3 ms·
Also, when you hit compaction at 200k tokens, that was probably when things were just getting good. The plan was in its final stage. The context had the hard-fo
by hombre_fatal 8mo ago
Also, when you hit compaction at 200k tokens, that was probably when things were just getting good. The plan was in its final stage. The context had the hard-fought nuances discovered in the final moment. Or the agent just discovered some tiny important details after a crazy 100k token deep dive or flailing death cycle.
Now you have to compact and you don’t know what will survive. And the built-in UI doesn’t give you good tools like deleting old messages to free up space.
I’ll appreciate the 1M token breathing room.
- roygbiv2 8mo agoI've found compactation kills the whole thing. Important debug steps completely missing and the AI loops back round thinking it's found a solution when we've already done that step.
- garciasn 8mo agoFor me, Claude was like that until about 2m ago. Now it rarely gets dumb after compaction like it did before.
- 8note 8mo agooh, ive found that something about compaction has been dropping everything that might be useful. exact opposite experience
- myrak 8mo ago[dead]
- s900mhz 8mo agoI find it useful to make Claude track the debugging session with a markdown file. It’s like a persistent memory for a long session over many context windows. Or make a subagent do the debugging and let the main agent orchestrate it over many subagent sessions.
- roygbiv2 8mo agoYeah I use a markdown to put progress in. It gets kinda long and convoluted a manual intervention is required every so often. Works though.