Weekly Review 5 January 2024

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

1. Generative AI has a large carbon footprint because of the electrical power it takes to train the models. Time for more server farms in countries like New Zealand, that get most of their electricity from renewables: https://www.theguardian.com/commentisfree/2023/dec/23/ai-chat-gpt-environmental-impact-energy-carbon-intensive-technology 
2. The arguments between AMD and Nvidia over benchmarking chip performance for AI: https://www.theregister.com/2023/12/21/nvidia_amd_benchmarks/?td=rt-3a
3. A lot of advances are being made in AI, but ethics in the field are still lagging behind: https://techcrunch.com/2023/12/23/this-week-in-ai-ai-ethics-keeps-falling-by-the-wayside/
4. Geoffrey Hinton is one of the progenitors of neural networks. Now they scare him: https://www.technologyreview.com/2023/05/02/1072528/geoffrey-hinton-google-why-scared-ai/
5. An overview of AI and neuromorphic chips: https://www.extremetech.com/extreme/333143-what-is-artificial-intelligence
6. Will quantum computers actually speed up machine learning? https://www.nature.com/articles/d41586-023-04007-0
7. What is going to follow generative AI? The founder of Deepmind says interactive AI: https://www.technologyreview.com/2023/09/15/1079624/deepmind-inflection-generative-ai-whats-next-mustafa-suleyman/ 
8. So the people who produce the training data for artificial intelligence are poorly trained, highly pressured, and it's surprising that they are using AI to do their jobs? Quality costs! https://www.technologyreview.com/2023/06/22/1075405/the-people-paid-to-train-ai-are-outsourcing-their-work-to-ai/
9. There needs to be ways of determining if an AI is conscious, and rules around how we treat them if they are: https://www.nature.com/articles/d41586-023-04047-6 
10. How content credentials will help guard against the abuse of AI generated material, especially around elections: https://spectrum.ieee.org/deepfakes-election
11. No, superintelligent AI is not going to just suddenly appear out of nowhere: https://www.nature.com/articles/d41586-023-04094-z
12. How a deep learning AI was used to discover new antibiotics for fighting drug-resistant bacteria: https://www.datanami.com/2023/12/21/mit-researchers-leverage-ai-to-identify-antibiotic-that-can-kill-drug-resistant-bacteria/
13. Some techniques for measuring the similarity between documents, an important step for learning in AI: https://www.kdnuggets.com/evaluating-methods-for-calculating-document-similarity
14. The quality of the data matter more than the size of the Neural Network: https://www.datasciencecentral.com/the-best-kept-secret-about-llms/