Hacker Newsroom AI for 26 May: Better AI Coding, Pope Leo On AI, Norway Sovereign LLM, Opaque AI Power

Hacker Newsroom AI for 26 May recaps 5 major AI Hacker News stories, moving through better ai coding, pope leo on ai, norway sovereign llm, opaque ai power. 1. Better AI Coding The next story is about using AI to write better code more slowly, with Nolan Lawson arguing that coding models are most valuable as tutors, reviewers, and iteration partners rather than slop cannons, and that matters because it reframes AI coding as a quality tool instead of a speed contest. Hacker News mostly agreed that this workflow can raise the quality bar, but the debate turned on whether the extra review loops preserve understanding or just replace typing with expensive supervision. Story link Hacker News discussion 2. Pope Leo On AI The next story is about Pope Leo the Fourteenth's first encyclical on AI, which argues that the technology must serve humanity rather than concentrated private power, and that matters because it pushes AI governance into moral, political, and international terms instead of treating it as a product roadmap. Hacker News reacted with a mix of surprise, admiration, and skepticism, with many readers saying the document sounded more serious about human dignity and long-term accountability than most mainstream tech commentary. Story link Hacker News discussion 3. Norway Sovereign LLM The next story is about Norway's National Library building a sovereign Norwegian language model with a two petabyte Huawei flash layer in front of a much larger archive and supercomputer pipeline, and that matters because it shows how local language AI depends as much on data plumbing, licensing, and governance as on raw GPU counts. Hacker News pushed back on the headline at first, then settled into a more technical argument about whether the storage was really for training, how much hardware is enough for a national model, and whether cultural sovereignty justifies the spend. Story link Hacker News discussion 4. Opaque AI Power The next story is about a second write-up on Pope Leo's AI manifesto that zeroes in on opaque algorithms controlled by a few firms and their risk of creating new forms of dehumanization, and that matters because it frames AI concentration itself as a civic and human problem rather than just a market outcome. Hacker News did not hold a separate debate on this submission for long because the comments were redirected to the original discussion, but the main reaction in that source thread centered on whether the Vatican is offering a clearer critique of private tech power than most governments or companies are willing to make. Story link Hacker News discussion 5. CVE 2026 28952 Apple macOS The next story is about Apple crediting Claude and Anthropic Research alongside Calif dot io for a macOS kernel vulnerability, and that matters because it is another concrete sign that AI-assisted security research is moving from marketing language into official vendor advisories. Hacker News did not treat this as magic bug hunting so much as a sign that large-scale automated auditing is becoming normal, with debate over whether this is just fuzzing with better tooling or the start of always-on model-driven exploit discovery. Story link Hacker News discussion That's it for today, I hope this is going to help you build some cool things.