Leo Casarsa

Principal Data Scientist at Quantumblack Germany. All views and opinions are my own. He/Him 🏳️‍🌈🇧🇷

[Leo Casarsa @leocasarsa]

@check24de ich hätte gerne ein Trikot und ein Handtuch

[Leo Casarsa @leocasarsa]

@GBRChris_A @Maanvis @KatyMontgomerie I disagree with you, but that's irrelevant for the conversation. They are vulnerable people and should be protected. They pose no harm to cis women, but cis men pose harm to them. Thus the solution of trans and cis women sharing toilets seem quite obvious, in my opinion.

[Leo Casarsa @leocasarsa]

On this very interesting paper, authors argue on the possibility of emergency of Theory of Mind (ToM) in large language models. However, I wonder if the paper is instead accidently pointing out limitations of the current tasks used to identify ToM.

[Cameron R. Wolfe, Ph.D. @cwolferesearch]

Are large language models (LLMs) capable of understanding others' mental states? Recent research indicates that the answer might be yes, showing that LLMs possess mental models comparable to those of young children on select tasks. 🧵 [1/8]

[Leo Casarsa @leocasarsa]

@wojczarnecki Have you thought about opening it up? I.e., creating a meta-version of the game which other people could customize to find romantic partners for themselves?

[Leo Casarsa @leocasarsa]

@AlexisCompBio @gebirgsziege I'm curious to learn what the evidence is for swearwords being an indicator of emotional involvement as the cause of the difference in the distributions.

[Leo Casarsa @leocasarsa]

@VictorTaelin @ronymmoura @paulohpontes @Agothosp @sseraphini Usa a publicidade a seu favor. Eu acabei de conhecer vocês por causa do hate e achei seu projeto super interessante. O mesmo vai acontecer com gente mais importante que eu 😉

[Leo Casarsa @leocasarsa]

@ararunaar @VictorTaelin Nah, it was fine exactly as it is

[Leo Casarsa @leocasarsa]

"Com a iniciativa, o Serrapilheira e a FAPERJ pretendem promover novas linhas de pesquisa em ecologia formuladas por pós-doutores negros e indígenas que almejam obter, no médio prazo, posições permanentes."

https://t.co/cIOuF6fy77

[Leo Casarsa @leocasarsa]

@m_elantkowski @FrankRHutter To endorse the point above - when I first read the tweet, the language made me think it was satire. Glad to see it's for real and interested to learn more!

[Leo Casarsa @leocasarsa]

@daniela_witten Was the logistic regression run on R or Python? It's only statistics in the first case 😝

[César A. Hidalgo @cesifoti]

If you built your own government program, would it be similar to that of Lula, Bolsonaro, or something in between?

https://t.co/lzr3PxTacw is a site where you can create a program combining proposals from different candidates. It also shows you ... /1

[Leo Casarsa @leocasarsa]

@MayLiSilva2 Fuck! I'm so sorry, May-Li!

[Leo Casarsa @leocasarsa]

@MuhammadAslam37 @PhTaheri I don't understand your argument - are you implying that gatekeeping poorer students from PhD programs is a good solution, since it decreases competition?

[Leo Casarsa @leocasarsa]

@TeaBagTroopers @terekacollector @PassarHariky @vvvorvvtorvitor @QuattroRudi @kasmith87 @tarasowski Getting back at the original comment (which is by no means centered in Germany), I fully endorse that we should be building more cities as Berlin. And I also grant that Berlin has room to improve.

[Leo Casarsa @leocasarsa]

@TeaBagTroopers @terekacollector @PassarHariky @vvvorvvtorvitor @QuattroRudi @kasmith87 @tarasowski I'm really sorry for your accident. That's awful and shouldn't have happened.

That said, there's a large difference between "there are worse places" and "most places are much worse". Berlin can be a global example for public mobility and not be good enough at it simultaneously.

[Caio Fazuely Gomes | #BLM @caiocgomes]

Me falaram que eu tinha interpretado errado. Esse rapaz interpretou da mesma maneira que eu.

[Leo Casarsa @leocasarsa]

@zhuo1108 @md_anderson03 @academic_exit Good bosses do care about their team members - that's the not-so-secret secret to high performing teams. The problem is that academia does not select for good bosses, it selects for high performing individuals.

[Senior PowerPoint Engineer @ryxcommar]

You're doing data science all wrong.

Here are 5 tools every data scientist needs to be using, a 🧵: