I think they "need" to fail because, standing back, they serve no real need other than the acceleration of the SV money machine. AI needs to end up as a useful component in various systems, the fantasy that the "everything machine" is desirable, inevitable or beneficial needs to be ended.
If the law doesn't protect creators and users online, how can they thrive as the AI rampage continues? One option: make a little internet which works the way we want it to, and invite everyone in. Robots can go to the tradesman's entrance round the back. dominicyoung.uk/2025/10/07/g...
Ha ha I guess I need to work out which image Bluesky picks and add one!
Setting up a new Wordpress for something and I rediscovered this old one I made a few years ago as a sort of DIY Instagram....
ihaveseen.blog
This one doesn't seem to coincide with anything sensical at all
If you DM me I'll send you a link to something you might find interesting...
It's a real world example of an idea which was riddled with fatal flaws. If you start with user needs and real world outcomes from subscription it's very hard to conclude that nobody wants zero-commitment access to whatever media they want, whenever they want it. Like in shops...
Saying "micropayements don't work because Blendle didn't work" is like flapping your arms and concluding that heavier-than-air flight is impossible. There are many reasons why Blendle would never have worked. The fact remains that most users hate paywalls and would like access without commitment.
Ha! You're quoting Shirky! The fact that media peoples thinking and imagination hasn't moved on tells it's own story. They have been continuously atrophying, demanding that people who don't want to and never will, subscribe, and taking smug comfort in decades old dogma. No wonder they're dying.
Yes. And we have built it. People really like it. The more publishers adopt it the faster it will grow, because the one login and payment method works across all participating sites. There's an orthodoxy which says subscription is the only form of payment which works. But it's not true. @axate.com
Many bots and crawlers don’t identify themselves when they access sites. Some of them go to great lengths to avoid detection. I’m interested to hear some reasons why this behaviour is legitimate and reasonable and/or reasons why identifying themselves would be problematic…
It's time for publishers to declare independence from Google.
pressgazette.co.uk/comment-anal...
But AI companies are huge businesses and have a huge bargaining chip - a share in their future success. Existing in a state of war won't help anyone. It's time to stop posturing, start collaborating and negotiating...
22/22
Whatever happens, creators should and will own a bigger share of the future value of AI systems built using their work. AI companies which went early and chose to ignore copyright might be disadvantaged versus newer, compliant competitors.
21/22
That still leaves the problem of the infringing that has already happened. It can be resolved in a couple of ways. One is to write big compensation cheques. Another is to give a bigger share of the future revenues to the creators whose work was stolen
20/22
AI companies which choose to pay will find themselves with all the best content and better AI products as a result. Others will find themselves locked out.
19/22
This time around that approach seems doomed to fail, because their need for content is ongoing and continuous. If they refuse to pay for it then they'll lose access to it and nothing they or any politician can do will be able to force people to keep working for free.
18/22
The other reason is greed. West coast investors are used to making big bets on "winner takes all" outcomes. They like huge scale, huge margins and zero input costs; the model which made Facebook and Google. To put it another way, they don't like sharing. They want it all.
17/22
Obviously copying that has already been done is a huge headache for AI companies. Potentially a multi-trillion headache. They can't afford not to keep playing chicken with this, and credulous politicians are making them hopeful that the problem might be magicked away
16/22
So why aren't more creators and AI companies teaming up to unlock all this value and build licensing frameworks which work for both, beyond the piecemeal deals which have been happening so far? Two reasons, in my view. One is the big numbers at the top of this thread.
15/22
Nor would any creative business would have to take a chance on guessing the future value of the AI market and gambling that they got it right. Being able to share in the value their content helps create would make the whole thing less risky and more rewarding.
14/22
Adopting a royalty-based approach to licensing would work brilliantly for AI companies and creators alike. No AI company would have to shoulder multi-billion dollar up-front payments, keeping the market open for all.
13/22
Companies do it, too, when they sell shares. If you invest early in a startup, for example, you pay relatively little for the shares and can't be sure what they will be worth in future. If the company does well, your share of it is worth more - maybe a lot more.
12/22
This helps share risk and reward, and aligns incentives. Book authors who sell hundreds of thousands of books make a lot more money than those who sell fewer. So the question of "how much should AI pay?" is easily answered: it depends on how much they make.
11/22
The problem of doing a deal without being sure of its future value is another one the creative industry is used to solving. They call it "royalties" - rather than a fixed, one-off, amount paid up-front, creators get a share of the revenue their work helps to generate.
10/22
The issue is price. How much should AI be paying? Nobody knows, because nobody really knows how AI will generate revenue, or how much of it there will be. Hyperbolic predictions of multiple trillions have yet to come true.
9/22
