Weekly Review 13 December 2024

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

1. As skeptical as I am about shoving AI into everything, this might actually be useful: https://www.extremetech.com/computing/creators-of-arc-browser-tease-new-ai-browser
2. So OpenAI won't be running ad-supported AI just yet: https://www.theregister.com/2024/12/02/openai_mulls_other_revenue_streams/
3. AGI in six years? I suppose if you keep saying something will happen in six years, eventually you'll be right: https://www.popularmechanics.com/technology/robots/a63057078/when-the-singularity-will-happen/
4. I really don't want an AI shopping for me: https://techcrunch.com/2024/12/02/the-race-is-on-to-make-ai-agents-do-your-online-shopping-for-you/
5. The role of AI in project management: https://www.datasciencecentral.com/revolutionizing-product-management-with-ai-exploring-innovation-machine-learning-and-ethical-practices-across-the-product-lifecycle/
6. How much Google's AI can tell from your photos: https://arstechnica.com/ai/2024/12/new-website-shows-you-how-much-google-ai-can-learn-from-your-photos/
7. The problem with expecting people to verify the decisions made by AI like facial recognition is that people are lazy: https://www.nzherald.co.nz/nz/rotorua-woman-fights-facial-recognition-tech-after-wrongful-supermarket-trespass/GGK7BYBC7RBLHB2MK2SOJD6UFQ/
8. Google seems to be taking the position that creators should opt-out of having their creations used to train AI, rather than Google asking for permission: https://www.theguardian.com/technology/2024/dec/03/it-feels-like-a-startup-energy-googles-uk-boss-on-the-advent-of-ai
9. AI is making its way into online courses, even though they're not that accurate about what they teach: https://techcrunch.com/2024/12/05/openai-wants-to-pair-online-courses-with-chatbots/
10. AI is a tool, it can be a helper or an enemy to artists: https://www.theverge.com/2024/12/5/24306548/stormy-pyeatte-friend-or-faux-ai-interview
11. AI can out-perform human doctors in diagnosing patients: https://www.computerworld.com/article/3613982/will-ai-help-doctors-decide-whether-you-live-or-die.html
12. Most open AI is not open, and not the same as open source software: https://www.theregister.com/2024/12/02/open_ai_research/
13. Makes we wonder if AI are creating the ads for AI: https://techcrunch.com/2024/12/03/the-abject-weirdness-of-ai-ads/
14. The first steps in benchmarking AI safety: https://www.theregister.com/2024/12/05/mlcommons_ai_safety_benchmark/
15. An AI is not the same as a search engine: https://www.theverge.com/2024/12/5/24313222/chatgpt-pardon-biden-bush-esquire
16. Education, ICT and finance are the areas most likely to be affected by AI, according to an OECD report: https://www.theregister.com/2024/12/03/ai_oecd_report/
17. How to fine tune your large language model AI: https://www.kdnuggets.com/top-5-tips-fine-tuning-llms
18. An AI weather prediction model: https://www.theregister.com/2024/12/05/google_deepmind_weather_model/
19. Emotion recognition is a fairly old application of AI: https://techcrunch.com/2024/12/05/google-says-its-new-open-models-can-identify-emotions-and-that-has-experts-worried/
20. British workers mostly perceive AI as chatbots: https://www.theregister.com/2024/12/01/uk_workplace_ai/
21. Of course AI is going to be used in weapons: https://arstechnica.com/ai/2024/12/openai-and-anduril-team-up-to-build-ai-powered-drone-defense-systems/
22. What kind of 'glitch' would exclude one particular name from an AI? And is it the only name affected? https://www.theguardian.com/technology/2024/dec/03/chatgpts-refusal-to-acknowledge-david-mayer-down-to-glitch-says-openai
23. Turns out AI didn't interfere in elections this year as much as people feared: https://www.computerworld.com/article/3618188/meta-ai-created-less-than-1-of-the-disinformation-around-2024-elections.html
24. An  overview of AI agents: https://www.computerworld.com/article/3617392/what-are-ai-agents-and-why-are-they-now-so-pervasive.html