开启时会模糊预览图，关闭后正常显示
RT @indie_maker_fox: 🎁 MkSaaS官网更新 & 分销模式大改进！ MkSaaS使用的是Affonso平台进行分销，最近它更新了一个功能，Program Coupon，这个功能真的很有用。 过去，购买者点击了推荐人的 aff 链接，只有推荐人可以拿到…
🔥 The best AI SaaS boilerplate - https://t.co/VyNtTs0jSX 🚀 The best directory boilerplate with AI - https://t.co/wEvJ1Dd8aR 🎉 https://t.co/bh1RxeERuY & https://t.co/zubXJCoY92 & https://t.co/tfQf8T7gGF
It should be very easy to find me in #Bengaluru. I am right next to Bombay Vadaa Paav😅
AI researcher & teacher @SCAI_ASU. Former President of @RealAAAI; Chair of @AAAS Sec T. Here to tweach #AI. YouTube Ch: https://t.co/4beUPOmf6y Bsky: rao2z
昨天说泄露的NVIDIA-Nemotron-3-Nano-30B-A3B结果今天发布啦！ 这是一个结合Mamba-2 和 Transformer的MoE模型，可以开启或者关闭推理，知识截止日期是2025年11月28日。 最亮眼的成绩应该是数学能力测试 AIME25 刷到了89.1，如果让这个模型使用工具，则能刷到99.2 （AIME25有几道题使用编程方法来解会更简单）。编程能力测试LiveCodeBench (v6) 则是68.3，从这个分数上来看属于能用的水平。科学推理能力 GPQA则是 73。 作为对比，Qwen3-30B-A3B-Thinking-2507 的AIME25成绩是85，LiveCodeBench (v6) 则是66. GPQA是73.4. 基本上Nemotron-3-Nano-30B-A3B是针对数学能力优化的模型。 不过Nemotron-3-Nano-30B-A3B还有个最大的好处是，它把预训练和后训练数据集全都开源了，所以作为用来学习和研究的模型是个不错的选择。 模型地址：
性能参数/1
As production gets faster and faster (writing, code, clips, research etc) the limiting factor will be our speed of consumption - How quickly we can read and learn - How quickly we scroll and watch And how those things guide our ability to reason and make decisions
Attention will be all you need
Lazy take: since watching Zootopia 2 is definitely not an «in-kind transfer» or anything of the sort but just straight up discretionary consumption, this can be used to eyeball actual PPP/disposable income
We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1
If it's Tuesday, it must be @MSFTResearch India day..🤔 (Hello #Bengaluru👋)
AI researcher & teacher @SCAI_ASU. Former President of @RealAAAI; Chair of @AAAS Sec T. Here to tweach #AI. YouTube Ch: https://t.co/4beUPOmf6y Bsky: rao2z