A consistent theme we’ll talk about today is the idea of taking cutting-edge research and making it a reality. Here are a few examples of research efforts making their way into new and existing products
RESEARCH
-

Future of AI Thinking: Livestream Insights
By
–
Tune into the livestream to hear how we’re thinking about the future of AI https://t.co/XrdymSTRIA
— Google AI (@GoogleAI) 20 mai 2025Tune into the livestream to hear how we’re thinking about the future of AI
-

ARC-AGI-2 Paper Published: Design, Challenges, Human and Model Performance
By
–
We just posted our paper on ARC-AGI-2 — covering its design principles, what makes it more challenging, a detailed analysis of human performance, and current model performance.
-
Sovereign α-AGI: Self-Evolving On-Chain Agentic Swarms
By
–
"α‑AGI Sovereign" Autonomous, on-chain Meta-Agentic enterprises deploying self-evolving agentic swarms—strategically converting global inefficiencies into unprecedented value, implicitly reshaping the very foundations of economic reality. GitHub :
-
Adjoint Sampling: New Scalable Algorithm for Reward-Based Generative Models
By
–
Introducing Adjoint Sampling, a new learning algorithm that trains generative models based on scalar rewards.
— AI at Meta (@AIatMeta) 20 mai 2025
Based on theoretical foundations developed by FAIR, Adjoint Sampling leads to a highly scalable practical algorithm, and can become the foundation for further research… pic.twitter.com/RyVVzbEfBjIntroducing Adjoint Sampling, a new learning algorithm that trains generative models based on scalar rewards.
Based on theoretical foundations developed by FAIR, Adjoint Sampling leads to a highly scalable practical algorithm, and can become the foundation for further research -

Simula 67: First Language With Classes and Abstract Data Types
By
–
In May 1967, two computer scientists presented a paper about Simula 67, the first programming language to include abstract data types and classes: http://
bit.ly/2RqFkGn -

Visual Planning: Thinking Only with Images Research
By
–
Visual Planning: Let's Think Only with Images Xu et al.: https://
arxiv.org/abs/2505.11409 #ArtificialIntelligence #DeepLearning #MachineLearning -

Keras Release: Model Sharding, Muon Optimizers, New Ops
By
–
New Keras release is out — includes saved model file sharding for very large models, the Muon optimizers, new ops and many performance improvements
-
Stability AI Releases SV4D2.0 Video Generation Model
By
–
Weights, code, and paper @HuggingFace
: https://
huggingface.co/stabilityai/sv
4d2.0
… @Github
: https://
github.com/Stability-AI/g
enerative-models
… @arXiv
: https://
arxiv.org/abs/2503.16396 (4/4) -
4D Asset Generation Advances for Professional Production Workflows
By
–
This upgrade marks progress toward 4D asset generation for professional production workflows, from generating sprite sheets for in-game characters, to supporting assets for film and virtual worlds. Multi-view generation, however, remains complex due to the inherent ambiguity of