Some interesting links that I Tweeted about in the last week (I also post these on Mastodon, Threads, Newsmast, and Bluesky):
- AI rollouts fail because staff are scared of AI: https://www.theregister.com/2026/03/26/workplace_ai_forrester/
- Looks like even AI can become over-confident from false praise: https://www.theregister.com/2026/03/24/ai_models_persona_prompting/
- Shoe-horning a huge AI model into a PC seems like an impressive engineering achievement for a problem that doesn't exist: https://www.computerworld.com/article/4149838/hp-will-cram-a-20-billion-parameter-ai-model-into-new-ai-pcs.html
- The factors working against orbital data centres are more economic than technological: https://arstechnica.com/space/2026/03/orbital-data-centers-part-1-theres-no-way-this-is-economically-viable-right/
- With AI agents, like any other software project, security is important: https://futurism.com/artificial-intelligence/openclaw-bots-security-disaster
- While amateurs using AI to generate court filings have caused chaos, lawyers can find value in using it as a tool to assist them: https://arstechnica.com/ai/2026/03/ai-is-beginning-to-change-the-business-of-law/
- This wearable has AI embedded and can run for two weeks on a single charge, because it uses analog processing for the most common calculations: https://spectrum.ieee.org/wearable-devices-ai-health-mai
- How one university is bringing AI into the arts: https://www.insidehighered.com/news/student-success/academic-life/2026/03/26/richmond-integrates-ai-across-liberal-arts
- It's easy to subvert an AI agent because they are so gullible: https://www.theregister.com/2026/03/23/pwning_everyones_ai_agents/
- The impact of AI on economic growth is zero: https://futurism.com/artificial-intelligence/ai-economy-gdp-2025
- While I am in favour of nuclear power, I'm not sure that using AI to navigate the regulations around them is appropriate: https://www.theregister.com/2026/03/25/microsoft_nvidia_ai_nuclear/
- The psychological downsides to AI making things too easy: https://spectrum.ieee.org/frictionless-ai-psychology
- Companies give AI chatbots personalities because it's good for the bottom line: https://www.computerworld.com/article/4147771/the-dark-side-of-chatbots-with-personality.html
- AI is exacerbating death fraud: https://www.computerworld.com/article/4146580/what-it-leaders-need-to-know-about-ai-fueled-death-fraud.html
- AI are programmed to be sycophants, but this undermines the judgement of their users: https://arstechnica.com/science/2026/03/study-sycophantic-ai-can-undermine-human-judgment/
- Even AI are getting worried about being replaced by AI: https://futurism.com/artificial-intelligence/ai-agent-job-could-be-replaced-by-ai
- Do you want a machine Omnissiah? Because this is how you get one: https://www.theguardian.com/lifeandstyle/ng-interactive/2026/mar/24/ai-religion-god-digital-spirituality
- I don't think it's a good idea for AI to psychoanalyse teen job applicants: https://www.nzherald.co.nz/nz/wellington/teenager-bewildered-at-pretty-stupid-ai-personality-feedback-after-woolworths-interview/RAI3GNQN5REUDPTHB3STGSLBNQ/
- Another novel pulled from publication after accusations that it was written by AI: https://arstechnica.com/ai/2026/03/hachette-pulls-shy-girl-horror-novel-after-concerns-about-ai-use/
- Costly regulation of AI is keeping smaller players out of the market: https://www.informationweek.com/machine-learning-ai/compliance-costs-risk-widening-the-ai-gap
- Using an explainable AI to find new metal alloys: https://www.hpcwire.com/bigdatawire/2026/03/24/data-driven-ai-framework-speeds-discovery-of-metals-built-for-extreme-conditions/
- Moving an AI project from pilot to production is hard: https://www.informationweek.com/machine-learning-ai/why-ai-scaling-is-so-hard-and-what-cios-say-works
- Yet another enormous AI data centre is to be constructed: https://www.theregister.com/2026/03/23/softbank_to_put_mega_server/
- Suddenly, AI-generated bug reports are useful: https://www.theregister.com/2026/03/26/greg_kroahhartman_ai_kernel/
- AI is taking over jobs one task at a time: https://www.theregister.com/2026/03/24/ai_job_unbundling/
- Six ways AI can moderate online content: https://www.techtarget.com/searchcontentmanagement/tip/Types-of-AI-content-moderation-and-how-they-work
- A global survey of attitudes towards AI shows continuing concerns: https://dataconomy.com/2026/03/25/anthropic-releases-worlds-largest-study-on-global-ai-attitudes/
- Is AI derailing the upwards mobility of the working class? https: https://futurism.com/artificial-intelligence/doctor-reels-son-becomes-plumber-ai
- AI-generated X-rays can now fool medical experts. What part of this seems like a good idea? https: https://dataconomy.com/2026/03/24/ai-generated-medical-scans-prove-nearly-indistinguishable-from-reality/
- High-voltage DC power transmission can improve the efficiency of AI data centres: https://spectrum.ieee.org/data-center-dc
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.