AI slashes ads by boosting intention (+ weekly update)
AI slashes ads by boosting intention (+ weekly update)
A look at a possible relationship between AI, intention and ads
Disclaimer: this article assumes we’ll use AIs that have our backs (what people call ‘aligned’). That’s a very rosy assumption, and one I’ll test elsewhere.
We’ve got a computer in front of us. How do we get from our Intention — something that lives in our brain — to some kind of end Result on the computer?
Like this, more or less:
Intention: our brain creates one.
Action: we Speak or Move to bring our Intention into the real world.
Input: a camera, microphone, mouse, touchscreen or keyboard picks up our Action and turns it into a digital Input.
Result: the computer turns the digital Input into some kind of Result, and something happens.
We do this loop thousands of times a day.
That sounds OK, but our poor Intention sadly doesn’t do very well. Unfortunately, it gets weaker at every single stage. Look at how it drops off:
That’s pretty bad. By the time we get to our Result, we are hardly aligned with our initial Intention at all.
You might think I’m exaggerating. But if anything, I might be underplaying the situation. I can think of countless times when my Intention Alignment drops straight to 0% at the very first stage! I decide to do something, I open my phone to move towards the Action — but I get distracted and do something completely unaligned with what I intended.
That’s how, most of the time, we end up in the big red box — the Crapgret Zone. When we end up there, we see and do crap we didn’t intend to see and do, and we (sometimes) regret it.
As you can see from the handy blue bubble, ads live in the Crapgret Zone. We rarely intend to consume ads — they are part of the crap that we end up seeing when our Intention Alignment is not doing so well.
A lot has been tried to shrink the Crapgret Zone. Ad blockers, website blockers, time trackers, electric shockers. Now it’s AI’s turn.
AI and intention
So what happens when we start using AIs to turn an Intention into a Result?
Let’s take a look:
Two things have changed.
First, let’s look at the AI Tunnel. All that Input — audio, keystrokes, visual data — does not directly cause a Result. Instead, it goes through the AI Tunnel where it is translated into a far more complex Input — one that is scarily faithful to your Intention.
The reason that’s possible is that the AI is extremely experienced when it comes to Inputs. In fact, it’s been trained on billions of them. It has a good idea of what the Intention behind them might be — because your Intentions aren’t so different from those of others.
The AI’s more complex Input goes on to cause something to happen on the computer — something much more complex than you’d have managed in a single loop without the AI. Like creating an entire piece of music based on humming a riff, or summoning the exact YouTube video you want with just some vague instructions.
Second, we change our Actions in anticipation of the AI Tunnel. We get an Intention Confidence Boost, as an angelic voice says to us from on high, ‘It’s OK, you can just say what you want’.
(We begin to use prompts, in other words.)
And we become more and more confident with time. As AIs improve and we anticipate a more powerful AI Tunnel, we become more likely to use an Action simply to state our rawest Intention, instead of using the Action to fulfill just one part of the Intention — which we had to do before AI.
How does our Intention Alignment look now? A lot better:
At the Action stage, we’re much more aligned with our Intention because of that Intention Confidence Boost (which, by the way, is something we did all by ourselves — a bit like Ron Weasley with his placebo Felix Felicis. We should pat ourselves on the back).
At the Input stage, the AI Tunnel gives us a more complex Input that’s better-aligned with our Intention.
And finally, we get a nice big Intentionality Pay-Off at the Result stage. Maybe now we can loop hundreds of times a day, instead of thousands. Woohoo!
Side note 1: because of the Intention Confidence Boost, we’ve started telling computers what we want, instead of laboriously using computers to get what we want. Information has been surfaced in the real world that we previously left in the brain. That has some useful second order effects for us and for the AI:
We start thinking more in terms of intentions
The AI remembers our intentions and can align us better with them in future
Side note 2: how might things change from this point on? Let’s think about it in terms of drop-offs 1, 2, and 3 on the chart.
Drop-off 1 will shrink as we give better prompts.
Drop-off 2 will shrink as AI gets better at translating our Input.
Drop-off 3 will shrink as AI gets access to more controls, for instance through APIs that allow it to control more aspects of our operating systems.
Boost intention, slash ads
That’s all very well, but aren’t we meant to be talking about ads?
Yes — look!
The Crapgret Zone has shrunk! That’s a big deal, because as we already know, it’s where ads make their home. Now there’s much less space for ads to set up camp, which is bad news for the adtech industry.
But maybe they shouldn’t have made themselves completely reliant on the existence of a big Crapgret Zone.
This is my first model of how the AI vs. ads scenario could play out. It’s basic and full of assumptions, but I’ll be diving deeper in future posts. Stay tuned!
Weekly update
Challenges: Various pieces of work ongoing in the background. On the linknames side, I followed up with Gilga after a reader kindly told me that Donald Glover used a matching domain for his album Because the Internet! The domain is becausetheinter.net, and that’s incredibly cool because my pitch to DG suggests doing something similar (except in my version, he’d have simply called his album becausetheinter.net, and that would have been a linkname). Anyway, I still awkwardly own the Gil.ga domain, and that’s looking unlikely to change any time soon. Noice!
AI vs. ads: Launched a new series of articles about how AI can (maybe) destroy the adtech industry — you just read the second in the series.
Stats: 194 subscribers, +32 from last week. If you’re new, thank you so much for subscribing. I’ll keep trying to write useful and interesting things about Big Tech.
Link that’s guaranteed to make you feel immoral: Right here.
Link that’s guaranteed to make you wonder at the world: Right here.
This week’s writing
That’s all for now — have a great weekend and I’ll be back next week, probably with more on AI vs. ads.