Shanshan Wang

@cooltxeel

Applied AI @Tensorlake | agents, sandboxes, orchestration, multimodal AI

Dallas, TX

Joined April 2014

Replies Reposts Media

Shanshan Wang
@cooltxeel
Sep 3
It’s a little funny that prompting best practices are now model-specific, and even version-specific... You’d think general-purpose AI models should get easier to use over time, but somehow we still need to learn the quirks of every new version. 🤣
Shanshan Wang
@cooltxeel
Sep 3
Interesting results. NoRA-init seems worth a quick test to see how it goes.
J.L
@Jo1uck
Sep 1
1/ How can we make Low-Rank Matrices behave more like Full-Rank Matrices—and achieve higher rank efficiency? We release our new paper: NoRA: Normalized Low-Rank Adaptation. Paper: arxiv.org/abs/2608.31036
Shanshan Wang
@cooltxeel
Aug 31
One thing I noticed after building several sandbox integrations: The harness/application still ends up managing parts of the sandbox lifecycle and the session-to-sandbox binding: reconnecting, resuming, handling restarts, and tracking which sandbox belongs to which session. For
Shanshan Wang
@cooltxeel
Aug 28
We just released the Tensorlake-OpenCode plugin v0.3.0 🚀 npmjs.com/package/tensor… It supports more tool calls and subagents. Most importantly, it now enables project sync to the sandbox using cloud volume and Git, so your work stays persistent across local and sandboxes.
00:00
3
Shanshan Wang
@cooltxeel
Aug 26
Wow, this is fantastic. Going to try it.
Unsloth AI
@UnslothAI
Aug 25
You can now fine-tune Qwen3.8-27B for free with our notebook! 🔥 Local training works on 24GB VRAM. Unsloth trains Qwen3.8-27B 1.5x faster with 50% less VRAM. GitHub: github.com/unslothai/unsl… Qwen3.8-27B Notebooks + Guide: unsloth.ai/docs/models/qw…