The Myth of Endless Supply of Free Stuff
“Why would anyone pay for that?” he said, blowing a cloud of smoke up my
Apologies Sir But You Seem to Have It Backwards
When I organized PaynPoint I had a bizarre conversation with someone from 500.co.
“Why would anyone pay for that?” he asked, snapping his gum, and referring to scraping content from the web. “They already get it free.”
“They’ll pay for it because AI companies don’t have a business if they don’t have content to train on,” I told him. “Content is the raw material for AI.”
He blew a raspberry. “That’s not how it works,” he replied. “Payments come from the demand side. If you don’t get the demand side to agree to pay, then you can’t implement payments on the supply side.”
Which I thought was an incredible thing to say. It’s one of those things that’s half-right, but in such a twisted way you feel crazy having to explain how backwards it is.
Since when does a payment obligation come from the demand side?
You don’t pay for pizza because you want to, you pay because the supplier makes you.
Transactions Originate from Supply, Not Demand
You don’t go to the store and insist “I’m not leaving until you take my money!”
Payments are imposed by supplier, not voluntarily adopted by the desirer.
If demand refuses payment, the supplier refuses to give the supply to the demanding party.
Unless it’s the internet, where — for some reason — nobody thought to do that before.
When the price of a thing is zero, demand is unbounded. When the price of a thing is infinite, demand is zero.
Between zero and infinite is where economics happen. And economics tells us that there’s an equilibrium price where demand and supply are equivalent.
Suppliers put prices on their products to fund their effort to deliver the product, but also to limit the demand for their product to a volume that they can supply against. When demand exceeds supply, you raise prices to find that equilibrium.
For digital content, the product is infinitely reproducible, and the same product can be resold time and again. Which means that there’s no fundamental upper bound on the volume of the product a supplier can supply!
Which is an extremely good thing for digital content, if the supplier prices it.
Most websites… don’t.
Free Scraping is an Expensive Problem
Scraping content to train AI has created no small amount of contention. Wired has a AI copyright lawsuit tracker (thanks to Kate Knibbs ) that last I checked, had 27 lawsuits between major media owners and AI companies. Individual lawsuits are themselves worth billions.
That’s a lot of money for something they got “for free”.
AI companies spent around $3b to acquire content just in 2024. The AI content market was estimated to reach $10b by around 2030 even before the flood of AI copyright lawsuits.
That’s a lot of money to spend on something that you can get “for free”.
In 2024, AI companies got proactive about licensing content for training. reddit has three AI licensing deals that made up about 20% of reddit’s 2024 revenues. As of six months ago, OpenAI was licensing content from seven major publishers.
If AI companies can get the content they need for free, then why are AI companies paying so much for content?
A lot of content is scraped by third parties and sold on marketplaces by someone, and that someone is usually not the person who owns the content. Which means AI companies are paying for content, just not paying the originator of the content.
While an AI company is already paying something to get content, and someone to get the content, they usually aren’t paying the source of the content.
PaynPoint is here to change that.
If You Liked it Then You Shoulda Put a Ring On It
Copyright owners and content creators have started to realize that AI companies don’t have a business without content to train their AI on.
领英推荐
An AI company doesn’t have AI without content to train on, so an AI company must obtain content to produce and sell AI. That’s why you need a way to price your content!
This realization translates all existing content of every sort — newspaper archives, old magazines, professional research journals, fanzines, video, music, images — into a goldmine for licensing to AI companies. Your website included.
The problem is that the internet wasn’t set up like that. The internet was set up for people, and it was assumed that pretty much all internet traffic was from a human.
Which, like many things, used to be true, but isn’t true anymore.
Another hangover false assumption was that bot traffic is beneficial for your website, which also used to be (kind of) true, since most of the bots, originally, were search engine indexers and later search engine optimizers.
But that’s not true anymore either, because 1/3 of all internet traffic is malicious bots, and another 1/4 is content scrapers, acquiring content to feed into AI models.
In our experience reviewing our users’ web traffic, about 1.5% of bot traffic actually benefits the website. Why subsidize 200 bots when only benefit from 3 of them?
This is where robots.nxt comes in.
We monitor your web traffic and segment it into humans and bots. We pass the humans on invisibly, they don’t notice anything happened. We ignore requests from the bad bots, and apply rules to requests from commercial bots.
One of those rules is adding a price to serving pages.
Pay Per View Internet for Bots
robots.nxt is like a point-of-sale system for your website. It makes your website pay per view for bots. Scrapers can’t fill a shopping cart with stuff and walk out anymore.
When we see a content scraper, instead of just giving away your content for free, we make the bot pay you. The price is up to you. What pages, routes, files, or other content you charge bots for is up to you. We’re just giving you the tools to charge for it.
Content scraping is “free” from your website, with AI companies paying for infrastructure to scrape you, or buying your content from an intermediary who got it from your website without paying, while you pay the costs to serve them that content.
And if half your traffic is bots, then half your web hosting costs are wasted on bots.
This is no small market, AI companies are already paying $3 billion per year to get content. And they’re paying major websites for their content.
So why aren’t they paying you? Well… because you’re just giving it away.
These AI companies are major corporations. They make enormous profits (well, not OpenAI, not yet).
They make those profits on your back. From your content.
Why not you? Why shouldn’t you get paid for your content? Why give away the raw material that AI depends on, for free, so that someone else can make money from your work?
Your website costs you money. And your website is supposed to make you money.
Every bot that scrapes you is not only lost revenue, but an operational expense you bear to serve that content to the bot so they can make money off of you!
Why are you subsidizing AI? Why are you subsidizing massive, profitable corporations?
Payments are imposed by the supply-side, and you’re the supplier of content that AI needs.
robots.nxt is a point-of-sale system for your website that makes your website pay-per-view for bots so that you can get paid for your content.
Turn on robots.nxt and see how many bots you’re subsidizing, how much it’s costing you, and how much money you can make by charging AI companies for your content.
It takes about five minutes to set up.
What do you have to lose?