Daniel Severo

3.8K posts

Daniel Severo banner
Daniel Severo

Daniel Severo

@_dsevero

Researcher at Meta Superintelligence Labs - FAIR / PhD @UofT @VectorInst / Previously @GoogleAI @UFSC 🇧🇷

Montréal, Québec Katılım Mart 2011
1.1K Takip Edilen2.5K Takipçiler
Sabitlenmiş Tweet
Daniel Severo
Daniel Severo@_dsevero·
More positions opening up for research scientist (all levels) in Montreal! Req: published papers at top conferences (ICML/NeurIPS/ICLR/CVPR etc) on world modelling or action-conditioned video gen metacareers.com/profile/job_de… Feel free to DM me if you meet the requirements above!
Daniel Severo@_dsevero

The FAIR Montréal office is expanding! We're looking for folks to join us who are exciting about world modelling. Hiring for both recent grads as well as more experienced researchers. Feel free to reach out, but please read the posting first :) metacareers.com/profile/job_de…

English
1
3
46
10.1K
Daniel Severo
Daniel Severo@_dsevero·
@dudelbem O que achas de usar IA em peer review? NeurIPS 2026 tem auto review com uma LLM direto no open review. É interessante. Melhor do que a revisão do hipotético reviewer não-remurado mediano, mas muito pior do que os top 5 reviews que ja recebi.
Português
0
0
0
338
Luiz-Eduardo Del-Bem
Eu não consigo deixar de pensar o quanto o sistema atual de peer-review baseado em trabalho não-remunerado está quebrado. Já passou da hora de um sistema mais profissional.
Português
7
12
195
6K
Daniel Severo retweetledi
Jeannette Bohg
Jeannette Bohg@leto__jean·
World models are one of the most exciting frontiers for robotics and embodied AI right now and we're growing the team at Meta FAIR to push on it. Hiring Research Scientists in: 🇺🇸 Menlo Park 🇨🇦 Montreal Links below 👇 or find me at #RSS2026
English
4
12
168
24.2K
Daniel Severo
Daniel Severo@_dsevero·
Deus deve ser argentino mesmo, não é possível
Português
0
0
2
285
Daniel Severo
Daniel Severo@_dsevero·
@HaghifamMahdi Yeah and now they reversed the USAs red card and I guess that’s fine and not dirty at all 😂😂
English
0
0
1
32
Mahdi Haghifam
Mahdi Haghifam@HaghifamMahdi·
@_dsevero Exactly 🎯 I am surprised by the way people reacted to the match compared to 2022 Final. Argentine did the same thing.
English
1
0
0
45
Daniel Severo
Daniel Severo@_dsevero·
Clearly football is a very different sport for the european colonialists and south americans. Paraguay played as they always do, completely acceptable in SA given how superior France is. Brazil has been in the same situation as France many times (long ago). Just deal with it
English
1
0
2
440
Daniel Severo
Daniel Severo@_dsevero·
Classic comment from those who think statistics win matches in football. They don’t. It’s about the moment / opportunity and largely vibes. You know nothing
Rasmus Jarlov@RasmusJarlov

@Nomentor1 It was 40 degrees and sunny most of the game and the referee was ridiculous for not giving Paraguy one or two red cards. Normally, France would have won by five goals.

English
0
0
2
436
Daniel Severo retweetledi
Manudarg
Manudarg@ManUdArg·
No puedo estar más orgulloso del futbol paraguayo. Están haciendo enojar a los franceses solo haciéndolos jugar un partido normal de eliminatorias.
Español
399
1.7K
19.4K
246.3K
Daniel Severo
Daniel Severo@_dsevero·
This WC performance by Cabo Verde is gonna boost tourism. What a legendary investment.
English
1
0
3
409
Daniel Severo
Daniel Severo@_dsevero·
Today, Latin America has defeated the Axis of evil.
English
0
0
4
270
Daniel Severo
Daniel Severo@_dsevero·
The prettiest beach in Canada is uglier than the ugliest beach in Florianópolis.
English
1
0
2
418
Daniel Severo
Daniel Severo@_dsevero·
@poolio Your work inspired me a lot during my PhD! Excited to see what’s next. Congrats on the journey
English
0
0
1
811
Ben Poole
Ben Poole@poolio·
After 15 years (4 internships and exactly 8 years full-time), today is my last day at Google. Growing up as a scientist at Brain and DeepMind was an incredible privilege, and I'm so thankful for the people who made it such a special place. I went from editing decision trees by hand as an intern, to exploring representation learning, latent-variable models, and diffusion as an AI researcher, to kicking off a new wave in generative 3D with DreamFusion, to scaling up generative models with Veo, Genie, and Omni. We're just beginning to build systems that can understand and simulate the real world, and I'm excited to see what’s next 🧠🐸🚀
Ben Poole tweet mediaBen Poole tweet media
English
111
31
1.5K
119.7K
Quentin Garrido
Quentin Garrido@garridoq_·
After 4.5 formative years at FAIR, I am thrilled to join AMI Labs as a Member of Technical Staff ! I'm looking forward to pushing the boundaries of JEPA and World Models, alongside an amazing team of familiar and new faces 🚀
Quentin Garrido tweet media
English
13
4
152
14.3K
Daniel Severo
Daniel Severo@_dsevero·
Nothing more depressing than watching anglophones narrate football/soccer. Are you guys even watching the game? Are you being held hostage? What's wrong?
English
0
0
7
432
Daniel Severo retweetledi
Josh Woodward
Josh Woodward@joshwoodward·
🇧🇷World-changing AI companies are coming from Brazil. That’s why we’ve officially expanded our Google AI Futures Fund to Brazil, partnering with venture capital leader Monashees to launch the Gama Fund. We’re looking for an elite cohort of deep tech founders and will offer: - Early access to Google DeepMind models - Up to $2M in co-investment - $350k in Google Cloud & Gemini credits - Direct co-development with Google engineers at our new IPT Open campus hub Apply today: gamafund.com
Josh Woodward tweet media
English
54
54
598
52.8K
Daniel Severo
Daniel Severo@_dsevero·
This is not representative of the technical abilities of Brazilian engineers/scientists. This is not a reputable institution. They have no relevance nationally. Please ignore they exist, it’s just political BS :)
IplanRio@IplanRio_rj

NOTE ON RIO 3.5 OPEN In recent days, Rio 3.5 Open has received far more attention than we anticipated. Along with it came analyses and, of course, criticisms and questions. First, we want to clarify that the model is not foundational, trained from scratch, nor was it ever communicated as such. It is a post-training project built on open models, following classical approaches and some experiments. We started with open baseweights and applied various techniques, including merging, OPD, and finally used inference with SwiReasoning. It was precisely thanks to the community's attention that we identified an operational error in the publication process. We ended up making available an intermediate checkpoint that had not yet completed all the final validation and optimization steps. This generated interpretations that, looking back now, we fully understand. The checkpoint has been removed. We tried to recover the final model, but it was not possible. It will only be released after the new training and all external validations are completed. We also want to correct an important attribution point. Our team used public models provided by Alibaba, through Qwen 3.5, and by Nex-AGI, through Nex-N2 Pro, as a basis. In the initial documentation, we did not include Nex's important contribution. Correctly recognizing who builds these foundations is part of the open development process. Thank you, Nex, for your work and for contributing to advancing the state of the art in open models. It is worth contextualizing that there was no official release of that version of the model. The project ended up gaining traction organically and unexpectedly while it was still undergoing independent validations. In any case, this shows that there is interest and that Brazil has more space in this conversation than we usually imagine. We hope to see more initiatives emerging in Latin America, India, Africa, and other places that seek to expand their technological sovereignty, especially at a time when Fable has been closed to the rest of the world and access to frontier models has become part of the global strategic debate. Rio 3.5 Open is just the beginning. We will correct what is necessary, continue developing openly, and share what we learn along the way. Our goal is to show that the Brazilian public sector can also learn, build openly, and contribute to the forefront of current technological research.

English
0
1
3
405
Mat Velloso
Mat Velloso@matvelloso·
Just to recap: 1-Announced a SOTA model 2-Used a model without attribution, added a system prompt to omit the model name 3-Gets caught, aknowleged the mistake 4-Uploads the wrong model 5-Deletes the wrong model 6-Says the right model is now lost, so they will have to train it again 7-Mixed stories whether the funding came from a private or a government initiative 8-Coincidentally, elections are in October and Rio just announced a US$ 500M deal for a data center: temporealrj.com/data-centers-n…
IplanRio@IplanRio_rj

NOTE ON RIO 3.5 OPEN In recent days, Rio 3.5 Open has received far more attention than we anticipated. Along with it came analyses and, of course, criticisms and questions. First, we want to clarify that the model is not foundational, trained from scratch, nor was it ever communicated as such. It is a post-training project built on open models, following classical approaches and some experiments. We started with open baseweights and applied various techniques, including merging, OPD, and finally used inference with SwiReasoning. It was precisely thanks to the community's attention that we identified an operational error in the publication process. We ended up making available an intermediate checkpoint that had not yet completed all the final validation and optimization steps. This generated interpretations that, looking back now, we fully understand. The checkpoint has been removed. We tried to recover the final model, but it was not possible. It will only be released after the new training and all external validations are completed. We also want to correct an important attribution point. Our team used public models provided by Alibaba, through Qwen 3.5, and by Nex-AGI, through Nex-N2 Pro, as a basis. In the initial documentation, we did not include Nex's important contribution. Correctly recognizing who builds these foundations is part of the open development process. Thank you, Nex, for your work and for contributing to advancing the state of the art in open models. It is worth contextualizing that there was no official release of that version of the model. The project ended up gaining traction organically and unexpectedly while it was still undergoing independent validations. In any case, this shows that there is interest and that Brazil has more space in this conversation than we usually imagine. We hope to see more initiatives emerging in Latin America, India, Africa, and other places that seek to expand their technological sovereignty, especially at a time when Fable has been closed to the rest of the world and access to frontier models has become part of the global strategic debate. Rio 3.5 Open is just the beginning. We will correct what is necessary, continue developing openly, and share what we learn along the way. Our goal is to show that the Brazilian public sector can also learn, build openly, and contribute to the forefront of current technological research.

English
13
19
119
10.1K