Keep on to blur preview images; turn off to show them clearly
Thread
We're sharing/showcasing best of @github projects/repos. Follow to stay in loop. Promoting Open-Source Contributions. UNOFFICIAL, but followed by github
RT @doodlestein: As I've been working more today with my coding agents while they coordinate work using my new MCP Agent Mail system, I'm s…
Former Quant Investor, now building @lumera (formerly called Pastel Network) | My Open Source Projects: https://t.co/9qbOCDlaqM
3/3 settings that give you control over preserving your agents tool calls and thinking blocks You can read more here.
AI @amazon. All views personal!
预热一下， 明天下午1点开始。 以后每周五都会有一场这样的圆桌会， 和大家一起交流学习。
Growth Coach｜Helping creators build their personal brand on X 增长教练｜帮助创作者在X上打造个人品牌 公众号：PandaTalk8 X 增长群群主
2/3 It also impacted Interleaved thinking with Tool use (off course)
3/3 settings that give you control over preserving your agents tool calls and thinking blocks You can read more here.
1/3 What just Anthropic did that is great for Agents? It was a big problem to benefit from prompt caching, while using Extended Thinking, Interleaved Thinking with Tool calling etc. They used to make a choice for users, that on every next 'user message' they used to clear "Extended thinking" blocks to manage the context length (to prevent it from growing rapidly) However, this used to result in cache invalidation (because cached input doesn't any longer match new input - negating any benefits of caching). Now Anthropic gives control to users on how they want old "Extended Thinking" and "Tool use" blocks want to be preserved.
2/3 It also impacted Interleaved thinking with Tool use (off course)