If my eyeballs are being resold to advertisers then it had better be worth my time

20.07, Friday 28 Jul 2023

Ok let’s say you’ve got two physically identical factories. One is doing better than the other – how do you characterise the difference?

The delta might be Marx’s concept of general intellect (Wikipedia): a combination of technological expertise and social intellect. The machines are the same, the processes are the same, the inputs are the same. What varies is how well people get to know the quirks of the tools and develop their skills of cooperation, and that overall ability is something we get for free as humans because we are social, intellectual animals.

(Look I don’t understand Marx very well, I know. I’m cribbing badly from an amazing talk I saw years ago by the philosopher Adam Arvidsson – here’s a draft of chapter 2 of his book The Ethical Economy (Amazon): Ethics and General Intellect.)

General intellect - people’s skilled cooperative productivity - can walk out the door so one way of reading industrial history is that you make the processes and tools so specific that it is bound to a single factory.

And actually there’s always this tussle over who owns the value of general intellect - capital or labour, right - and you can see internet platforms in a similar way. What was the value of Instagram when it sold for a billion dollars? It wasn’t the tech, and it probably wasn’t the ad dollars (although Facebook realised that was a way of extracting value): it was this community of practice that was effective at growing itself. You can’t build that directly, maybe you shouldn’t be able to own that, and yet that was what the founders sold in the acquisition. (I wrote about the Instagram acquisition through this lens back in 2012.)

Now imagine two social media networks, microblogging platforms if you like. One is vibrant and growing and one isn’t. The difference, possibly, is the effectiveness of the general intellect.

It’s not just ownership and appropriation of general intellect.

One read on the history of the internet is that it’s the history of startups figuring out how to wrap their arms around something and turn it into value before anyone else realises that there was even value there to begin with.

(And, as it happens, this vampiric nature was the topic of the talk by Arvidsson that I attended way back in 2006. Here are my raw notes..)

AI training data for example.

Like GitHub Copilot, WHICH I LOVE and couldn’t work without, but let’s be clear it is trained on the source code of open source software, and at least some of that software would not have granted that permission in their license had anyone thought of it ahead of time.

But who knew you could train an AI from the free-to-view patterns in bulk code as passed through a wood-chipper and reassembled with uh math, the abilities of said AI being something that you could then charge rent for?

ANYWAY so there’s the YouTube generosity superstar MrBeast.

e.g. MrBeast a.k.a. Jimmy Donaldson arranged many free cataract surgeries and filmed the result. The end product was an eight-minute video called “1,000 Blind People See for the First Time.” – 152 million views. Here’s his channel.

Audience means ad revenue which is recycled into bigger giveaways means content means audience, etc. What a flywheel.

Donaldson has built a YouTube empire on this kind of quasi philanthropy, in which he crafts spectacles around surprise cash giveaways (“Giving a Random Homeless Man $10,000”), contests with expensive prizes (“Last to Leave $800,000 Island Keeps It”) and other lavish, if not particularly sensible, gifts (“Tipping Waitresses With Real Gold Bars”).

(Here’s a paywall-dodging link.)


  • MrBeast randomly gives things away: a Lamborghini handed off to a randomly chosen Uber rider, an entire house deeded to an unsuspecting Domino’s delivery person – which is arbitrary and amazing and also kinda led by proximity to MrBeast.
  • Also MrBeast has a chocolate brand called Feastables: fans have been acting as volunteer brand representatives, tidying up the supermarket displays and posting photos of their acts of service to social media – which is a kind of way of earning proximity. This is performed of course vaguely in the hope that you will be struck by a magnanimous miracle lightning bolt.

I mean, buy a lottery ticket or tidy MrBeast’s chocolate bars?

The random-yet-efficacious cause-and-effect is simultaneously how one trains a dolphin, and also the end point of this give-and-take evolutionary cascade is how we end up with gods. Capricious benevolent variable-interval operant conditioning gods that you pray to (or tweet at) for intervention re your cataracts. Gods or billionaires, same same.

BUT, this is not my point.

From that article I learned of the concept of the audience commodity.

Vincent Miller, from the media studies department at the University of Kent in the UK:

“The interesting thing that he was doing was saying: ‘You don’t have to give up anything. All you have to do is watch. And I make so much money from each one of you who views these things.’“

And so!

Miller’s interest in MrBeast resulted in a new academic paper, written with Eddy Hogg, in which Miller places MrBeast in the context of a media-studies concept called the “audience commodity,” the idea that media consumption is essentially a form of labor, because people spend time creating a valuable commodity - an audience - that is then sold to advertisers.


If media consumption is labour then what is the equivalent of minimum wage?

So let’s accept that this is a generally applicable result, and viewers/users are engaged in the labour of building a valuable, being audience commodity, in exchange for the reward of consuming entertaining media (which might be sharing photos or watching TV, whatever it is).

Do users see themselves as workers?

Reddit users do – the mods recently went on strike.

Do users see media consumption as compensation, analogous to wages? That’s where my minimum wage equivalency angle comes in.

A minimum wage is put in place to ensure that workers are given fair exchange for their labour, and also to prevent any “race to the bottom” dynamic as inter-company competition pushes for lower and lower compensation.

I’m not sure how you assess the actual value of media consumption though.

Maybe via a tribunal? They could assess it, just the same as that committee that sets age ratings on movies. They would set a minimum level of being entertained/other value. Too boring? Not equitable in the building of audience commodity! You are taking unfair advantage of your users! Back to the App Store! Try again!

Or a minimum level of being fulfilled.

See, it’s hard to assess the compensation value of dopamine pump Skinner boxes. If you find a mobile game addictive, is that true audience value you are receiving, or some kind of fairy-glamoured fake gold that turns to dry leaves by the next morning?

It is hard to do the assessment.

So I’ve been wondering how to assay the true value of consumed media, and all I can think of is deathbed regrets. People seem to have some kind of special clear-eyed retrospective judgement on how they’ve spent their minutes, approaching the end.

Let’s build on that!

What you would do is to run AI simulations of whole populations, exposed to this particular media.

Then ask the sims on their virtual deathbeds whether or not they felt they had wasted their lives. In contributing to the audience commodity of Candy Crush, you would ask, in that particular labour exchange, say, looking back at it all, your whole span, how self-actualised do you feel? One to five stars.

The percentage of self-adjudicated wasted lives is then inversely proportional to value in labour value exchange.

If the level of regret is too high, then the media itself is a waste of time. People are being exploiting for their time spent building the media owner’s audience commodity. Ban it.

Also fingers crossed that the AI populations that we use for such mass automated assessment don’t possess sentience.


If you enjoyed this post, please consider sharing it by email or on social media. Here’s the link. Thanks, —Matt.