Weekly Review 14 November 2025

Some interesting links that I Tweeted about in the last week (I also post these on Mastodon, Threads, Newsmast, and Bluesky):

1. When the AI bubble bursts, how much damage is it going to cause? https://www.stuff.co.nz/business/360877272/fears-ai-bubble-correction-mount-despite-wall-st-rises
2. No, AI written ransomware is not in the wild: https://www.theregister.com/2025/11/03/mit_sloan_updates_ai_ransomware_paper/
3. It looks like most tests for AI safety are fundamentally flawed: https://www.theguardian.com/technology/2025/nov/04/experts-find-flaws-hundreds-tests-check-ai-safety-effectiveness
4. What should CIO do when IT staff get laid off in favour of AI? https://www.informationweek.com/machine-learning-ai/when-ai-is-the-reason-for-mass-layoffs-how-must-cios-respond-
5. A lot of money has been spent on data centres for AI. Will it pay off? https://www.theguardian.com/technology/2025/nov/02/global-datacentre-boom-investment-debt
6. Is prompt injection an unfixable security flaw in AI browsers? https://www.theregister.com/2025/10/28/ai_browsers_prompt_injection/
7. Are orbital AI data centres really feasible? https://arstechnica.com/google/2025/11/meet-project-suncatcher-googles-plan-to-put-ai-data-centers-in-space/
8. Grok's AI encyclopedia either reproduces Wikipedia's material or produces complete garbage: https://www.theguardian.com/technology/2025/nov/03/grokipedia-academics-assess-elon-musk-ai-powered-encyclopedia
9. A class of students were caught using AI to cheat, and used AI to write their apology letters: https://arstechnica.com/culture/2025/10/when-caught-cheating-in-college-dont-apologize-with-ai/
10. arXiv bans certain types of submissions after rise of AI generated slop: https://www.404media.co/arxiv-changes-rules-after-getting-spammed-with-ai-generated-research-papers/
11. When students use AI to do their homework, they become lass able to do the work themselves: https://www.theregister.com/2025/11/03/chatgpt_real_understanding/
12. AI browsers can do some useful things, but cannot be trusted: https://www.computerworld.com/article/4083528/ai-web-browsers-are-cool-helpful-and-utterly-untrustworthy.html
13. More AI slop advertising: https://dataconomy.com/2025/11/04/coca-colas-new-ai-generated-christmas-ad-shows-why-generative-video-still-struggles-with-realism/
14. Which AI markets are dominated by current players, and which are still open: https://techcrunch.com/2025/11/03/elad-gil-on-which-ai-markets-have-winners-and-which-are-still-wide-open/
15. Microsoft apologises for its confusing pricing around its AI offerings: https://www.stuff.co.nz/nz-news/360878414/microsoft-apologises-offer-refunds-after-ai-price-hike
16. A decentralised AI, that doesn't need a data centre: https://www.theregister.com/2025/11/02/fortytwo_dcentralized_ai/
17. When the big names in finance start to bet against AI, it's a pretty strong sign the bubble is close to bursting: https://www.rnz.co.nz/news/world/577884/concern-on-share-markets-as-the-big-short-s-michael-burry-bets-against-top-ai-stocks
18. Yet another AI model is yanked from public use after it hallucinates facts, this time about a US senator: https://www.theverge.com/news/812376/google-removes-gemma-senator-blackburn-hallucination
19. UK court rules AI does not store copyrighted images: https://www.theguardian.com/media/2025/nov/04/stabilty-ai-high-court-getty-images-copyright
20. The latest copyright ruling on AI does not really set a useful precedent: https://www.theregister.com/2025/11/04/uk_court_getty_stability_ai/
21. Large scale AI requires large scale nuclear: https://www.theregister.com/2025/10/31/datacenter_biz_and_nuke_startup/
22. LLM AI still can't explain their internal reasoning: https://arstechnica.com/ai/2025/11/llms-show-a-highly-unreliable-capacity-to-describe-their-own-internal-processes/
23. Massive amounts of money are being poured into AI, mostly for data centres to run them: https://www.theguardian.com/global/2025/nov/04/the-mind-boggling-valuations-of-ai-companies
24. Developer roles for AI agents are less in demand as AI takes over more roles: https://www.theregister.com/2025/11/03/demand_for_software_skills_ai_jobs/
25. Students fear that AI is destroying their ability to study, but use it anyway: https://www.theguardian.com/technology/2025/oct/15/pupils-fear-ai-eroding-study-ability-research
26. Energy is now the bottleneck of large scale AI rollout, not chip shortages: https://techcrunch.com/2025/11/03/altman-and-nadella-need-more-power-for-ai-but-theyre-not-sure-how-much/
27. Enzyme design by AI-in this case an enzyme to break down and recycle polyurethane: https://arstechnica.com/science/2025/10/polyurethane-is-the-latest-polymer-broken-down-by-designer-enzymes/
28. An AI lawnmower: https://www.theregister.com/2025/11/01/robotic_lawnmower_ai/