Safety built into LLMs creates new risks. “malware developers added nuclear & biological weapons text to to their spyware. Goal? To trigger LLM safety refusals… so that their spyware wouldn’t be analyzed by an AI security scanner.” ~ learn more
The Promised Land for education has been 1:1 tutoring since Aristotle tutored the young prince Alexander the Great. “Exponentials are hard for even math teachers to internalize.” The piece argues that most edtech is still procedural, while LLMs are the […]
Karpathy is the AI sherpa our generation needs. “It’s so interesting to watch an agent relentlessly work at something. They never get tired, they never get demoralized, they just keep going and trying things where a person would have given […]
Goodbye SEO, hello AEO. The shift from SEO to Answer Engine Optimization (AEO) is underway. “As humans shift FROM ‘keywording’ for 10-blue links to do all the work themselves TO typing resolution questions to have the LLMs do nearly all […]
Which table format is best? Reminder that LLMs are a new tool and we need to try hard to learn how to get the best out of them. “Markdown-KV came out top, hitting 60.7% accuracy and landing roughly 16 points […]
A bitter lesson for the bitter lesson fanboys. This is a really great interview by Dwarkesh Patel, who I think is really on a roll the past ~year. “Richard Sutton is the father of reinforcement learning, winner of the 2024 […]
Microsoft launches Copilot AI function in Excel, but warns not to use it. I certainly am not yet ready to trust LLMs with my spreadsheets. “However, Microsoft specifically warns not to use it for “any task requiring accuracy or reproducibility,” […]
Coding with LLMs in the summer of 2025. “Despite the large interest in agents that can code alone, right now you can maximize your impact as a software developer by using LLMs in an explicit way, staying in the loop. […]
A million tiny whispers. This is a good podcast interview with a very clever content pro I used to work with, Logan Freeman. “Logan shares tactical approaches for optimizing content for LLMs (large language models), including using FAQ schemas, focusing […]
The last six months in LLMs, illustrated by pelicans on bicycles. “There are plenty of benchmarks full of numbers. I don’t get much value out of those numbers. There are leaderboards, but I’ve been losing some trust in those recently. […]