So that’s what it is! I was wondering why reducing context and summarising still makes it make mistakes and forget the steering. And couldn’t find explanation to why it starts ignoring instructions when context is not full at all. How did you find that tool call is what degrades it? Isn’t this a biggest problem there is and not just “design tension”?

Well it's a selection bias.

If an athlete breaks a world record, they're likely to do it again. Even though it's incredibly hard to break a world record.

It’s not quite clear that this project is- there’s no “Claude code” a program. There’s tui/gui app, harness, prompts, and LLM. so this is a harness part?

It's the harness/orchestration layer — the part that runs the agent loop, dispatches tool calls, and manages context.

Which kind of mosquito? European or Asian?

Strangely, I expected to see multiple large previous discussions about movies here, yet the largest has only nine replies. https://news.ycombinator.com/item?id=1021736

Yes, ChatGPT just recently started to add these engagement phrased follow-ups; “If you want, I can also show you one very common sign people miss that tells you…”

It’s Australia and bureaucracy is vindictive here. He’s got life in jail now and the dog was eurhanised.

The author likes to look at every concept from all sides, yet seemingly not aware about Token Notation (TOON) and almost wishing something like that existed…