AI advancements take center stage, with Lex Fridman’s latest episode exploring massive AI clusters (and everything from DeepSeek to xAI), OpenAI unveiling its new O3-mini “reasoning” model, and a Forbes piece delving into Google’s Notebook LM project. We also have a look at LLM jailbreak tactics, the emerging limits of chatbot software, and a worrying exploit where hackers can plant “false memories” in ChatGPT to steal data indefinitely.

Shifting gears, GDC 2025 highlights the role of urban planning in game design, while an inventive “tiny vane” tech promises big fuel savings in aviation. In finance news, El Salvador has dropped Bitcoin as legal tender after a turbulent run, and an innovative Polish city deploys mussels to track water quality. Finally, one researcher’s deep dive into location tracking through in-app ads shows just how much privacy we might be giving up.


AI:

Miscellaneous:

  1. GDC 2025 - Urban Planning in Games -> https://www.youtube.com/watch?v=Q3Scw3dzxWE
  2. Tiny vanes glued to planes promise big savings -> https://newatlas.com/military/ticky-little-vanes-bring-big-savings-us-air-force/
  3. El Salvador Abandons Bitcoin as Legal Tender After Failed Experiment -> https://ticotimes.net/2025/02/02/el-salvador-abandons-bitcoin-as-legal-tender-after-failed-experiment
  4. This Polish city is using mussels to monitor water quality -> https://www.awa.asn.au/resources/latest-news/technology/innovation/polish-city-using-mussels-monitor-water-quality
  5. Everyone knows your location: tracking myself down through in-app ads -> https://timsh.org/tracking-myself-down-through-in-app-ads/

12 / 02 / 25 | Weekly Logs #1

Exploring massive AI clusters (and everything from DeepSeek to xAI), OpenAI unveiling its new O3-mini “reasoning” model, and a Forbes piece delving into Google’s Notebook LM project. We also have a look at LLM jailbreak tactics, and planting “false memories” in ChatGPT