When The Internet Becomes Chat
AI will fundamentally change the structure of the web. That's a good thing.
Welcome to People vs Algorithms #58.
I look for patterns in media, business and culture. My POV is informed by 30 years of leadership in media and advertising businesses.
I haven't read famed producer Rick Rubin's new book "The Creative Act: A Way of Being" and probably won't. But it's fun to listen to him ramble on about his Buddhist like dedication to creative process, among other interesting music things. Rubin is doing a remarkable amount of press lately, most recently stopping by Ezra Klein's podcast. It's worth a listen.
Early in the episode they talk about the things that emerge in our world, things that "unfold," simultaneously (like say, punk rock, with Ramones in the US and the Clash in UK) because for one cosmic reason or another, people tap into the same energy, at the same time: "If you have an idea you are excited about and you don't bring it to life it's not uncommon for idea to finds its voice through another maker, this isn't because the other artist stole your idea, but because the idea's time has come."
As I listened, it occurred to me that the same thing is happening to the internet right now. AI breakthroughs are leading to a broader reconsideration of how we create and navigate information. Ground zero for that change is the spine that holds the internet together, search. Google's near monopoly of that position, is the obvious nexus of our attention. The consequences to its business are enormous, as is the ripple it will cause to businesses that rely on it for oxygen.
I suspect you feel it too. The internet has become a cluttered mess of misaligned incentives. The pressure is mounting to evolve at a structural level. AI is creating a crack, and the economic incentives to disrupt Google's position are tantalizing. Especially for Microsoft. Having dumped billions and years into their search competitor, Bing, only to pick up a few points of market share, Satya Nadella and the teams at Microsoft are giddy about the potential of this moment. "From now on, the [gross margin] of search is going to drop forever," was the money line from a Financial Times interview, noting this moment was the start of a new "race" in the search market that would upend "the largest software category on planet earth."
Microsoft, and everybody else, is hard at work figuring out what AI means to their business. But the search part is of particular interest, because it is the primary choke point and control center for the internet and for trillions of dollars in commerce that flow from its network of links. As goes search, so goes the internet.
Indeed, AI is much more than just a cool new chat thing that sits on top of the old way of searching. It will force changes in how we interface with content and applications at a fundamental level. Like in 1974, when Joey Ramone walked on stage at CBGB, the audience is about to experience something new.
The value chain NOW
Let's consider how things work today. You want to get information on something. You open your browser. The sequence that follows looks something like this:
Search (Almost always Google, unless looking for a product on Amazon.)
Result (Via Google's Search Engine Results Page (SERP).)
Link (The monetizable lynch pin of all web things.)
Page (The canonical thing at the end of a link that the world creates to say something or sell something.)
Transaction (What the page creator hopes you do to pay for everything.)
We are well trained on the routine. At it’s center of the system is the Web Page. It is the foundation of the open web. Feeds and proprietary formats (Instagram posts) are Pages in the app world, but lets leave those for a bit. Someone has created that Page for a reason, perhaps out of the human need to share, express and connect, but more likely for commercial reasons. The Page is usually trying to sell you something. Like a car or car insurance. If that Page is "media", it endeavors to sell you something on behalf of someone else.
Of course, if there was no commercial motivation to create the Page, there would be far fewer of them. People and companies need incentives to create Pages. We feed Google our Pages. Google feeds us traffic in return. Traffic feeds us.
The value chain NEXT
So what happens when AI takes up more and more space inside of the search starting point. This week both Microsoft and Google showed what that might look like. Google’s new AI thing is called “Bard” and sometimes it makes mistakes as all of this stuff does. Hey, so does the internet.
It looks like this (here’s a link if it doesn’t render):
This new behavior will start to replace more of what we used to do. NEXT we have:
Ask (Less a “search”, more a “question”)
Answer (The machine surfaces a custom response, you ask again, clarify, refine, repeat.)
Citation (AI provides reference points inside of the Answer above. Citation is a new way of saying "link” and an important hedge in inevitable instances when the answer is inaccurate or subjective, which it will often be. See how Bing is handling these here.)
Page (The place you go when you need more than the Answer. Feels increasingly clumsy next the the frictionless and personalized back and forth of a chat exchange.)
Transaction (Same as above.)
Inevitably, the NEXT value chain pushes energy and attention from Pages to Answers. The sources that support the Answer, let’s call them Citations, still proffer downstream traffic, but this stream ebbs. The obvious consequence, search becomes a less reliable source of Page traffic for everyone.
And, something else happens.
The chat interface becomes a far stronger point of gravity in how we use the internet and, with that force, more of the experience gets pulled into its orbit. After we have conditioned the user to ask a question and receive the response right there and then, the next step of visiting a web page feels awkward and discontinuous. The natural UX inclination will be to pull subsequent steps closer to the existing user Q&A / chat construct. Smart chat becomes the interface to much more of all of our digital experiences. The internet begins to conform to its shape. You see the same energy in Discord, Reddit, Snap and plain old texting.
So, let's say I want to find out what to do in New York this weekend. In the old days a Google search would yield a list of sites, the user would engage in the usual internet hunt and peck to find inspiration and make a plan.
I tried the same use case above on Bing's new Open.AI powered chat feature that sits alongside its search engine. The AI responded immediately with a respectable set of choices, including Free Fridays at MOMA. I asked it to tell me more about Free Friday. It responded with detailed information, museum hours, ticket reservation info, etc. It conveniently offered Friday night activities at the Met as an alternative.
As the tech matures, one would expect further queries that connect the entire use case. Reserve tickets by typing "yes, let's go." "Put the results in my calendar" and "send Jenny a note to let here know when we are going" and "notify everybody an hour before we have to leave," are all natural follow ups. Then "find a good Italian restaurant nearby." You get the idea.
Let's speculate as to the consequences for for media and beyond.
Interestingly, what the internet did to media in the first wave, it does once again, in ways that are far more profound and destabilizing. In the first round, the internet pulled out media’s spine, robbing it of its curatorial value proposition. Articles become detached from issues and more disconnected from the brands they represent. Distribution became a hit and run affair, largely dependent on a headlines ability to drive attention inside of a feed of mismatched media things. Media became a search-driven fishing expedition. Economics were shifted from the value of an audience inside of a branded bundle of content to click value at the article level.
The next wave fragments media at a semantic level.
Pages will still exist but will be far less important. Content sucking bots will pull the relevant stuff to a personalized narrative response. More time will be spent in Q&A mode than Page mode. The easy stuff, specifically data driven feeds, have already been absorbed into the search feed today inside of Google’s Answer Box - weather, sports scores, product listings, movie times and so on. How-to's, product reviews, travel content, health information will follow. Google would have and could do far more of this already, baring the economics of their click economy and pressure from dependent entities that fall below the search query.
Now, the dam has broken open.
Surely in the future a content creator can opt out, erect a robot fence around their IP. Once again, they will face tough choices — the promise of being referenced as a Citation will entice traffic hungry publishers into the next Faustian bargain.
Video will become an even bigger part of all medias communication repertoire, because video cannot be flattened into a semantic pancake by the AI. Video holds it's position as the last surface for a brand advertising message.
More and more of what was once page level functionality will migrate in a component-like fashion to the "AI chat box." Video players, basic forms, product SKUs, purchase flows and so on. Functional components will swirl around the chat experience powered by new API integration frameworks. Search Engine Optimization morphs to Chat Bot Optimization.
The dominant chat position will be deliriously powerful, perhaps even more than Google is today. Except, open frameworks will enable more people to roll their own AI chat experiences. Chat will become a far bigger part of every interface, to content, to shopping and utilities. Vertical solutions will emerge underpinned by protected data sets. But, distribution power laws will still rule. We will have a couple of big ones with brand power and, vitally, the ability to manage complexities of monetization.
The cost of the computationally intense AI query will force hard decisions on Google, Bing and others, who will introduce new subscription offerings and push hard to innovate new advertising solutions to defer the expense. One way or another Google’s reliable “ATM” will come under severe pressure. The point here, even more subscription clutter for the consumer, more pressure on media companies looking to sell them.
The surface area for advertising will shrink dramatically, at least the way we used to think about it in terms of graphical units, measured in viewable pixel area. Advertising will still be important as it always will be, but as an integrated part of the human / machine back and forth, measured only by its transactional value. RIP old fashioned banners.
We can separate navigational scenarios online roughly as "search" and "social." Both share something that will define the future of our interaction with each other and machines. Algorithmic driven content, interfaces that shift from clicks and pages to a back and forth interactive volley with machines that do more and more to shape and surface deeply customized content, accomplish tasks, automate routines, and step us closer to a transaction. The future looks like the past but far more personalization; feeds with streams of content, ever more tightly connected to communication, the machine doing more of the downstream work for us.
I do not mean to be hyperbolic here. Change takes time. Journalism on Pages will exist, advertising will always morph to suit the next frontier. But a lot changes when 1) the starting navigation point shifts fundamentally, 2) more of the downstream human searching activity is replaced by automation, 3) the financial incentives of creating pages of content for an old system atrophy.
As Rick would say, "As soon as a convention is established, the most interesting work would likely be the one that doesn’t follow it."
Have a great weekend…/ Troy
Uniqueness, differentiation and building the next media brand
We discuss “uniqueness” and what it means as to media as the algorithmic age moves into new territory. Good stuff, I promise. Listen here.
The Clash, unfolding…
“When we came to the U.S., Mick stumbled upon a music shop in Brooklyn that carried the music of Grand Master Flash and the Furious Five, the Sugar Hill Gang...these groups were radically changing music and they changed everything for us.”
The song was inspired by old school hip hop acts from New York City, like the Sugarhill Gang and Grandmaster Flash & the Furious Five. Rap was still a new and emerging music genre at the time, and the band, especially Mick Jones, was very impressed with it, so much so that Jones took to carrying a boombox around and got the nickname "Whack Attack".
"The Magnificent Seven" was recorded in April 1980 at Electric Lady Studios in New York City, built around a funky bass loop played by Norman Watt-Roy of the Blockheads. Joe Strummer wrote the words on the spot, a technique that was also used to create Sandinista!'s other rap track, "Lightning Strikes (Not Once But Twice)". This white rap single is earlier than Blondie's "Rapture" by six months.