Let's generate 'cold media,' shall we?
In “Fragment the media! Embrace the shards!” I argued that we need to shift in seeing ourselves not as ‘consumers’ but as ‘generators’ of content.
“Every action you take — every article you read and share (psst!), every newsletter you pay for (ahem!), etc. — is a vote to have more of that thing in the world. It’s a vote for the kind of information ecosystem you want to inhabit. You’re not ‘consuming’ content, you’re generating more of it.”
I’d like to add a second urging: generate cold, not hot media. Let’s dig in.
Marshall McLuhan famously argued that the medium matters more than the content it carries: the medium is the message. E.g. TV has made all news into entertainment. It’s a useful framework, but it begs the question: what kind of message do different mediums offer? Luckily, in an overlooked and forgotten chapter of Understanding Media: The Extensions of Man, McLuhan identifies the effect of a medium. In short, the effect hinges on whether the medium is ‘hot’ or ‘cold’ or as McLuhan writes, “the effect of a medium is in what it omits and what we supply.”
Think of hot and cold as relative to how much participation it requires of you. Hot media is high definition, high information, high stimulation — and so it requires less participation from you to make sense of it. It resolves any uncertainty or ambiguity inherent in the information and offers a directive or buttoned up presentation in its place. Cold media offers low information and low stimulation, and, in turn, requires more participation from the user. It maintains the uncertainty and depends on the user engaging with it. Meaning is made out of this dynamic back-and-forth, not in passively receiving information.
A lecture is hot, a seminar is cooler. A book is hotter than a dialogue. YouTube videos are hot, while Facebook, which requires more active participation (e.g. posting, sharing, etc.) is a li’l cooler. Images are hot, while charts are somewhat cooler. Phone calls, which require you interpret silences or words left unsaid, are even cooler. Texts, group chats, and private forums might be even cooler than that — you have to interpret the pacing of the information, the use of emojis, etc. By contrast, email is warmer than instant messaging. For example, I want Untangled to be a dialogue but it’s daunting to respond to a 2,000 word essay that took gobs of hours to write, research, outline, write again, edit, and then finalize. By the time you receive an issue of Untangled, it’s pretty closed off. I try to balance hot elements (e.g. attention grabbing headings) and cold elements (e.g. an argument or perspective that leaves room for curiosity and imagination), creating just enough space for you to engage the material. Long story short: “…the hot form excludes, and the cool one includes.”
The early days of the web and social media was a cooler time. People wrote, posted, and shared information, which linked to other information, inviting inclusion and exploration. Then, algorithmic systems turned up the heat and started arranging what we engaged with, removing choice, and — let’s be real — flipping most of us from participants to passive recipients, even lurkers. As McLuhan writes, this is problematic because hot media “engenders specialism and fragmentation in living as in entertainment” and contends that “Were we to accept fully and directly every shock to our various structures of awareness, we would soon be nervous wrecks, doing double takes, and pressing panic buttons every minute.” Cool media disburses this shock by allowing us to explore and learn at our own pace.
Now, generative AI — when applied to both search information and images — is turning up the heat. As I wrote in ‘Will ChatGPT replace search?’, the authoritativeness of a direct answer coming from an information retrieval system poses a problem: the convenience of the direct answer collapses the motivation to do more research — and exploring, researching, and piecing together an answer to a query is an important part of understanding complexity and nuance. With a direct answer, we are meant to internalize the message, rather than get curious about it. We become passive recipients, not inquisitive. As Google AI Overview and startups like Perplexity embed these tools in their products and remove links and citations, even less engagement is possible.
This hot-cold framework is helpful, but it leaves a lot out. For example, McLuhan was writing at a different time, when the output was all that really mattered — producers, writers, or creatives either made, wrote, or created something. But our information environment has evolved in a way that swirls creators and audience members together. Both groups create, both groups consume. And, with the addition of data-driven technologies, such as LLMs, the training data and model weights we select all materially shape the outputs. You maybe rolling your eyes and thinking ‘Who’s we? Companies will never let us engage in this way!’ But they kind of already have. The results are imperfect and messy, but promising.
Divya Siddarth and the Collective Intelligence Project recently ran ‘alignment assemblies’ with OpenAI, Anthropic, and other companies. Basically, they showed how generative AI can be retrofitted into a participatory process from the jump, even if it’s outputs nurture a passive acceptive. CIP engaged a representative group of Americans in drafting a constitution — basically a list of principles — used to fine-tune Anthropic’s large language model, Claude. They found that “public model was less biased across the board, but just as capable at core tasks, as the researcher’s model” and further that “For every divisive statement, there were a hundred statements with near-total consensus.” In other words, it wasn’t a total cluster — people agreed more than they disagreed, even on fraught issues, and Claude is better because of it. (Whether Anthropic continues to use this version of Claude or any company adopts the recommendations of a participatory process is up to the company, and participation without power only gets you so far.
McLuhan was also writing at a time when there was a clear distinction between media producers and their audience. But those lines are blurry these days, and many of us — myself included — wear multiple hats at the same time. If we lean into the blurriness and relax our assumptions about expertise and credentials, we might come up with different ways of engaging people. For example, as Deepti Doshi explained in our conversation, New_Public is training ‘digital stewards’ in cultivating healthy local information environments. These are people who already spend tens of hours each week curating and sharing local information on messaging boards or in private groups — e.g. anyone who runs a subreddit.
If you’ve spent time on a messaging board or in a private neighborhood group, you know they can run hot. People often talk at (not with!) one another, and they can quickly devolve into shouting matches and name calling. These stewards don’t have formal credentials, and are people that have been doing free labor online for years because they care deeply about their respective online communities. Now, New_Public is offering them a small stipend, training, and space to co-create best practices with peers. Do these stewards replace trust and safety teams or local journalists? Absolutely not, but if we shift our assumptions about what’s needed to dispense critical local information, digital stewards might fill an important gap in the broader ecosystem.
Last, McLuhan was writing at a time when there wasn’t so much variation within mediums of the same category. Most TV was the same, most books, most radio etc. As a result, the kind of engagement was roughly the same too. But the internet and social media has blown that assumption out of the water. Commenting on Facebook is different from commenting on an Untangled post. The former encourages reactivity or reflection while the latter is an ABSOLUTE DELIGHT. You get the point — the same kind of participation, when mediated by different platform cultures and affordances, result in different kinds of participatory experiences.
We’re living in a hot media environment that is only getting hotter. But we can stop listening to the confident sounding blowhards who project a hollow confidence and instead embrace humility and uncertainty. We can engage with our neighbors with curiosity, and make meaning in the back-and-forth. We can support media platforms that invite real dialogue and create cultures that invite more questions than answers. We can follow and promote creators who hold space for ambiguity and not-knowing. There are actually many paths for participation now, but they are perhaps less obvious. We can shape the message of the medium.
Comments