@thedankoe

Article

Everyone is racing to be successful right now....

@waterloo_intern

Article

...

@LunaGitana0333

Thread

La historia de Roma se divide en tres grandes períodos: La monarquía, la república y el imperio. La Monarquía: Según la leyenda, Roma fue fundada en el año 753 a.C. por Rómulo y Remo, dos gemelos criados por una loba. Los primeros reyes de Roma fueron de origen etrusco y fueron los encargados d...

@WealthCoachMak

Thread

If I had $75,000 and I wanted to keep my trading strategy really simple with not much time and effort? I would pick 1 HIGH quality stock $AMZN Here would be my system:...

@crismartinj

Thread

KISSINGER: EL NOBEL DE LA PAZ QUE SEMBRABA GUERRAS El 22 de agosto de 1973, Richard Nixon nombró secretario de Estado a Henry Kissinger. No llegaba un diplomático cualquiera. Llegaba al Departamento de Estado uno de los hombres fundamentales de Bilderberg, miembro de su Comité Directivo, íntimament...

@GoodAMLiberty

Thread

Ron Paul was booed at the Republican debate for saying Iran didn't have a weapon (they didn't). Then he made one of my all time favorite debate moments:   "If they're so determined to go to war, the only thing I plead with you for if this is the case, is do it properly. Ask the people and ask the...

@navaneethvb

Thread

Apparently, one way to make LLM inference almost 2x faster is to make the CPU stop talking to the GPU so much. And this is where CUDA graphs come in, but before getting into the details let us actually understand the problem first. An LLM forward pass is not one giant GPU operation. Underneath it,...

@aaravwattal

Thread

As we scale to trillion parameter models and gigawatt-scale training, we don't do interesting work on just one chip anymore. Everything interesting happens across thousands of them, and moving data between those chips has become a fundamental bottleneck of both training and inference. I spent this ...