Sublime
An inspiration engine for ideas
> We find that LLMs’ intelligence [...] almost linearly correlates with their ability to compress external text corpora
> compression efficiency [...] serves as a reliable evaluation measure that is linearly associated with the model capabilities
IQ for LLMs just dropped https://t.co/B9UOeLctRS
We are now at the superhuman AI phase!
@eyelevelai's GroundX, and enterprise-grade RAG system, outperformed humans on DocBench — a benchmark that tests deep document comprehension.
GroundX is an open-source system that you can run on your servers (or any cloud provider, as long https://t.co/s6n67blQUK
随着特朗普发币，市场有点fomo总统币了。盯着各路总统X账户，给大家找出一些领导人的账户
@nayibbukele 萨尔瓦多总统
@JMilei 阿根廷总统
@narendramodi 印度总理纳伦德拉·莫迪
@LulaOficial 巴西总统路易斯·伊纳西奥·卢拉·达席尔瓦
@ZelenskyyUa 乌克兰总统弗拉基米尔·泽连斯基
@lopezobrador_ 墨西哥总统安德烈斯·曼努埃尔·洛佩斯·奥夫拉多尔
@RodrigoChaves... See more
GemFactoryx.comHypercycle
docs.google.com
AI Overlords
Shachaf Rodberg • 2 cards
Using the term CogSec already shows low cognitive security. It means you're already thinking about it as a military, which is a misstep in the wrong direction from the very start. If you wish to maintain cognitive security, think of it as a set of ingredients to bake an apple pie, not a technology you might deploy in a battle field. The former is l... See more
Captain Pleasure, Andrés Gómez Emilssonx.com
