Open Swimcast
Another little app I only built because it was essentially free to build - this one gets my podcast feed onto a pair of underwater swimming headphones.
Another little app I only built because it was essentially free to build - this one gets my podcast feed onto a pair of underwater swimming headphones.
Another little bespoke app I built for myself - this one translates English and German live so I can talk to my sister-in-law.
I gave a talk at AI Practitioner London on what 'good enough' looks like when you can no longer review every line an AI writes.
Continuing to explore Anthropic's Agent SDK - making tool calls more descriptive and learning about Python introspection.
Trying out Anthropic's new Claude Cowork mode.
Learning in public: experimenting with Anthropic's Agent SDK after hearing that anything Claude Code can do, you can do with the SDK.
One of the nice things about time off is the chance to play a little.
This is my recap of Hamel and Shreya's LLM evaluation course. I'm hoping I come back here in the future every time I need to remind myself of how to do this the right way.
Notes from the final lesson of Hamel and Shreya's LLM evaluation course - practical strategies for improving accuracy and reducing costs through prompt refinement, architecture changes, fine-tuning, and model cascades.