The coming de-enshittification boom (2025)

Venture capital in 2025 is all-in on AI, but generative AI's only proven use cases are slop and scams, according to a blog post by an author citing David Gerard and others. The post argues that Big Tech companies like Meta are battling an 'epidemic of scams' and that AI tools enable fraud, fake research, and low-quality content, while the tech oligopoly remains out of touch with the antisocial behaviors AI enables.

25 May 2025 the coming de-enshittification boom /de-enshittification/ I was going to call this how to prosper in the coming de-enshittification boom but I’m still trying to figure that out. David Gerard has a pretty good summary of today’s Internet business situation, in In 2025, venture capital can’t pretend everything is fine any more https://pivot-to-ai.com/2025/05/03/in-2025-venture-capital-cant-pretend-everything-is-fine-any-more/ . This year, the investors are all-in on AI. Crypto’s still dead, quantum isn’t taking off—AI’s the last game in the casino. And the Big AI companies are having trouble making anything useful even for stuff like customer support where corporate decision-makers are already willing to tolerate low quality https://fortune.com/2025/05/09/klarna-ai-humans-return-on-investment/ . So far, generative AI does have two winning use cases: slop and scams. Widely available generative AI tools are able to generate content that’s similar enough to real content to be plausible, but different enough that it can beat the algorithms used for detecting infringement and fraud. The AI/crime mess is most obvious at the Big Tech companies. Meta Battles an ‘Epidemic of Scams’ as Criminals Flood Instagram and Facebook https://www.wsj.com/tech/meta-fraud-facebook-instagram-813363c8 by Jeff Horwitz and Angel Au-Yeung at The Wall Street Journal Google Ads Shitshow Report 2024 https://blog.zgp.org/google-ads-shitshow-report-2024/ You can almost stop reading this report at the release date. A report on the crime level of Google ads that got released on the same day as huge antitrust news isn’t going to be good. And it’s not just those companies. Sometimes it looks like we’re in a doom spiral of fake research https://pivot-to-ai.com/2025/05/17/how-to-make-a-splash-in-ai-economics-fake-your-data/ , bogus articles https://balloon-juice.com/2025/05/17/chasing-the-electric-pangolin-open-thread/ and legal briefs https://substack.com/@garymarcus/p-164360492 , auto-generated accessibility problems https://adrianroselli.com/2025/05/do-not-publish-your-designs-on-the-web-with-figma-sites.html , AI slop in educational apps https://pivot-to-ai.com/2025/05/04/duolingo-replaces-its-contractors-with-ai-courses-with-slop/ and even classrooms https://thoughtshrapnel.com/2025/05/14/the-classroom-ai-doom-loop.html , and, worst of all, as Asma Derja covers in The Weakest Link in AI https://asmaderja.substack.com/p/the-weakest-link-in-ai , an enforcement of Big Tech points of view. Before AI can understand the world, someone has to tell it what the world means….Labelers in Kenya, the Philippines, and Venezuela are handed instruction sets written in California, reviewed in Dublin, and told to mark what counts as toxic, hateful, threatening, or real. They don’t decide. They execute. Rands https://randsinrepose.com/archives/minimum-viable-curiousity/ writes, I am deeply suspicious of AI, especially after watching decades of social networks monetize our attention while teaching us to ignore facts and truth, minimizing our desire to understand….AI does an incredible job of confidently sounding like it knows what it’s talking about, so it’s easy to imagine what it will do in the hands of those who want to manipulate you. Soatok writes, in Tech Companies Apparently Do Not Understand Why We Dislike AI https://soatok.blog/2025/05/04/tech-companies-apparently-do-not-understand-why-we-dislike-ai/ . What concerns me about AI has very little to do with the technology in and of itself. I’m concerned about the kind of antisocial behaviors that AI will enable. The tech oligopoly is even more out of touch now than IBM or Microsoft ever got during their peak days. Big Tech companies have enough power to get away with being mean and dishonest and half-assing things. Most people don’t. The result is that the technologies that Big Tech builds, with the assumption that a certain level of meanness, dishonesty, and half-assing are acceptable, are less and less relevant to more and more people. Maybe we’re not in a doom spiral after all? What if we’re back to something more like the dismal Microsoft monopoly from which came the Internet and free software boom of the 1990s? The green shoots of de-enshittification are popping up in lots of places once you start to look for them. Here’s a de-enshittification sighting in a toot from Hugo-award-winning author Charles Stross https://wandering.shop/@cstross/114541747269406389 explaining his switch from Microsoft Office 356 to LibreOffice https://www.libreoffice.org/ : CoPilot in Office would open me up to accusations of breach of contract—my book contracts warrant that they’re all my own work: CoPilot brings that into question. Generative AI slop has real consequences for authors https://www.404media.co/authors-are-accidentally-leaving-ai-prompts-in-their-novels/ because of reputation risks and contract terms. A Big Tech company’s AI service can output copyright-infringing slop for a user—what are they going to do about it? An individual author can’t turn in copyright-infringing slop to a big publishing company. More sightings… 12 years of Ghost https://john.onolan.org/12/ by John O’Nolan. As ever, we’re competing with prominent, VC-funded platforms with tens of millions of dollars and hundreds of employees, playing zero-sum games to try and control the entire media industry. Now, though, it feels like there’s more appetite than ever for something different. What makes a tech gadget calm? This certification firm has an 81-point checklist https://arstechnica.com/gadgets/2025/05/what-makes-a-tech-gadget-calm-this-certification-firm-has-an-81-point-checklist/ by Scharon Harding. Currently, six products are Calm Tech-certified How to Improve Your Site’s Typography https://scottk.mba/how-to-improve-your-sites-typography/ by Scott Knight. More people are making personal web sites again. Just a QR Code http://blog.fnaard.com/2025/05/just-qr-code.html by Gabriel M. Schuyler You type something, it makes a QR code. You can also set the size, error correction, and colors. And that’s it. Just a QR code. Slop, surveillance, and extreme right politics are all part of the same big movement, so political concerns are a big motivator to avoid the first two. The networker: Microsoft shutting down email accounts of Trump’s foes should be worrying to us all https://observer.co.uk/news/columnists/article/the-networker-microsoft-shutting-down-email-accounts-of-trumps-foes-should-be-worrying-to-us-all by John Naughton on The Observer. Khan, who is a British subject, has already lost access to his bank accounts, which is what happens when you’re sanctioned. More interestingly, though, Microsoft has also blocked his email account….The reality is that it’s a global company based in the US. More: Trump’s trade war risks splintering the Internet, experts warn https://arstechnica.com/tech-policy/2025/05/trumps-trade-war-risks-splintering-the-internet-experts-warn/ by Ashley Belanger. the Eurostack, so hot right now https://blog.zgp.org/mlp-2025-02-08/ Enterprise Policies for Firefox Customization: A Beginner’s Guide https://www.quippd.com/writing/2025/05/18/enterprise-policies-for-firefox-customization-a-beginners-guide.html by Asif Youssuff. I wanted to post something to enable people who are currently using dodgy user.js files and forks without security teams to use the mainline version of Firefox, just tweaked in known safe ways. See Data and Firefox Suggest https://blog.mozilla.org/data/2025/05/07/data-and-firefox-suggest/ for info from Mozilla on one of their ad features. Firefox is still good, but they do end up in weird risky advertising rat holes https://blog.zgp.org/the-advertising-cartel-coming-to-your-web-browser/ sometimes, so it helps to have an independent tool to fix the settings. Bonus link: When Your Webcam Doesn’t Work: Solving Firefox and PipeWire Issues https://jgrulich.cz/2024/12/13/when-your-webcam-doesnt-work-solving-firefox-and-pipewire-issues/ by Jan Grulich