Getting Chegged
Getting Chegged
One companies misfortunes are a warning for media. But it's not just AI. Media's troubles are deeply generational.
Welcome to People vs Algorithms #67.
I look for patterns in media, business and culture. My POV is informed by 30 years of leadership in media and advertising businesses.
Sometimes it’s nice to read in the browser.
Chegg, a subscription ed-tech company that offers students test prep, tutoring, textbook rentals etc, lost half its market cap this week on reports that a new competitor was stealing market share. That competitor… ChatGPT. It appears students have been substituting a smart free thing that helps them do better at school for a paid thing that offered the same promise. For Chegg, the fuse between threat and financial result was short. The market acted swiftly.
The significance of the moment may be broader than Chegg’s stock performance. From here forward, the term getting “Chegged” can only mean the moment your business gets upended by AI.
The feeling is going around. Many businesses, digital media among them, fear they are about to be Chegged. Even the mighty Google, the stabilizing, monopolistic backbone of the web. Microsoft would love to Chegg Google. OpenAI would too.
And in this week’s hottest tech story, a leaked memo from a Google employee revealed some inside the company fear both they, and OpenAI, are about to get Chegged by more efficient open source AI models that are much cheaper to run, almost as good and faster to evolve. “We Have No Moat, And Neither Does OpenAI,” is worth a read as is this counter position.
From where I sit, OpenAI, Google, Microsoft and Meta will have sustaining advantages in bringing AI services to market, leveraging cloud computing, enormous capital requirements, existing distribution advantages and the necessary commercial infrastructure to profit from them. On the consumer side, power will collapse around a couple of players as it did with search and social. Media will be left with a marginally new set of gatekeepers, same as ever.
Certainly, digital media is feeling Cheggy, though AI is not yet to blame for all of it. The ad market was terribly soft in Q1. It’s not entirely clear where good news will come next. Everybody is slashing costs. Old titans like Vice have fallen. And the shuttering of BuzzFeed News provided an emotional bookend to an era of social media distribution dependency.
Unlike its peers, Meta may have squeaked out a decent quarter, but it does seem the era of “big social” is ending, collapsing in on a single, frenetic vertical-video TikTok inspired black hole. Don’t worry, these companies will fade more slowly than coal power. But the idea that they are even a remotely reliable media distribution channel are long gone. Seems this is the great lesson of BuzzFeed. The notion of "free social distribution" will sit alongside idioms like "the check is in the mail" as a misplaced wisdom too many of us relied on as a north star.
The moment provided a perfect opportunity for BuzzFeed News founding editor, Ben Smith, to visit every podcast I listen to, promoting his well-timed book documenting the traffic-addicted era. His new project Semafor, and others like it, are playing a more old-fashioned distribution game powered by hard-earned email subscribers and live events. If there is any consensus on where to go next, this seems like it.
Now it seems, social media was just a prologue to a much longer and painful story of media disruption. It's AI that will truly confound long held notions of how media is made and distributed and the underlying economics that tie the whole thing together. AI will certainly exacerbate trends that so many established media companies ignored for too long: 1) Attention is finite, content is virtually infinite, consumers learn to cope with this reality; 2) Consumers have little regard for "professional" conventions media companies regard as sacred; 3) YOU, the audience, are the center of this world; 4) Structural distribution advantages are fleeting on the network.
How do media brands avoid being Chegged? This is probably the most perplexing moment I have witnessed in 20 years. Fundamental questions emerge. How will chat impact the flow of traffic from search? How will the consumer attention shift between chat and page content, a vital issue if your business depends on any manifestation of modern digital advertising. Can media brands find opportunity inside of chat models — as trainers, personalities or trust validators? How will IP protections emerge in the new era?
I wanted a 2x2 to provide perspective even if the issues transcend two dimensional characterization. My simplistic attempt:
Left to right we have "open ended" to "structured". Think of the far left as "AI chat," interactions with smart systems that extracts content from a global corpus of knowledge. This is by its essence, infinitely personalized for you. On the opposite end is structured media. This is an editorial page, a crafted piece of video. Digital or otherwise, it involves creating against a creative spark and unanticipated needs.
Along the vertical I've just connected "fact-based" or "evergreen" content on the bottom with narrative stuff that involves human imagination and perspective on top.
Some observations:
The lower left is where tools like ChatGPT have the most immediate impact. Users typically begin this journey with questions. AI can provide answers with respectable accuracy and a lot less friction. This will quickly disrupt search distribution where it is easier to pull answers next to the query than send the user out via a link. It will be interesting to see how consumers begin to acclimate to chat as a replacement for an old way of getting information. Feels troubling for commodity content creators.
Structured responses may persist in areas where complex presentation of information next to a brand creates trust. This too will atrophy as chatbots mature in multimodal information presentation and better information citations to pull brands into its vortex. There are a ton of questions, not the least of which is who gets paid. Needless to say, there will be a compression of sources that here-to-fore found the Search Engine Results Page (SERP) to be a reliable source of distribution value.
This GIF below tells a story. It is a compressed content creation process from a company called Cohesive. To me it demonstrates how futile much of page-based content creation will become. This stuff will never sustain a media business model. Media as glorified prompt writing is pretty low value.
Chat will attempt to mimic more nuanced human interaction in the top left corner. Branded chatbots will proliferate trained on the proprietary data sets. "Yourbrand.GPT" is possible today but not hugely useful outside of utility-first service constructs. Chatting with a chatbot can be highly inefficient and annoying. PI is the latest example, a new personal assistant from Reid Hoffman backed Inflection.AI. PI stands for "personal intelligence." Personally, I would much rather talk to a human. Many brands, media included, will try to make hay here. Most of the efforts, at least for the foreseeable future, will amount to novelty.
The tension points between AI automation and professional / creative industries, as reflected in the current WGA strike, are going to define labor / capital relations for a long time to come. This tension is represented by a pressure to move the bottom left to the top right.
Stories on pages, like movies on streamers will sustain. Chat has its limits. We will continue to embrace thoughtfully constructed content from trusted sources that create culture with originality, filling in new knowledge gaps, rich with human judgment and emotion.
What’s not captured in the 2x2 are profound generational changes in media consumption that are just very difficult to build value around. I was reminded of these this week when I forwarded a NYT article, "How The Legend of Zelda Changed the Game” to a gamer friend who consumes media like a 19 year old.
The presentation of the content was ambitious — a combination of text, imagery, video clips and interaction cutely designed to reflect the interactivity of a video game. Interacting with the elements on the page accrued "progress points." Well executed in a paternalistic, NYT feature-y way. This was the kind of content publishers once aspired to create, a style of multimedia storytelling that could only exist online. Today, few but dominant subscription-funded market leaders could justify the cost to create them. It certainly would not pencil as ad supported media.
But to get to the point…. his response to my share was a one word, <summarize>. I know... rude… very rude.
I recognize that one never try to fit media consumption preferences in tidy little categories, less extrapolate one person's habits across a vast marketplace of behaviors. But there was something more here. Indeed, part of it was an enthusiast's reaction to something written for a nube.
But his bigger issue was format: "I love this shit as you know from a technical and artistic standpoint. But I've never read one of these articles," adding, "If I want to rabbit-hole into a topic, I'll go into the Reddit forums and look at that. Or watch one of the many great YouTube videos that people make to go deep on topics."
Precisely the kind of behavior we see in many younger media consumers, a head scratcher for many in the media business looking to design products that resonate with a new generation. Institutional strategies to profitably capture attention online have proved challenging to even the most enterprising. Ask Buzzfeed. Pressure to make online media business work have only contorted incumbents into service layers that sit nervously on top of Google.
Media values form and structure. He wanted something different, an unadulterated kind of media that has become second nature to digital natives. Straight to the vein.
The point is, generational pressures will collide with AI in ways that will force a complete reengineering of digital media. To me those pressures are characterized by very low signal to noise, informal delivery and impatience for anything that adds overhead, a strong preference for first person contact, and attraction to media that converges practitioners expertise with personality.
AI will collapse much of the value in existing media models, particularly the kind that stuffed Google for the past decade. New models will need AI to automate everything except that which makes human created media nuanced, weird and wonderful. This is an extremely difficult discussion to have inside of a tense union discussion, FWIW.
An ability to automate connections between more and more parts of our digital lives with AI, which one way or another will steadily reduce the pressure on content to monetize with advertising alone. Media that answers questions becomes a front end for transactions.
It’s definitely a moment, one that brings me to a Sam Altman (OpenAI’s CEO) quote from a short blog post some five years ago. Commenting on the inevitable “merge” of machine and human intelligence, he warned: "Although the merge has already begun, it’s going to get a lot weirder. We will be the first species ever to design our own descendants. My guess is that we can either be the biological bootloader for digital intelligence and then fade into an evolutionary tree branch, or we can figure out what a successful merge looks like." Nobody wants to be a bootloader. I suppose that’s like being Chegged. Sage advice.
Have a great week…/ Troy
ON THE PODCAST
The New Interface
The commanding heights of the digital economy are control of the interface. It’s the interface through which the vast expanses of the tangle of data that makes up the digital economy are accessible. Think of the most powerful digital businesses and they are, in essence, interface businesses. Google is the interface for accessing information. Facebook is a social interface. TikTok is an entertainment interface. This week, Troy, Alex and Brian discuss how AI is already changing the interface for accessing information and implication that has for the media business and the digital economy as a whole.
ON DECK
Dick and Karen
A prodigious pair.
Karen Anne Carpenter (March 2, 1950 – February 4, 1983) was an American singer and drummer, who formed half of the sibling duo the Carpenters alongside her older brother Richard. With a distinctive three-octave contralto range, she was praised by her peers for her vocal skills. Carpenter's struggle with and eventual death of heart failure related to her years-long struggle with anorexia would later raise attention and awareness of eating disorders and body dysmorphia and their possible causes.
Carpenter was born in New Haven, Connecticut, and moved to Downey, California in 1963 with her family. She began to study the drums in high school and joined the Long Beach State choir after graduating. After several years of touring and recording, the Carpenters were signed to A&M Records in 1969, achieving enormous commercial and critical success throughout the 1970s. Initially, Carpenter was the band's full-time drummer, but she gradually took the role of frontwoman as her drumming was reduced to a handful of live showcases or tracks on albums. From then on she found her appearance under constant scrutiny and developed anorexia as a way to cope with the massive pressure to look slim on stage.
At the age of 32, Carpenter died of heart failure due to complications from anorexia nervosa, which was little-known outside celebrity circles at the time, and her death led to increased visibility and awareness of eating disorders. Interest in her life and death has spawned numerous documentaries and movies. Carpenter's work continues to attract praise, including appearing on Rolling Stone's 2010 list of the 100 greatest singers of all time.
Source: Wikipedia