Weekly Review 7 February 2025
Some interesting links that I Tweeted about in the last week (I also post these on Mastodon, Threads, Newsmast, and Bluesky):An AI will reflect the data it was trained on. When that data is taken froma highly-controlled environment like the Chinese internet, its responses will also exhibit those restrictions: https://www.stuff.co.nz/nz-news/360562380/watch-we-test-deepseek-censorship-it-doesnt-go-wellAI have no concept of reality, so such colossal errors are to be expected: https://dataconomy.com/2025/01/29/top-ai-screw-ups-2024/The number of workers with skills in AI does not match the number of skilled workers needed: https://www.techrepublic.com/article/ai-adoption-australia-google-accenture/There are a few tasks I perform regularly where an AI agent would be useful, but I would want to watch it quite carefully: https://dataconomy.com/2025/01/30/ai-can-now-click-scroll-and-type-for-you-but-is-that-a-good-thing/I'm old enough to remember the dot.com crash. But after the crash, the internet didn't go away. The same with AI, after the bubble bursts, AI will still be here and still be used: https://www.computerworld.com/article/3811828/the-ai-bust-is-here.htmlHow AI is shaping future work: https://www.kdnuggets.com/role-ai-shaping-future-workDeepSeek shows a more efficient way of doing AI, but I doubt its long-term impact on the AI industry will be all that significant: https://www.bigdatawire.com/2025/01/29/ai-lessons-learned-from-deepseeks-meteoric-rise/How to successfully deploy AI at the edge of the cloud: https://www.informationweek.com/machine-learning-ai/ai-projects-at-the-edge-how-to-plan-for-successPeople are fed up with AI companies stealing their content to use as training data: https://arstechnica.com/tech-policy/2025/01/ai-haters-build-tarpits-to-trap-and-trick-ai-scrapers-that-ignore-robots-txt/Like most other fields, AI can both help and harm gamblers: https://www.informationweek.com/machine-learning-ai/how-ai-can-help-or-deceive-gamblersThe changes that speech-to-speech AI will bring: https://www.informationweek.com/machine-learning-ai/speech-to-speech-ai-empowering-a-more-connected-worldThe computing demands of AI will grow so large that distributed data centres are needed: https://www.theregister.com/2025/01/24/build_bigger_ai_datacenters/Stalkers are now using AI as a tool to torture their victims: https://www.theguardian.com/technology/2025/feb/01/stalking-ai-chatbot-impersonatorThe confluence between AI and biotech: https://www.informationweek.com/machine-learning-ai/ai-s-key-role-in-the-emerging-bio-revolutionMicrosoft is using AI to detect scams on web pages: https://www.extremetech.com/computing/microsoft-introduces-ai-powered-scareware-blocker-for-edge-browserAnother lawyer using AI to write their briefs, this time in Australia: https://www.theguardian.com/australia-news/2025/feb/01/australian-lawyer-caught-using-chatgpt-filed-court-documents-referencing-non-existent-casesWhile AI can use words, they need to interact with the real world to learn what they mean: https://arstechnica.com/science/2025/02/to-help-ais-understand-the-world-researchers-put-them-in-a-robot/Rather than waiting for a grid connection, this startup uses off-grid renewables to power their AI data centre: https://www.theregister.com/2025/01/30/startup_datacenter_biogas/Content that is entirely generated by AI cannot be protected by copyright: https://arstechnica.com/tech-policy/2025/01/copyright-office-suggests-ai-copyright-debate-was-settled-in-1965/How AI can fill the workforce skills gap: https://www.datasciencecentral.com/the-role-of-ai-in-closing-the-skills-gap-in-the-business-world/The argument against AI agents: https://www.theregister.com/2025/01/27/ai_agents_automate_argument/While quantum computers pose a substantial threat to security, they can also boost AI: https://www.rnz.co.nz/news/political/540347/y2q-facing-up-to-the-threats-posed-by-mega-powerful-quantum-computersDeepSeek will cause a rethink of some investment into AI, but it won't disrupt the entire industry: https://www.theregister.com/2025/01/31/deepseek_implications/Hopefully this will reduce the amount of rubbish their AI comes out with, but I'm not confident about that prospect: https://www.theregister.com/2025/02/03/openai_unveils_deep_research_agent/
- An AI will reflect the data it was trained on. When that data is taken froma highly-controlled environment like the Chinese internet, its responses will also exhibit those restrictions: https://www.stuff.co.nz/nz-news/360562380/watch-we-test-deepseek-censorship-it-doesnt-go-well
- AI have no concept of reality, so such colossal errors are to be expected: https://dataconomy.com/2025/01/29/top-ai-screw-ups-2024/
- The number of workers with skills in AI does not match the number of skilled workers needed: https://www.techrepublic.com/article/ai-adoption-australia-google-accenture/
- There are a few tasks I perform regularly where an AI agent would be useful, but I would want to watch it quite carefully: https://dataconomy.com/2025/01/30/ai-can-now-click-scroll-and-type-for-you-but-is-that-a-good-thing/
- I'm old enough to remember the dot.com crash. But after the crash, the internet didn't go away. The same with AI, after the bubble bursts, AI will still be here and still be used: https://www.computerworld.com/article/3811828/the-ai-bust-is-here.html
- How AI is shaping future work: https://www.kdnuggets.com/role-ai-shaping-future-work
- DeepSeek shows a more efficient way of doing AI, but I doubt its long-term impact on the AI industry will be all that significant: https://www.bigdatawire.com/2025/01/29/ai-lessons-learned-from-deepseeks-meteoric-rise/
- How to successfully deploy AI at the edge of the cloud: https://www.informationweek.com/machine-learning-ai/ai-projects-at-the-edge-how-to-plan-for-success
- People are fed up with AI companies stealing their content to use as training data: https://arstechnica.com/tech-policy/2025/01/ai-haters-build-tarpits-to-trap-and-trick-ai-scrapers-that-ignore-robots-txt/
- Like most other fields, AI can both help and harm gamblers: https://www.informationweek.com/machine-learning-ai/how-ai-can-help-or-deceive-gamblers
- The changes that speech-to-speech AI will bring: https://www.informationweek.com/machine-learning-ai/speech-to-speech-ai-empowering-a-more-connected-world
- The computing demands of AI will grow so large that distributed data centres are needed: https://www.theregister.com/2025/01/24/build_bigger_ai_datacenters/
- Stalkers are now using AI as a tool to torture their victims: https://www.theguardian.com/technology/2025/feb/01/stalking-ai-chatbot-impersonator
- The confluence between AI and biotech: https://www.informationweek.com/machine-learning-ai/ai-s-key-role-in-the-emerging-bio-revolution
- Microsoft is using AI to detect scams on web pages: https://www.extremetech.com/computing/microsoft-introduces-ai-powered-scareware-blocker-for-edge-browser
- Another lawyer using AI to write their briefs, this time in Australia: https://www.theguardian.com/australia-news/2025/feb/01/australian-lawyer-caught-using-chatgpt-filed-court-documents-referencing-non-existent-cases
- While AI can use words, they need to interact with the real world to learn what they mean: https://arstechnica.com/science/2025/02/to-help-ais-understand-the-world-researchers-put-them-in-a-robot/
- Rather than waiting for a grid connection, this startup uses off-grid renewables to power their AI data centre: https://www.theregister.com/2025/01/30/startup_datacenter_biogas/
- Content that is entirely generated by AI cannot be protected by copyright: https://arstechnica.com/tech-policy/2025/01/copyright-office-suggests-ai-copyright-debate-was-settled-in-1965/
- How AI can fill the workforce skills gap: https://www.datasciencecentral.com/the-role-of-ai-in-closing-the-skills-gap-in-the-business-world/
- The argument against AI agents: https://www.theregister.com/2025/01/27/ai_agents_automate_argument/
- While quantum computers pose a substantial threat to security, they can also boost AI: https://www.rnz.co.nz/news/political/540347/y2q-facing-up-to-the-threats-posed-by-mega-powerful-quantum-computers
- DeepSeek will cause a rethink of some investment into AI, but it won't disrupt the entire industry: https://www.theregister.com/2025/01/31/deepseek_implications/
- Hopefully this will reduce the amount of rubbish their AI comes out with, but I'm not confident about that prospect: https://www.theregister.com/2025/02/03/openai_unveils_deep_research_agent/