When the Noise is the Signal

• May 19, 2026
• Blog
• 5 minutes reading time

I've always loved the signal-versus-noise framework from information theory. James Gleick's The Information was one of those books that genuinely changed how I see the world. It introduced me to the idea that the history of communication technology is essentially a long effort to improve the signal while reducing noise. The digital revolution itself was, at its core, a solution to that problem: hard-coding information in a way that was far less susceptible to degradation than analogue formats. The signal-to-noise concept helps to understand the world around us, and our information environment is in serious trouble, not because of the internet, but because of what's been done with it.

Noise as Product

It's easy to blame the internet for the feeling many of us are experiencing, where everything has gotten louder and more chaotic. But the internet itself isn't especially noisy. The people who actually build and maintain websites – who publish things under their own name, linked to their own identity – are relatively few. If anything, that culture of independent publishing was more alive in the early 2000s than it is now. What's changed is the platform layer: the walled gardens of Facebook, Instagram, TikTok, where you can't read or contribute without being part of the platform. Alongside the digital platforms, legacy media have undergone their own transformation over the same period.

That transformation has a clear architect. Rupert Murdoch figured out, earlier than almost anyone, that noise sells. What people were actually willing to spend time and money on wasn't the signal – the authoritative, considered journalism that newspapers held up as their reason for existing – it was the tabloid stuff. The gossip, the celebrity scandal, the beef, the outrage. He shifted newspapers toward that model through the 70s and 80s, then applied the same lesson to television in the 90s and 2000s. What Murdoch pioneered was that you could build an entire media apparatus around producing noise.

The result is that our relationship with the media has inverted. Noise used to be the thing you tolerated to get to the signal. The classifieds and the ads that paid for the journalism, and so they had a place. Murdoch realised you didn't need to bother – the noise is the product. Headlines exist for their own sake. Opinion has replaced reporting. The metric is engagement, not understanding – and engagement, it turns out, is most reliably generated by things that provoke rather than inform. We see this on a weekly – or more honestly, daily – basis as the outrage machine kicks off another cycle of noise.

A recent episode of the Lamestream podcast illustrated this perfectly. A story that deserved serious treatment – an Australian citizen denied entry to the United States without a criminal record, raising real questions about what's happening to the relationship between allies – collapsed almost immediately into tabloid coverage of the person's girlfriend's opinions. That became the headline. That got repeated across the media landscape, Murdoch and non-Murdoch alike. Because that's how infected the whole ecosystem has become. It's not just the outlets that started as noise machines. It's everything now.

McLuhan said the medium is the message. I'd update that: the medium doesn't matter, there's no message anymore. No one's really trying to communicate anything.

The newspaper is devoid of news. The TV schedule is reality show, single opinions, and panel shows of opinions. We don't even watch any more - our time divided by our second screen and the show becomes background noise. The social apps present an infinite scroll – there is no end, the algorithm always has more. Engagement is the only metric that matters. Engagement? Really? Are we actually engaged at all?

And with AI, this is all going to get worse before it gets better. Generative AI is, among other things, a noise amplification machine – capable of producing plausible-sounding content at a scale and speed that makes the current environment look restrained. The signal isn't going to get easier to find.

Finding the Signal

My own media diet has drifted a long way from the algorithmic mainstream. I'm not on the major platforms. What I do have is Mastodon – non-algorithmic, heavily curated, and slow in a way I've come to appreciate. My whole approach to social media has been about controlling what I consume: unfollowing ruthlessly, treating the feed as something to be shaped rather than something that happens to you. I've found that asymmetric following is underrated – I don't need to follow everyone who follows me, and vice versa. Often, the most useful signal comes not from the original account but from someone else boosting it, which means a human has already done some curatorial work.

There are still people out there looking for a signal in the noise. And I think the way you find it is the same way it always was: through other people. Not algorithms – algorithms just increase the noise. Through actual human connection, which is what the open web, at its best, always facilitates. What made the early internet genuinely exciting wasn't the technology – it was the people behind the pages. Individuals with names and lives and ideas, linked together by genuine interest. People as people, not as sound bites and opinions – but thoughts, feelings and contradictions. With pain and sadness, regrets and joy – so much joy, enough to affect you across the wires and tendrils that connect words and minds.

If we want the signal, we have to realise that it is carried by the people. That person behind the webpage, with a life as complex as yours, trying to make sense of the same world. That's the signal we're actually looking for. And I think we have an innate capacity to tune ourselves to find it, to filter out the rest, if we're deliberate about it.

I'm hoping that capacity doesn't get drowned out entirely. But I think it's going to take some work to protect it.

• Tim Klapdor
• Adelaide, Australia
• posts, Signal, Noise & Media