Friday, December 1, 2023

Weekly Review 1 December 2023

Some interesting links that I Tweeted about in the last week (I also post these on MastodonThreadsNewsmast and Post): 

  1. A brief(ish) summary of the Sam Altman saga at OpenAI: https://www.nzherald.co.nz/world/what-does-sam-altmans-firing-and-quick-reinstatement-mean-for-the-future-of-ai/KOIP6NE5HRDK7CYUV2J6LHICYY/
  2. More on the fallout and implications of the Sam Altman / OpenAI debacle: https://www.nature.com/articles/d41586-023-03700-4 
  3. There's nothing wrong with teaching Maori culture, or with wanting Maori and Pasifika kids to do better in education. But when New Zealand kids overall are slipping further down the international rankings, maybe prioritise English, maths and science? https://www.stuff.co.nz/national/education/301014381/damien-grant-our-education-system-is-constantly-failing-our-students
  4. When the code ChatGPT generates, is your code: https://spectrum.ieee.org/chatgpt-code
  5. The data sets used to generate facial recognition models lack diversity, especially of smaller populations like Polynesians. This is known. A biased data set will produce a biased AI. This also is known. Now it seems the general public are becoming more aware of this: https://www.nzherald.co.nz/kahu/maori-being-used-as-guinea-pigs-for-facial-recognition-tech-ethicist/L3UKTI5AT5BT7NFJD4CTAZJQOA/
  6. More jobs going from New Zealand universities. Just the outcome of a succession of governments that have failed to support higher education. https://www.rnz.co.nz/news/national/503209/all-the-jobs-likely-going-and-already-gone-from-new-zealand-s-universities-in-2023
  7. A framework for tuning the prompts used for generative AI: https://hbr.org/2023/11/improve-your-companys-use-of-ai-with-a-structured-approach-to-prompts 
  8. On how artificial intelligence is being used in art history-from detecting elements of style, to reconstructing lost works: https://www.nature.com/articles/d41586-023-03604-3
  9. How artificial intelligence and machine learning got to be so big at Facebook: https://www.technologyreview.com/2023/11/14/1083336/how-facebook-went-all-in-on-ai/
  10. Coca-Cola designed by AI is not something most people will want: https://www.stuff.co.nz/life-style/food-drink/301013178/newsable-i-tried-ai-coke-heres-what-it-tastes-like 
  11. Using AI and machine learning to improve estimates of rainfall: https://spectrum.ieee.org/weather-satellite 
  12. Online learning improves participation in higher education, especially for more marginalised groups. It does come with its own set of challenges, however: https://www.stuff.co.nz/pou-tiaki/133337278/online-learning-can-make-university-more-accessible-for-mori-report-finds
  13. Using AI to assist with the diagnosis, treatment and assessment of animal diseases: https://www.nzherald.co.nz/nz/its-a-rain-gauge-for-disease-an-otago-based-companys-ai-assisted-technology-has-a-trove-of-applications/WSUDJ4CE6FEVTBPUHNX5NH2WWQ/
  14. Are we now going to see an arms race between AI that generate fake data and AI that will detect fake data? https://www.nature.com/articles/d41586-023-03635-w 
  15. I wouldn't trust ChatGPT if it told me the sky is blue, I'm certainly not going to let it plot my data. But at least it will show you the Python code that generated the plot, so it's useful in that respect: https://www.kdnuggets.com/create-stunning-data-viz-in-seconds-with-chatgpt 
  16. Ultimately, AI is a tool. The danger with artificial intelligence is how people use it: https://www.theguardian.com/commentisfree/2023/nov/26/artificial-intelligence-harm-worry-about-people-control-openai
  17. An overview of machine learning and AI for microbiologists: https://www.nature.com/articles/s41579-023-00984-1
  18. Why does this article on the safety of artificial intelligence read like it was written by an AI? https://www.datasciencecentral.com/why-generative-ai-safety-research-is-beyond-alignment/
  19. How to use the new AI features in YouTube: https://dataconomy.com/2023/11/22/youtube-ai-features-and-how-to-use-them/ 
  20. Apple seems to have a bit more in the way of ethics than some other companies. Now Tim Cook is calling for governmental regulation of artificial intelligence: https://www.extremetech.com/internet/ai-needs-government-regulation-apple-ceo-says
  21. Detecting antibiotic resistant bacteria using artificial intelligence: https://www.datanami.com/2023/11/22/oxford-researchers-use-ai-to-detect-antibiotic-resistance-faster-than-gold-standard-testing/
  22. Now authors of nonfiction works are suing over the use of their writing to train AI: https://www.computerworld.com/article/3711321/nonfiction-authors-sue-openai-microsoft-over-copyright-infringement.html
  23. Resources for learning about generative AI and large language models: https://www.kdnuggets.com/a-comprehensive-list-of-resources-to-master-large-language-models 
  24. Using machine learning to control experimental fusion reactors: https://www.theregister.com/2023/11/23/ntt_ai_nuclear_fusion_control/ 
  25. The risks of using code generated by AI: https://www.informationweek.com/machine-learning-ai/chatgpt-use-sparks-code-development-risks
  26. Ten principles to follow in order to successfully use artificial intelligence in your business: https://www.datanami.com/2023/11/23/beyond-the-hype-10-core-principles-for-ai-success/
  27. I considered moving into high school teaching myself once. It is a much more stable career choice than post-doc research: https://www.nature.com/articles/d41586-023-03612-3

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.