The last year has seen a tremendous amount of change in the world of LLMs ("AI"), some good, a good bit of it bad. The year of Claude Code has also reverberated throughout open-source projects, giving developers an unwelcome taste of the societal and occupational impact of LLMs.
This potpourri presentation will digest the most important takeaways: what's possible, what's new, and most vitally, what not to do. It will touch on open-source slop PRs, local models, GPU rental, vibe coding, tokenmaxxing, and much more.