A series of proposals that will make Marc Andreesen furious and ensure I am never promoted by Substack. But the technology industry is massively broken, socializing the costs of its work while privatizing the benefits in fewer and fewer hands. Technology is only beneficial to the extent that it improves the lives of the people it affects — something that should be our guiding principle.
Some of these proposals may seem radical, and they may result in fewer dollars going to technology companies, especially internet companies. Oh well, says I. Maybe we can start having the smart people who work there focus less on how to sell you ads and more on real problems.
This is not a manifesto. All of these ideas probably have many devils hiding in their details. I doubt any of them are original to me. But it is long past time we stop pretending that the tech industry is special. It is not — it needs to be harnessed for good just like any other. And history clearly shows that letting business leaders do as they wish is a recipe for immiseration and repression.
It is not our duty to lay down and die. So perhaps the items below can help us stand up and live.
- Make people and AI companies responsible for the output of AI systems they publish. No more pretending that they shouldn’t be held liable when a large language model lies about medical issues or libels a real person. If you published it or your system produced it, you are responsible for the consequences.
- Ban selling personal data. No more data brokers invading your privacy without your knowledge, no more selling personal information to people who can harm you.
- No more personalized ads. Take away the incentive to sell ads to individuals in narrow categories and you’d get fewer systems that openly discriminate in job and housing listings, for example. There would have to be a discussion about how fine you can cut this before you get too personal, but we sold ads in newspapers and TVs to broad categories, successfully and profitably, before. No reason the internet cannot run on the same ideas.
- Treat internet products and recommendation algorithms as products, not speech. The people who produce these recommendations should be held liable when they go wrong, just like any other faulty product.
- Companies should have the default feed be people you follow. Recommendations can be allowed (subject to the limitations of points one and three), but they should be a separate feed and not the default.
- Imitative AI companies should be forced to get permission before using people’s work to replace them. And please don’t whine to me about fair use. Fair use is a vague, hardly defined concept in the law that largely dependent upon who has the most expensive lawyers in practice. If you are going to take someone’s work in order to replace them with a machine that copies said work, then you need their permission and compensation. that would generate less focus on replacing artists to make a quick buck with more focus on helping researchers find new cures for cancer, for example.
- Strong protections for reverse engineering. Remove the DCMA related limitations that prevent people from figuring out how systems and machines work so that they can be improved upon. We got Word on Macs, for example, after Apple reverse engineered the format. That kind of innovation is good for all of us and much more difficult today.
- In relation to number seven, enforce interoperability between companies. You should be able to easily take your social graph from Facebook to Twitter to Bluesky if you want and they should all allow cross communication. That would encourage companies to compete on service, not by locking you into their bubble. (Corey Doctorow has a book on this subject. It’s not a great book, more of an extended pamphlet with some flaws in its thinking, but on this issue, it is well done.)
- Make online realtors responsible for the quality of the products in their marketplaces. This is a bit obscure, but Amazon, as an example, has been able to claim that it is not the seller of products (despite its algorithms pushing specific products to users, despite its deep control of how sellers must operate on its platform, etc.) it is merely an aggregator and therefore should not be liable for the products sold. If you buy a product from, say, a Walmart store, and it is defective, Walmart has obligations to you that Amazon currently does not. This allows Amazon to profit from fraud and defective products in a manner other retailers cannot.
- Enforce anti-trust and other pro-competitive laws. Amazon should not be able to control its sellers for their own profits. Facebook shouldn’t be able to buy its competitors, like Instagram. The online ad sales should be controlled by Google and Facebook. Capitalism requires competition to have any hope of being anything other than a means of wealth extraction. Tech companies have been allowed ot ignore that basic rule of life for too long.
As I said, some of these may appear radical and all of them will be fought tooth and nail by the tech industry in order to protect their privileged status in the law and the market. I am under no illusions that enacting these will be quick or easy. But we need to start putting these companies under the lash of democratic control, just like we have with every other aspect of our market.
Business is business, whether it is a tech-based business or a local mom and pop store. Without democratic control of business and the market, we inevitably end up in societies that immiserate the majority, create a very few winners, and force all of us to pay the costs of their decisions. Letting the tech companies do what they want has resulted in a less free, poorer society. It is long past time we reigned them in.