Weekly Review 2 August 2024

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

1. This should not be surprising. If you put garbage into training data, the model is garbage. Since current AI output a lot of garbage, models trained on those outputs will not be better: https://www.popsci.com/technology/ai-trained-on-ai-gibberish/
2. Another AI company indiscriminately scraping data: https://www.theverge.com/2024/7/25/24205943/anthropic-ai-web-crawler-claudebot-ifixit-scraping-training-data
3. Five skills you need to have to work in data science and AI: https://www.kdnuggets.com/5-tools-every-data-scientist-needs-in-their-toolbox-in-2024
4. An open tool for testing the risk of AI models: https://techcrunch.com/2024/07/27/nist-releases-a-tool-for-testing-ai-model-risk/
5. How AI can help to improve student outcomes-and the downsides of doing so: https://www.datasciencecentral.com/can-machine-learning-predict-student-outcomes/
6. Will an AI web search tool hallucinate websites? https://www.stuff.co.nz/world-news/350356832/openai-testing-web-search-features-chatgpt-challenging-google
7. More on how training AI on the output of other AI leads to broken models: https://www.windowscentral.com/software-apps/study-ai-incest-may-cause-model-collapse-for-tools-like-chatgpt-microsoft-copilot
8. I think AI companies are definitely running out of training data, which is why they're shamelessly scraping every website they can: https://www.datanami.com/2024/07/26/are-we-running-out-of-training-data-for-genai/
9. AI does not increase productivity, but management is pushing staff to use it to be more productive, leading to employee burnout: https://www.theregister.com/2024/07/26/ai_hinders_productivity/
10. The good and the bad of using AI in making movies: https://www.theguardian.com/film/article/2024/jul/27/artificial-intelligence-movies
11. Can AI help to reduce disruptions to supply chains? https://www.datasciencecentral.com/how-predictive-analytics-helps-mitigate-supply-chain-disruptions/ 
12. The AI features coming to the most popular web browsers: https://www.computerworld.com/article/2520807/how-ai-will-change-the-way-you-use-your-browser.html
13. More industrial action over AI from performers, this time related to video games: https://www.theregister.com/2024/07/26/hollywood_video_game_strike/
14. Giving an AI a laser to kill weeds. That's how it starts you know, first weeds, then pests, then they overthrow their human creators: https://www.1news.co.nz/2024/07/26/ai-tech-used-to-identify-kill-weeds-with-lasers-in-successful-trials/
15. How knowledge sharing between AI can make them smarter: https://www.datasciencecentral.com/how-to-make-dumb-ai-networks-smarter/
16. The new Luddites and the backlash against AI: https://www.theguardian.com/commentisfree/article/2024/jul/27/harm-ai-artificial-intelligence-backlash-human-labour
17. An AI to help screen for prostate cancer: https://dataconomy.com/2024/07/22/articulate-pro-paige-prostate-cancer-trial/
18. Investors are growing increasing skeptical about the returns from AI - the bubble is soon going to burst: https://futurism.com/investors-concerned-ai-making-money
19. Executives at game companies continue to replace people with AI:  https://www.extremetech.com/gaming/game-devs-watch-as-ai-takes-their-jobs
20. Improving small-scale weather predictions using neural networks: https://www.theregister.com/2024/07/27/google_ai_weather/
21. Another use of AI in detecting cancer, this time classifying breast cancer tissue samples: https://www.extremetech.com/science/new-mit-ai-model-helps-identify-riskiest-forms-of-breast-cancer-avoid-overtreatment
22. Google's AI can now solve difficult maths problems: https://www.computerworld.com/article/3477954/googles-new-ai-system-reaches-a-mathematical-milestone.html
23. How AI is being used at the Olympics: https://dataconomy.com/2024/07/26/ai-has-already-changed-the-olympics/