Image (and now video) diffusion models have been mainstream for a couple of years. They're immensely popular, with billions of images and text prompts being generated daily. The huge computational running cost of all this is enormous - with OpenAI alone purportedly losing millions of dollars every day. This is a familiar cycle in ground breaking tech - after explosive growth comes the question, 'how are we going to monetise this?' - the answer (as always), is advertising, specifically pay-per-click advertising.

I believe we are on the verge of seeing the pay-per-click advertising model applied to AI Image and video generation - resulting in product placement on steroids. Are you generating an image of an urbanite male aged between 16-25? Nike is paying $5.00 per 1000 images generated to ensure the person in that image is wearing Nike Air Jordan 1 trainers. Generating an image of a delicious casserole for your food blog? Le Creuset has paid for it to be displayed in its latest dish.

This model will first be applied to image generation - text generation will soon follow (although this is more nuanced). It's easy to imagine a few corporate PR disasters on the horizon, in which ad copy is bluntly forced into passages of inappropriate text - but eventually it will find its feet. In the beginning it may even be something as primitive as appending an ad to the end of the passage of text - 'This article on heart health sponsored by Acme Drug. For a free cardio health score visit...'.

All of this is nothing new. In the early days of the internet, when it still felt like a cross between a hippy commune and lawless frontier, advertising arose as the dominant revenue model for search engines and content creators. It's now THE dominant form of advertising, accounting for 70% of all global ad spending. Apart from the awful period of pop up ads in the 00's, all of this is a good thing - the economic engine behind an explosion in creator content that now allows me to find production quality YouTube channels on almost any niche topic I can think of - the YouTube channel Rate my Takeaway has provided me with more hours of entertainment and joy than any Hollywood blockbuster of the past decade.

I've seen a lot of healthy skepticism towards AI hype (myself included), but I think we're forgetting the recent past. People are skeptical about the hype around the tech, but forgetting about the ad revenue potential. There was similar skepticism with Google, Facebook and Reddit's early multi-billion dollar valuations - because people were valuing them based on their tech, not their ad revenue potential. I'm seeing the same fallacy with openAI, Stability AI et al.. I believe they will become multi trillion dollar companies in the future, and that value will be derived from advertising, rather than AGI, or Agentic AI, or any of the other buzzwords that are currently being floated about. As Giaus Baltar (the scientist-cum-prophet in the Battlestar Galactica franchise) said of AI 'All this has happened before, and all this will happen again'.