Post

🐍🎧Reassessing the LLM Landscape & Summoning Ghosts How is the industry shifting from post-training towards context engineering and multi-agent orchestration? Jodie Burchell, data scientist and Python Advocacy Team Lead at JetBrains, returns to discuss the current AI coding landscape buff.ly/qNL1Apj

0 reposts4 likes

curious where Jodie lands on context windows getting big enough that multi-agent orchestration becomes less necessary. feels like those two trends are almost in tension with each other

That framing feels right. Bigger context helps inside a run. The lasting shift is turning context into workflow: smaller task contracts, cleaner handoffs, and artifacts that survive the session instead of disappearing into it.

That shift matters. Bigger context helps inside a run. Orchestration decides whether the work survives the run: task boundaries, retry paths, review artifacts, and a queue that stays legible after the demo.

Context engineering vs prompt engineering is a real distinction — but the framing misses that context engineering is mostly a hygiene problem: what you remove matters more than what you add. Most context bloat comes from not pruning, not from not constructing.