Hantmango

从pi的源码开始，用我熟悉的python重写一遍

SFT tutorial comes out tomorrow! It’s a ~45 minute video that will go through instruction post-training end to end Synthetic local training data gen -> unsloth finetuning -> evals -> packaging SLMs into narrow little harnesses bonus: low-level guidance/constrained decoding

Claude Opus 4.7に隠れてあまり話題になってないけど、Qwen3.6-35B-A3Bかなりすごいモデルなのでは？

Qwen3.6 35B-A3B dropped yesterday, so I ran it on 4 GPUs to see how it performs: 🟣 RTX 3090 — 49.78 tok/s, TTFT 852ms 🟡 RTX 4090 — 118.93 tok/s, TTFT 686ms 🟢 RTX 5090 — 160.37 tok/s, TTFT 409ms 🔵 DGX Spark — 59.98 tok/s, TTFT 228ms I went with ollama as the backend because honestly, it's the easiest way for most people to get started. One command, model pulled, done. I used Q4_K_M (24GB) across all four cards. The reason is the 3090 and 4090 don't support NVFP4 (only the 5090 and DGX Spark could use it). Keeping the same quant everywhere felt like the fairest way to compare. And yes, you can absolutely squeeze more performance out of every card with vLLM, SGLang, or TensorRT-LLM. But that's not what this test is about. This is just the out-of-the-box experience for folks who own a GPU and want to try the new model tonight.

Google DeepMind researcher argues that LLMs can never be conscious, not in 10 years or 100 years. "Expecting an algorithmic description to instantiate the quality it maps is like expecting the mathematical formula of gravity to physically exert weight."

这种开源教材站的用法是这样的：遇到什么具体问题，问AI这是哪个领域的，然后去教材站找到相应的课本，送进notebooklm里，然后问答或者学习或者让Gemini出个skill。 相当于是科班prompt/skill，有各种经过同行评议的理论依据，不是随便找几个疯疯癫癫的行业翘楚，或者诹一堆玄而又玄的不知所云。

副主任王昌林：着力扩大国内有效需求，将制定2026年—2030年扩大内需战略实施方案，推动符合条件的重大工程项目尽早开工建设。

Prompt 满屏流，生成替手修。 数年间不写一行 code。 树下开 IDE 犹未稳，能几次，又回眸。 旧 repo 枝头，故人还在否？ 新架构多是旧烦忧。 欲开 IDE 审架构，终不似，少年游。