Weekly notes
- Here’s how I use LLMs to help me write code: Simon Willison using Claude and other llms to help write tiny programs
- Impressions of SRECon Americas 2025: Lots of ai but not necessarily lots of good uses for them
- How the world’s leading breach expert got phished: Even the people who know how to be really careful make mistakes given the right circumstances
- Two decades of Git: A conversation with creator Linus Torvalds: Interview with Linux Torvalds on the 20th anniversary of git. Enjoyed this immensely. :)
- AI: Where in the Loop Should Humans Go?: Lots of stuff in here to think about and links to papers worth reading. How do people relate to llms - are people subordinate or do we consider them assistants? What happens to us and the way we think about systems and solve problems if we spend all our time trying to prompt right answers out of an llm? And what happens when an llm is given a task that’s outside its wheelhouse? This and more here. So good.