Stealing Reasoning Traces from Proprietary LLM APIs

Re:Stealing Reasoning Traces from Proprietary LLM APIs by supriyo-biswas on Tuesday August 11, 2026 @01:05PM (#49261253)

It’s about being able to change models mid-task. For example, I want to be able to plan using Fable but implement the plan using Sonnet, and that won’t work if this is implemented.

[ Parent | on HN ]

Re:Stealing Reasoning Traces from Proprietary LLM APIs by SaltyBackendGuy on Tuesday August 11, 2026 @01:14PM (#49261387)

Or even my fable credits run out mid task and need to switch back to opus >.<

[ Parent | on HN ]

Re:Stealing Reasoning Traces from Proprietary LLM APIs by QuercusMax on Tuesday August 11, 2026 @01:55PM (#49261981)

Prior to LLMs I never considered that I might have to make a resource-usage decision between hiring Star Trek's Data vs. his stupider brother B4...

https://memory-alpha.fandom.com/wiki/B-4

[ Parent | on HN ]

Re:Stealing Reasoning Traces from Proprietary LLM APIs by TeMPOraL on Tuesday August 11, 2026 @05:47PM (#49264952)

Star Trek is a post-scarcity society, those problems don't exist there unless you're in the middle of a crisis and on emergency power.

LLMs briefly seemed like this too, after subscriptions made the SOTA models too cheap to meter, but before they walked back on that and introduced quotas...

[ Parent | on HN ]

Re:Stealing Reasoning Traces from Proprietary LLM APIs by gs17 on Wednesday August 12, 2026 @04:32AM (#49269366)

Actually, that brings up a good reason they can't fix it. Fable falls back to Opus when the topic is too "unsafe". That behavior requires traces than can move between models!

[ Parent | on HN ]

Re:Stealing Reasoning Traces from Proprietary LLM APIs by crabette on Tuesday August 11, 2026 @02:17PM (#49262321)

For plan it's relatively easy, just make the plan the artifact. The point is to ingest knowledge with one model and use it in another, and that is not necessarily easily expressible in natural language.

[ Parent | on HN ]