Friction
Welcome to People vs Algorithms #64.
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.
Note: Monday again? This is not a Monday thing. This can’t be a Monday thing.
Barry Diller has me thinking about friction. The stuff that exists between intention and reward. Friction is a counter force that slows you down on the way to getting what you want, mostly for worse, but occasionally for better. Oddly enough, media depends on it. Not the kind of friction that makes for a good story. You need that too. But the kind that makes the business of media function. Like advertising. We never really had to worry about this prior to the internet because media lived happily in highly-controlled friction-filled delivery channels where people never really had a choice. But the internet changed that, pitting media against friction-hating technologists, with the consumer in the middle as a pleasure-seeking friction-avoiding optimizer.
The internet dislikes friction. It prefers utility. It’s defining feature is a “link.” Links connect everything together. Links are flow. Flow is the opposite of friction. Entrepreneurs use technology to remove friction wherever it is profitable to do so. Like creating platforms that connect the things that people like in one easy package — sorting through and consuming tons of content from a lot of different people, doing useful things around it, talking about it or making your own content. These kinds of tools remove friction for people and, as a consequence, grow very large. This creates endless friction between media and the platforms who now have all the power.
And now we have AI. It is the ultimate media friction remover because it's a layer that sits on top of everything, with jaw-dropping new capabilities to remove all the friction for a consumer. This is creating new battle lines between media and tech, ones that make the last ones look modest in comparison. Naturally, this is about to create plenty of new friction.
This time, Barry want's media companies to create much more friction in the courts before the AI's remove too much friction and consumers start to get accustomed to it. This is probably a good thing for society. Reflection is useful at a time like this. But it is unlikely to change things much because friction removers usually win.
They win because humans like to avoid friction. They like to engage in interpersonal kind, but they dislike the kind that makes doing things harder. The elimination of the bad kind of friction is a nice way to endear yourself to people. Which is why it is at the root of most product innovation. "We make X simple," or "X makes life easier" are core capitalist aphorisms related to friction removal. New friction removers replace old friction removers. In most cases, the one that destroys the most friction wins. Sometimes, we tolerate friction or even introduce it into a process because it forces us to savor or reflect. Sometimes friction is romantic. A wine cork has more friction that a screw cap. But this should not undermine my argument.
The modern internet economy is filled with examples of enterprising friction removal. Like how Uber did a admirable job of removing the friction of getting a car service. Or how Instacart removed the friction of filling your fridge, replacing schlepping to the store, walking aisles and filling a cart with scrolling and button pushing and answering the door when the guy arrives.
Elon wants to turn Twitter into X, the “everything app" which is essentially about removing friction between complimentary use cases. He has taken a couple of steps in this direction recently, first by allowing long Tweets, video and content paywalling which puts Twitter in competition with Substack, the company irritating Musk with its new Twitter-like "Notes" functionality that attempts to remove the friction between long form content (like this newsletter) and short missives that you find on Twitter. BTW, Twitter would die faster if it wasn’t for the friction of network effects. Next Twitter will inch toward integrated stock trading by partnering with eToro to remove the friction between reading a Tweet and buying a stock or crypto. The internet loves doing things like this.
Facebook removed the friction of reconnecting with your aunts and uncles and people from high school. And then pulled all the content they like into one infinite, frictionless stream. At first, Facebook’s idea of “bringing the world closer together” seemed like a good application of friction removal, but in retrospect it created new friction and was not great for people who made media and felt compelled to compete with others to feed free content into the social distribution machine. Barry does not want this to happen again. But it will keep happening because people hate friction.
Instagram founder, Kevin Systrom, hopes to remove friction with his new Artifact app, another in a long line of applications that gather all of your content interests together in one place, leveraging a fancy AI algorithm to ensure you don’t encounter friction when you want to read just the right thing. Now the app wants to remove the friction of discussion around content by pulling the commenting experience together in one place so the water cooler and media come to you in one convenient package. Platforms always like to remove the friction between content and conversation because talk is cheap and people like to talk to other people. BTW, his app exposes all of the junky friction in the internet media machine when you move from his smooth, ad-free venture-funded experience into the ad-pocked teenage face that is the reality of digital media, which, sadly, has never quite matured beyond its awkward economic puberty.
But friction removal is as old as the wheel. Newspapers replaced the friction of town criers and brought the news to your doorstep. Magazines removed the friction of having to explore vertical interests inside of mass market newspapers. Computers removed the friction of paper, physical delivery and space and time constraints. Phones removed the friction of immobility. Social removed the friction of sorting and retrieving news. Notifications removed the friction of having to seek stuff out.
AI is the ultimate friction remover because it does all the work for you. It is brutal in this regard. It does not ask why. It does not judge. It does not remember. This week we started to see some wild cases of how AI might remove friction in alarming ways.
AutoGPT is a new friction remover that uses the OpenAI API to string together multiple commands and add “memory” to the AI use case (Github repo here). Now your AI can do all kinds of things autonomously in pursuit of a goal without needing to be prompted at every step, which pretty much is the definition of a good human assistant. Just days later, seeking to highlight dangerous applications of this technology, an enterprising developer released ChaosGPT whose goal is "empowering GPT with internet and memory to destroy humanity." Don't worry, this is not going to work for a while. OpenAI's prompt sniffer thwarted the nefarious requests to its model. Evil AI's don't tire easily, so it sought to spread negative vibes on Twitter, even though this is not so effective anymore. But cray. (Note: to get a sense of the breakneck speed of new innovations spawning from ChatGPT, here’s a week’s worth.)
Now, even one of the "All In" podcast guys is asking for immediate regulatory intervention which is confusing for everyone. Don't worry, it was Chamath Palihapitiya and not David Sacks. The discussion created some friction on their popular podcast, mostly because nobody really understands what regulating AI really looks like. After all, you can run the stuff on all kinds of hardware, unconstrained by geographic legal boundaries and it's unclear at this moment at least how it might hijack anything other than our information space, which I suppose is worrying enough. But a regulatory response will come soon. Thoughtful people like Tim O'Reilly are starting to put forward ideas with as little negative innovation friction as possible to establish standards around AI accountability, transparency and bias, a seemingly herculean effort for a country that does not yet have coherent national online privacy legislation:
So too with AI safety. What we need is something equivalent to GAAP for AI and algorithmic systems more generally. Might we call it the Generally Accepted AI Principles? We need an independent standards body to oversee the standards, regulatory agencies equivalent to the SEC and ESMA to enforce them, and an ecosystem of auditors that is empowered to dig in and make sure that companies and their products are making accurate disclosures.
But if we are to create GAAP for AI, there is a lesson to be learned from the evolution of GAAP itself. The systems of accounting that we take for granted today and use to hold companies accountable were originally developed by medieval merchants for their own use. They were not imposed from without, but were adopted because they allowed merchants to track and manage their own trading ventures. They are universally used by businesses today for the same reason.
So, what better place to start with developing regulations for AI than with the management and control frameworks used by the companies that are developing and deploying advanced AI systems?
If the regulatory response looks anything like GDPR with its incessant website tracking notifications, I would happily opt for complete subjugation to the robot overlords. In my mind, the only thing that's gonna counter bad people getting access to frictionless armies of AI is to have good people sniffing them out with better AI, fortified by smart AI-powered security systems. Regulators will necessarily apply broad accountability and moral strictures on these systems and the companies behind them, but my money is on the AI powered swat teams inside of government and nextgen cyber security industry paid to protect connection points to vulnerable systems. Is all of this even worth it? It’s impossible to tell from here. Our inherent love of technology and the non-linear potential in areas like climate change and health, not to mention the threat of others using it against us, will justify our enthusiastic embrace.
But back to media. Barry has wisely suggested that media companies not fall into the next trap and surrender valuable IP to tech companies who seek ultimate content commoditization inside of the next generation of delightfully friction-free consumer applications. I want this too. But appreciate that this is not a battle against the technology platforms, of which AI is the latest and by far the most disruptive, it's a war against the anti-friction forces that motivate people to embrace the next thing. We love tech especially when it does all the work for us.
Modern digital media is filled with friction that runs counter to everything modern technology seeks to erase. Disruptive advertising is a massive force of friction. Content discovery is too. The more content we make, the more we advantage the tool makers that streamline creation, access and friction-free AI-powered consumption. Like TikTok.
Subscription seeks to remove friction with economic alignment between the creator and consumer. It does this fine until everyone wants to do it and the management of subs becomes an overwhelmingly friction-filled feat for the consumer. Platforms solve this with the bundle. AI simplifies by abstracting what matters to you. Media vs friction.
Robert Oppenheimer, the father of the atomic bomb, deliberating on the morality of technology at a very decisive moment, undoubtedly trying to make himself feel better, commented "Technology happens because it is possible.” Sam Altman seemingly found comfort in this idea too, comparing OpenAI's efforts to pioneer AGI (Artificial General Intelligence), with the ambition and moral turpitude of the Manhattan Project. But if technology happens because it's possible, it is adopted because we like it. We like it because it makes our life easier. It makes life easier by removing friction. No amount of legal or policy counter friction is going to make things like they used to be.
To see the shape of the future, start with abundance. Abundance is also the opposite of friction. Tools of creation will enable a quality of storytelling across all mediums, from text to video to gaming, empowering anybody with imagination. AI is not going to invent the next Succession but it will take a lot of friction out of bringing narratives to life. Abundance will suffocate advertising that attempts to put friction in front of unaligned consumer intention. And, the best tool makers win because the future is participatory. Tech platforms will continue to triumph as the glue that holds a democratic future media ecosystem together.
Barry suggests we cannot make the same mistakes we made last time by letting tech win. Fighting the tide is one way to meet the threat. While we do that, in the long run learning to surf is a much better idea, even if the waves feel overwhelming right now.
Have a good week.../ Troy
ON THE PODCAST
How to do ideas
This week we discuss how to generate ideas and take them through to execution. Brian and Alex are great at this in different ways. I wanted to understand how they do it. We break down the phases, from thinking to making to editing. Brian advocates for procrastination.
Coming later in the week, more on friction. Click to listen.
ON DECK
GenX new wave nostalgia
XTC was a big part of my Canadian commonwealth musical upbringing. This sweet song surfaced this week and it always kills me. Enjoy.
XTC were an English rock band formed in Swindon in 1972. Fronted by songwriters Andy Partridge (guitars, vocals) and Colin Moulding (bass, vocals), the band gained popularity during the rise of punk and new wave in the 1970s, later playing in a variety of styles that ranged from angular guitar riffs to elaborately arranged pop. Partly because the group did not fit into contemporary trends, they achieved only sporadic commercial success in the UK and US, but attracted a considerable cult following. They have since been recognised for their influence on post-punk, Britpop and later power pop acts.
Partridge and Moulding first met in the early 1970s and subsequently formed a glam outfit with drummer Terry Chambers. The band's name and line-up changed frequently, and it was not until 1975 that the band was known as XTC. In 1977, the group debuted on Virgin Records and were subsequently noted for their energetic live performances and their refusal to play conventional punk rock, instead synthesising influences from ska, 1960s pop, dub music and avant-garde. The single "Making Plans for Nigel" (1979) marked their commercial breakthrough and heralded the reverberating drum sound associated with 1980s popular music. More»
Source: Wikipedia
Also: a recent interview in the Guardian. “‘My dream had died’: XTC’s Andy Partridge on mental illness, battling the music industry and losing his muse”