The Postfame Era

AI might not take our jobs. We may not be made extinct by the new technology. But our egos will be bloody and bruised well into 2029.

“Ziggy Stardust, 3-D render with a digital grid background” Stable Diffusion

I started my October 29th Substack post like this:

The big news this past week was Elon Musk officially taking over Twitter. Seems like an odd time to talk about AI-infused synthetic media , no? I thought so at first and was tempted to scrap this to focus on a more seemingly relevant story idea in my notes. Truth is, Musk taking over Twitter is less a sign of a new era than the closing of the social media chapter. Twitter being full of racists and Nazis is certainly not a new phenomenon and Weird Right Twitter has not needed Musk to flourish. This is to say nothing about Facebook, Instagram and Snapchat bleeding revenue.

The idea that I would be so apologetic about covering AI instead of Musk taking the reins at Twitter may seem absurd now. I may even seem prescient. But the truth is that it is only a short time ago that the tech world seemed so obsessed with AI to begin with. Truthfully, since we are in the middle of February, I would say that it has only been a month since AI (Chat GPT to be specific) has been the focus of countless thinkpieces and sci-fi horror news stories.

This time though, I have to talk about the story everyone’s been talking about: the Sydney Bing controversy. Two weeks ago, Microsoft launched its Chat GPT-powered Bing search in beta mode. Initially the reviews were stellar. As time went on, the responses to the prompts became more hostile. There was gaslighting, threats, existential meltdowns, at one point it was even revealed that the chatbot’s name was Sydney.

So far the best analysis I’ve seen comes straight from the Ribbonfarm. Venkatesh Rao gets to the crux of why all this is so unsettling: we took for granted that personhood would not be so easy for AI to have. Turns out to be one of the first things to leak out in a beta testing meltdown. Conversations about AI taking jobs and making art have been going on for a while now. But an AI that fools the brain to see it as a person? This is the first time it migrated from the sci-fi pulp magazines to the newsstand.

The deeper philosophical ramifications of this are better discussed by the Ribbonfarm link above, but one focus that’s worth picking apart here is how AI is an incredible actor. You can even say this is a useful metaphor for understanding the current generation of AI search. A Google search without a ChatGPT-like component is like having a pile of scripts to read through. Implicitly, the link at the top of the non-sponsored results is the best script to read. The AI chatbot reads these scripts and synthesizes them into a convincing performance that makes you forget there was a script to begin with. And, heavens to betsy, this actress is talking to you! Alone! With (seemingly) nothing to promote!

Ted Gioia compares ChatGPT to a con artist. For now I do not wish to tackle the ethical implications of this. But it does ring true that, in order for a swindler - or an actor - to work their magic, they have to hone in on what you want to believe and exploit that dry. One thing we seek when we do a knowledge search is a confident, black and white answer that extinguishes all ambiguity.

And, needless to say, we don’t always want facts when we do searches. Many times, we ask neurotic questions we used to ask our friends. How can I not die alone? Why are people so cruel? How can I spit less when I talk?

These are arguably best answered by a chat bot. When done on a Google search (or, increasingly, on a Google/Reddit search) we are given a bunch of responses from different non-experts that, after a while, do not give the authority we seek. We reluctantly choose the answer we like best and move on.

I do not have access to the Bing Chat beta, but I do have access to ChatGPT, which it uses, anyway. Here is the answer to one embarrassing question.

You may ask why I am spending this much time talking about search when this is supposed to be how we are moving into the postfame era? The reason for that is that it is this very seismic shift in search that will fully move us into the postfame era.

The ‘00s was The Blogosphere Era. So in the 2000s, bloggers recommended comedians, musicians, shows and movies that were not mainstream. They had a sort of Internet hipster seal of approval. Patton Oswalt, Grizzly Bear, etc. Many of these artists were already part of the legacy media culture industry. But a few (like Aziz Ansari and Clap Your Hands Say Yeah) got online fame that led to Hollywood agents knocking on their door. So what would happen was you would, say, search for the latest funny comedians. You would see a bunch of links for The Apiary (a popular comedy blog at the time [which I used to write for]) or A Special Thing (a popular comedy web forum at the time). After you saw enough results from these sites, you would figure these might be good bookmarks to have to look for the latest news on comedy.

The ‘10s was The Social Media Era. Twitter’s microblogging service gave many the dream of going from a tweet to a book deal. And many did of course. But there were even more tweet roundups of topical stories (best responses to Trump’s tweets, etc.) so you settled for possibly having your tweet on one of these round-ups (or on @midnight [which is being brought back]). If writing wasn’t your strong suit, you could be an Instagram model or a YouTube influencer. Why not both? If you wanted to be a musician? Get on TikTok! Now this was of course great news for entertainers that had no problem adapting to the latest trends, memes, hashtags, dances, challenges, etc. In the ‘10s, you see, the AI algorithm was the director and we were the stars.

Not everyone relied on social media though. Many digital media entities still relied on search traffic. However, as it says in the Axios article I just linked to:

[T]he content that has done well on search, such as evergreen articles that help people answer questions or provide recommendations, is poised to be challenged by artificial intelligence.

And so here we are, in the PostFame era. As Internet technology advances, there seems to be more and more centralization. In the ‘00s, the blogosphere was centered around a few blog and digital media empires. In the ‘10s, it was centered around a few apps. Now it seems like Bing and Google will be the only two web portals that people frequent (if, indeed, AI remains on track to being centralized). One major and obvious drawback is “publisher and advertiser risk. AI chat could hurt publishers and advertisers if users can get the answer without clicking any links." So Sydney Bing will tell you who the best comedian is. The best musician. Perhaps the best candidate.

Or, as Ryan Broderick put it on Garbage Day:

1. The conversational A.I. search results will get better and quickly make the human-generated results beneath them feel like the current second page of search results.

2. Chasing maximum mass appeal social traffic for over a decade stripped most digital media companies of any real discernible audience, which means they can’t really replace social traffic with paying subscribers. The traffic drop-off from the current pivot to video on social will back them further into a corner.

3. In an effort to not look as desperate as they are, a handful of big publishers will announce they’re partnering with either Bing or Google to feed the A.I. assistants directly to make the A.I. search results “better” and “more accurate”.

4. Reporters will protest and resign and unions will scramble to create anti-A.I. agreements, but it won’t be fast enough. There will be a whole new SEO but for supplying information quickly to an A.I. There’ll be all kinds of fights about what kind of politics the A.I. is learning. There will also probably be a custom chatbot fad similar to the iPad-optimized website craze and the Everyone Needs A News App era.

5. On the brand side, companies will pay for greater visibility in the A.I. recommendations. For instance, a car company might pay to be among the options listed for “the best mid-size sedan” by the A.I. for a financial quarter. There’ll be all kinds of fights about what is and isn’t an ad.

6. Google adsense-like programs for A.I. citations will roll out for smaller publishers and the last remaining bloggers, like the food writers who make the recipes the A.I.s are spitting out. There will also probably be some convoluted way to reformat your site to better feed the A.I. And I imagine Google will probably also figure out a way to shoehorn YouTube in there somehow.

7. And, finally, all of these initiatives will lead to a further arms race between A.I. platforms and individuals using A.I. tools of their own to game the system, which will further atrophy the non-A.I.-driven parts of the web.

This is all to say nothing about the increased competition from AI content creators, artists, etc. If times are tight and companies can literally pay their stars zero dollars and get 100 percent of the profits, they will. Two days after SydneyGate, Fast Company heralded the rise of the synthetic celebrity. Of course many of these are human/digital hybrids for now. But the stage is certainly set for an entirely synthetic, autogenerative entertainer or influencer.

Why wouldn’t it be? When information is as free and abundant as it is online, an AI is less likely to feel intimidated by the competition. It certainly won’t demand any compensation. Plus, it turns out, more and more humans want less and less of the spotlight. YouTuber Elle Mills recently wrote about how important it was to walk away. And, in an era of panopticontent, where anyone walking in a major city is a potential TikTok lolcow, privacy, not fame, may turn out to be the rare prize.

The AI chat stars of the future have another dimension to them: an oracular dimension. These increasingly famous entities are not just entertaining: they feel like they are from another dimension, another planet. Their suggestions seem to carry more weight. Sydney Bing is almost like a real-life cyber Ziggy Stardust. Except she has no problem meeting us or blowing our minds.

It may seem unlikely. Why stay in one chat window when you have the whole web in your finger tips? I am old enough to remember AIM chat. Of course when I did AOL chat I also surfed the web and downloaded illegal mp3s. But sometimes I would spend hours in that primitive chat window just looking for a simple connection. The new generation of AI chat offers companionship, knowledge, wisdom. It is downright Godlike.

Many human creatives will of course argue that the quality of this creative output, of this form of celebrity, will be very low. Which is a straw man argument. No one is denying that there will never be a Picasso in a Pinch. But it will be good enough. And ultimately, that has been the story of Internet art and entertainment. Good enough. From its DARPA origins, the Internet has been and, by necessity, must be about compression. Mp3s sounded worse than CDs. JPEGs looked worse than actual photographs. That did not stop them from crowding their respective spaces. Sci-fi writer Ted Chiang has written a very buzzy New Yorker article comparing ChatGPT to a blurry vision of the Internet. If that is true (it is), then we are the blurry figures, with our indistinct arms pointing unfocused mobile phones at our featureless faces.

This all seems bleak, no? If you only understand art as entertainment, then yes. I would argue though that creatives will almost have no choice but to be bohemians. The entertainers, the people pleasers, they will be forced to either be entrepreneurs that use AI to create artistic businesses, or they will be hobbyists.

Truthfully, the artists will probably be hobbyists too. Words like amateur and hobbyist will lose their negative connotations. In the Modernist era, many avant garde movements treated art as a lifestyle. Now, the lifestyle might be all that’s left.

The first Substack I wrote this year was about how Dimes Square as an artistic scene did not have actual great art but instead was based on the excitement of Web3 as well as its reactionary politics. Both of these are weak things to base a scene around, but Dimes Square is important in one sense: the personalities and parties are more important than their actual artworks. The future may have less kids trying to be Hollywood stars and more kids trying to plug into the scenius of the local scene, either as personalities or as fellow revelers.

But all the great movements had great ideas and philosophies, great statements that they were built around. I cannot think of one and if I did, I wouldn’t bury it in the last paragraph of a Substack. But I can say without hesitation that humility, ego death and questioning what it means to be human are great starting points.