种族主义人渣、新纳粹马斯克正在宣扬白人要残酷对待非白人，否则白人成了少数，就会遭到非白人1千倍的残酷对待，就会遭到屠杀。加州现在白人就是少数，遭屠杀了吗？

Ben&Jerry冰淇淋的Ben在国会听证会上抗议美国政府帮助以色列在加沙搞种族灭绝和饿死加沙儿童，被逮捕。

柳智宇相信轮回是其信仰自由，却要胡扯有“科学依据”骗人，还要抬出“国外研究”（哪个国外，印度吗？）吓唬人，如此弱智，活该出家，何必还俗。

关于摄入欧米茄6脂肪酸亚油酸太多的风险，我以前已经说过，要尽量选择亚油酸含量低的植物油，但此人却直接造谣食用植物油中亚油酸含量最低、比橄榄油中的亚油酸含量还低的菜籽油，把种子油一棍子打死，他家喜欢危害大得多的猪油还是牛油？

The Trump administration has drafted a memo to Congress outlining its intent to end nearly all federal funding for public media, which includes NPR and PBS, according to a White House official who spoke to NPR.

川普新时代的美国文革：得克萨斯的共和党参议员认真学习伟大领袖川主席著作。

NIH突然取消了至少33项与疫苗有关的资助，并计划取消所有关于mRNA疫苗的研究资助，因为川粉害怕mRNA疫苗会改他们的基因。我早说了，川普、小肯尼迪上台必然会打压疫苗研究。这是不可估量的损失，不仅美国放弃了最先进的疫苗技术，而且对全人类都会是灾难。把川普选上台的川粉是反人类的罪人。

马斯克要求所有联邦雇员向他汇报过去一周的工作、否则就等于辞职的电子邮件就三句简单的话。美国国务院、财政部、国防部、联邦调查局都赶快给员工发函告诉他们不要回复该电子邮件。不知中央情报局、国家安全局员工要不要回复？联邦雇员工会也建议工会成员不要回复。马斯克要把他们都开除吗？

马斯克的传记作者说经过多年的调查，马斯克的智力平庸，智商在100-110，绝对不会超过110，而不像他自吹的140-170。哪里用得着花那么长时间调查。从马斯克高中毕业成绩数学B、语文60，就知道他只是一个装天才的蠢货。哪有高智商的人中学数学考不好的？

Woof

Very happy to show that we can do RL finetuning on 405B models with open-source code, beat Llama 405B instruct with their base model, and beat DeepSeek V3 too. Enjoy building off this teams hard work.

Here's Tulu 3 405B. A holiday present from @hamishivi.bsky.social @vwxyzjn.bsky.social and team.

Here is Tülu 3 405B 🐫 our open-source post-training model that surpasses the performance of DeepSeek-V3! It demonstrates that our recipe, which includes RVLR scales to 405B - with performance on par with GPT-4o, & surpassing prior open-weight post-trained models of the same size including Llama 3.1.

"But if an an AI company wants to fine-tune a general-purpose open weight model for a specific purpose, and that process uses a significant amount of computing, they would need to apply for a US govt license to do so in a Tier 2 country"

Controlling who in the world can finetune on what is dumb

Quick recap on the state of reasoning -- can LMs reason (I say yes, just different than humans)? How?
My talk at the NeurIPS Latent Space live event (pre o3).
Slides: https://buff.ly/40hsoTx
Post: https://buff.ly/40i2rDC
YouTube: https://buff.ly/40k8GH3

"The bitter lesson is over" in some NeurIPS discussions is a bad take.

Much like Moore's law being "over", we will continue to find other ways to continue to scale progress in these models by leveraging more, more efficient, more targeted compute.