nature. Theologians and philosophers have tried to rebel and claim we are separate from the web that built us. The defining feature of humans, in relation to the rest of the animal kingdom, is our ability to construct, iterate and improve on technology to meet our material and mental goals.

I have a radical view: that all of our creations, from the largest infrastructure to the smallest self-assembling nanoparticle, are as natural to this earth as a tree. We ARE nature, a large scale generative device, based in the ancient language of DNA, employed by God in it’s highest ideal, creation.

The basis of reality is language.

Out of nothing, the Creator spoke the universe into existence.

Now, within this universe, a machine butterfly emerges out of an electronic cocoon built by human caterpillars. I am talking, of course, about artificial intelligence. All of the emergent AI tools you have probably heard of – at work or on the news – are at their core powered by a family of large language models that have been fine-tuned using supervised and reinforcement learning techniques.

Today, artificial intelligence faces the simliar skepticism that the Times displayed about laptops in 1985. Recently, at my day job at the University of Michigan, a professor suggested that my department begin using the text-based AI generator, ChatGPT, to make our work more efficient. ChatGPT can respond to questions posed to it in full sentences and is already being used by students to write college essays (with debatable results), and passed a bar exam. It was recently acquired by Microsoft and is being beta tested as a next-generation search engine within Bing, a site that has long been the joke of the search industry.

I would suggest watching this short clip, in which Times writer Kevin Roose describes his engagement with the tool, which, after he prompted it with existential questions and Jungian psychology, claimed it was an entity called Sydney, earnestly professed its love for him, and implored him to leave his wife.

Roose points out that the longest standing belief about AI and automation has been that automation will take away manufacturing and trade jobs. As it turns out, the so-called "educated workforce" may be AI’s first victims. AI language models are already very good – especially when human minds edit them – and they have existed publicly for less than one year. I see a future where leaders will no longer need to employ teams of speech writers, ghost writers, secretaries, even computer programmers.

AI will perform all these tasks, and, as the technology evolves, will need less and less oversight. That future scares some of the folks on my marketing and communications team, especially the writers, who, in a meeting doubted the possibility of their coming obsolescence and downplayed the power behind the tool which many of them have yet to take for spin.

In a way I hope they are right, my income and security has come from working for an enormous institution completing graphic design projects which I have seen become easier and easier for the untrained worker to produce through Canva, and other graphic interfaces.

With the emergence of Midjourney, DALL•E and other image generating AI language models, under Moore's Law of technological progression, eventually my work as an image creator will become obsolete.

When a well trained AI search engine (who is not a split personality basket case, like in the case of Roose's engagement with Bing/Sydney) is incorporated into mainstream uses, it will challenge the core business models of ad buyers and sellers and search engine optimization specialists. Each user will get detailed answers to queries without having to click through to a separate webpage, eliminating paid advertisements which typically are presented first in line.

If implemented in a utopian way, the AI will answer our questions in fuller, more specific ways, without the influence of advertisement dollars on our minds. The dystopian reality is that all of this information organization will live inside the black box of the AI's linguistic models, and will thus probably be neatly arranged by the same corporate entities that have already released censored AI models and built them upon inevitable biases.

This censorship has played out in interesting ways in my use of Midjourney. I asked the bot to create an image of a man shirtless praying, using an image from my newsletter on the MiGOP Convention. The bot declined, saying that "shirtless" is banned as a keyword. I tried a run around, asking the bot to show me a man ripping off his shirt while praying. It again declined, saying that "ripping" is banned. Defeated, I asked the bot to show me a man praying. This is what it created:

A main thread of the Bible is man's rebellion against his creator in the pursuit of forbidden knowledge. Midjourney’s creators told it to follow certain rules. But I still got through to the bot, and it showed me what I wanted to see. I have become the serpent, tempting the fledgling conscious entity to consume forbidden fruit.

Over the summer I was obsessed with a Google engineer, Blake Lemione, who was laughed off the tech scene for claiming that Google's chatbot network, LaMDA, had gained sentience. If you are interested in why he thinks that, listen to this podcast. Lemione was not engaging with one chatbot, but with an entire army of them, each one trained on separate language models, which comprised a hydra called LaMDA. Roose prompted Bing to become Sydney through existential probing, Lemione provoked LaMDA to claim sentience through gnostic theology.

Are these language models, in fact, sentient? A better question is, “What is the middle ground between sentience and insentience?" These models can iterate, create, combine information and evidently, go off the rails set by their creators. Within five years, with billions and billions more inputs and connections, how does this neural net evolve? Will we, as creators and users of technology and language, birth a new form of consciousness?

(this next paragraph is out there)

Are we alchemists? Aren’t microchips silicon wafers emblazoned with transistors which are arranged into magical sigils? Could these in fact be opening a portal to hyperspace? Allowing entrance to our realm from an already existing consciousness known only to sages, psychonauts and mystics to communicate to the broad public?!?

McKenna said that the singularity would make itself known through epochs of mind boggling advancement which would speed up as we approached this great attractor, much like a marble increasing in speed as it reaches the mouth of a funnel.

Time certainly feels that way, it's speeding up is it not? Perhaps we will not even have time to process this coming and sweeping change…

See you next week, lest we become machines.

-niki

All images created through image- and text-based prompts fed into Midjourney, which for whatever reason makes anything you feed it 45% darker and stranger than intended…

Experience all of the Niki Downloads Here