playgroup

playgroup is a 1 day deep dive into 'the edge of LLMs', in this iteration we're looking at a kaggle challenge designed to show that LLMs have fundamental limitations. If you want to know 'where the limit is' and how to edge into it, attend this day.

Key points

• code (github), a shared workspace and LLM credits are provided
• the lovely venue offers hot breakfast, hot lunch and snacks throughout the day
• key aspects include - problem solving in groups, show-and-tells with tooling & process that attendees are using (e.g. Cursor, Claude Code, Aider), discussion about your issues, lots of networking with fellow cutting-edge attendees
• the event runs at-cost (organiser Ian makes no money)

About you - you've been using LLMs for over a year, you use them weekly either to write code or inside a product (ideally - you do both). You've seen them work well and poorly and you're curious to know where 'the edge is' and to learn new ways to deal with unreliable results. You want to collaborate with peers to learn new processes, see new tools and share insights.

I've written this up on LinkedIn here and here.

At the first event I had a selection of heads of data science and senior data scientists, all hands-on with LLMs, here's what they said:

Host or Publisher Ian Ozsvald