A ton of interesting tools are being built and open-sourced everyday, using LLMs at their core. Many of these tools are helpful and interesting, and some are changing how we work.

Something I’ve seen often in organizations is a desire to become “AI-enabled” by purchasing tools for the company. Doing this is like strapping a rocket to a skateboarder and saying “see you on the Moon.”

The skateboarder is now a fine mist over three counties. The skateboard achieved orbit. Call this “early traction.”

It’s too big a change, too fast for many people. The median result of AI use is confidently wrong work, almost indistinguishable from good work. This is a bad outcome for everyone.

Everyone’s learning curve for AI looks a bit different, and the best people I’ve seen using the tools are not always the earliest adopters. If you haven’t yet found useful applications of AI in your work and aren’t in an environment that is receptive to that feedback, I’d love to have a conversation.

It’s not too late to learn or to ask questions.