Weekly Review 14 June 2024

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

1. Europe is getting into the AI fighter / wingman space as well: https://www.theregister.com/2024/06/05/airbus_wingman_drone/
2. Just because it has AI in it, doesn't mean it's actually useful: https://www.theverge.com/2024/6/6/24172718/humane-ai-pin-selling-hp-rumor-1-billion
3. Biological brains are so much more energy efficient than computers. And it's not like they're less reliable than today's AI: https://www.theregister.com/2024/06/05/a_high_five_for_stanford/
4. More warnings on the danger AI poses. But nothing about their energy demands contributing to climate change! https://dataconomy.com/2024/06/05/another-day-another-ai-warning-that-authorities-wont-care/
5. Employees at all companies, not just AI companies, should be able to blow the whistle on dangerous activities: https://www.stuff.co.nz/world-news/350301251/ai-employees-warn-technologys-dangers-call-sweeping-company-changes
6. Most of what CEO do can now be done by AI: https://futurism.com/the-byte/ceos-easily-replaced-with-ai
7. More specialised hardware for AI: https://www.nature.com/articles/d41586-024-01544-0
8. AI is very energy hungry. Wouldn't it be better to focus on proven energy technologies, rather than fusion? https://www.theregister.com/2024/06/06/openai_negotiating_nuclear_power_deal/
9. AI driven atmospheric and pollution modelling: https://www.nature.com/articles/d41586-024-01677-2
10. DuckDuckGo's AI chatbot has an off switch. And there was much rejoicing. https://www.theregister.com/2024/06/06/duckduckgo_ai_chat/
11. Things to do to stop your AI chatbot going off the rails: https://www.datanami.com/2024/06/05/whos-watching-your-genai-bot/
12. Don't want Microsoft's Recall AI spying on everything you do? Use Linux instead: https://www.lifewire.com/install-linux-avoid-ai-8655664
13. The impact of AI on language education: https://www.insidehighered.com/news/tech-innovation/artificial-intelligence/2024/06/06/ai-adds-hope-and-concern-foreign-language
14. An AI note taking app from Google: https://www.theverge.com/2024/6/6/24172422/google-notebooklm-ai-gemini-pro-chatbot
15. More on the demand for AI company workers to feel safe about being whistleblowers: https://www.techrepublic.com/article/openai-google-deepmind-right-to-warn/
16. Opt out if you don't want Meta to use your data to train its AI. Privacy lawyers start rubbing their hands in glee: https://www.theregister.com/2024/06/06/meta_ai_complaints/
17. So how similar does an AI generated voice have to be to a person's natural voice for it to be a copy? https://futurism.com/the-byte/forensic-similarities-openai-voice-scarlett-johansson
18. The data gathered by Microsoft's Recall AI can be easily stolen: https://www.wired.com/story/total-recall-windows-recall-ai/
19. How to build an LLM AI application with Python: https://www.kdnuggets.com/beginners-guide-to-building-llm-apps-with-python
20. How AI could evolve in the coming years: https://bigthink.com/the-future/3-phases-of-ai-evolution-this-century/
21. When AI bots are able to make Dad jokes, and cynical / sarcastic comments, maybe then they'll be able to replace me in meetings: https://www.theguardian.com/technology/article/2024/jun/05/the-future-is-sending-ai-avatars-to-meetings-for-us-says-zoom-boss
22. AI is more likely to cause human extinction via accelerating climate change, but I don't think these tech employees mean that: https://www.computerworld.com/article/2138467/once-again-ai-industry-employees-warn-the-tech-could-lead-to-human-extinction.html
23. Interpreting neural networks has always been difficult, but the enormous deep neural networks used by large language models are likely to remain uninterpretable for some time:  https://futurism.com/sam-altman-admits-openai-understand-ai
24. The impact of large language model AI on three different professionals: https://spectrum.ieee.org/large-language-models-2668430044