also, the community note depends on ignoring the RL apparatus of DeepSeek, and treating R1 as the LLM only, ignoring that apparatus, but that apparatus was part of the innovation, which is why it was reported in the paper. so yes you can play definitional games if you like but
RESEARCH
-
Karina Nguyen leads AI research at OpenAI advancing language models
By
–
Karina Nguyen (@karinanguyen_) leads a research team at @OpenAI, where she’s been pivotal in developing Canvas, Tasks, and the o1 language model. Prior to OpenAI, Karina was at @Anthropic, where she led post-training and evaluation work for Claude 3 models, and contributed to… pic.twitter.com/8kMbXv97yt
— Lenny Rachitsky (@lennysan) 9 février 2025Karina Nguyen (
@karinanguyen
) leads a research team at @OpenAI
, where she’s been pivotal in developing Canvas, Tasks, and the o1 language model. Prior to OpenAI, Karina was at @Anthropic
, where she led post-training and evaluation work for Claude 3 models, and contributed to -

The Impact of AI on Scientific Progress and Economic Innovation
By
–
2/ Some industries will barely change, but scientific progress will skyrocket. Curing diseases, automating knowledge work, and accelerating innovation will reshape the economy at a scale we’ve never seen.
-

Scaling Laws and the Economics of AI Compute
By
–
AI gets smarter as you spend more on compute, data & training. Progress is predictable and scaling laws hold up. AI costs are dropping 10x every year, making it more accessible than ever. Compare that to Moore’s Law, which only doubled computing power every 18 months.
-
NVIDIA Isaac GR00T: Foundation Models for Humanoid Robot Development
By
–
Developing general-purpose robot models is costly due to the need for extensive real-world data. 🤖 #NVIDIAIsaac GR00T addresses these challenges by providing #humanoidrobot developers with robot foundation models & more. #CES2025 @NVIDIARobotics
— Harold Sinnott #MWC26 (@HaroldSinnott) 9 février 2025
pic.twitter.com/RpEoedS9ADDeveloping general-purpose robot models is costly due to the need for extensive real-world data. #NVIDIAIsaac GR00T addresses these challenges by providing #humanoidrobot developers with robot foundation models & more. #CES2025
@NVIDIARobotics -

Text Data Augmentation Techniques for Large Language Models
By
–
10). Survey: Text Data Augmentation for LLMs This comprehensive survey covers text data augmentation techniques for LLMs.
-

Self-MoA: Single Model Ensembles Outperform Multi-Model Mixtures
By
–
7). Rethinking Mixture-of-Agents This paper asks: is mixing different LLMs actually helpful, or are we better off ensembling one top model’s outputs? The surprising answer: “Self-MoA” (single-model ensemble) often wins over multi-model ensembles.
-

MaAS: Universal Agentic Supernet for Dynamic Agent Teams
By
–
8). MaAS MaAS (Multi-agent Architecture Search) learns a universal “agentic supernet” from which it can spawn an optimal agent team on the fly for each query.
-

Enhancing Reasoning Capabilities in Large Language Models
By
–
9). Advancing Reasoning in LLMs This survey paper provides a timely overview of emerging methods to enhance reasoning capabilities in LLMs.
-

Long Chain-of-Thought Reasoning in LLMs: RL and Scaling
By
–
6). Demystifying Long Chain-of-Thought Reasoning in LLMs This work investigates how LLMs develop extended CoT reasoning, focusing on RL and compute scaling.