leegang
1.4K posts

leegang
@bestmrlee
https://t.co/fQ7RS94AgO | AI tools that treat scientists like their own kids (tired researchers, come let mom fix it) Antai AI MBA • ex-DXY & Ping An Good Doctor • INTP












@karpathy @jeffreyhuber Mine is similar LLM is the CPU. The context filesystem is RAM plus disk. Tools are the peripherals. The agent runtime/harness is the operating system itself, responsible for scheduling, permissions, execution loops, and state management.





我最近开源的五个 Claw 项目👇 1. WeClaw - 在微信 ClawBot 连接多个 Agent 的桥接器 github.com/fastclaw-ai/we… 2. ClawHost - 基于集群的 OpenClaw 多租户部署面板 github.com/fastclaw-ai/cl… 3. ChatClaw - Agent 团队版 OpenClaw Web UI github.com/fastclaw-ai/ch… 4. FastClaw - 云原生部署友好的 OpenClaw 替代框架 github.com/fastclaw-ai/fa… 5. AnyClaw - 给 Agent 使用的互联网工具转接头 github.com/fastclaw-ai/an… ------ 你在用哪个?你最看好哪个?🤔

Introducing TurboQuant: Our new compression algorithm that reduces LLM key-value cache memory by at least 6x and delivers up to 8x speedup, all with zero accuracy loss, redefining AI efficiency. Read the blog to learn how it achieves these results: goo.gle/4bsq2qI

💀Fine-tuning is DEAD Their new paper Agentic Context Engineering (ACE) - A self-improving AI learns entirely from execution feedback. 87% reduction in adaptation latency. Open-source model. Beats proprietary enterprise agents. → Generator does the work → Reflector analyzes every failure → Curator applies surgical "delta updates" — one rule at a time → Zero human labels. Zero retraining. Zero fine-tuning bills. You no longer pay to retrain a model every time your use case shifts. The prompt becomes a living playbook that rewrites itself. Tested on AppWorld leaderboard. Matched the most expensive enterprise agents. On the hardest benchmarks, it beat them. I think the economics of AI just changed.


Introducing TurboQuant: Our new compression algorithm that reduces LLM key-value cache memory by at least 6x and delivers up to 8x speedup, all with zero accuracy loss, redefining AI efficiency. Read the blog to learn how it achieves these results: goo.gle/4bsq2qI









