At ISSCC, the premier conference for chip design, Igor Arsovski will explain our fast AI inference, which leverages a network of SRAM-fed LPUs to enable strong scaling across thousands of devices, providing access to >2TB of SRAM within <3 microseconds. https://
hubs.la/Q035rR630
RESEARCH
-

Groq’s Fast AI Inference with SRAM-Fed LPUs at ISSCC
By
–
-

Multi-Robot Collaboration in Autonomous Walking Robots by Boston Dynamics
By
–
Exploring multi-robot collaboration in #Autonomous walking #Robots
— Ronald van Loon (@Ronald_vanLoon) 6 février 2025
by @BostonDynamics #Robotics #RPA #MachineLearning #ArtificialIntelligence #ML #MI
cc: @karpathy @alvinfoo @HaroldSinnott pic.twitter.com/gHnQRs8YnzExploring multi-robot collaboration in #Autonomous walking #Robots
by @BostonDynamics #Robotics #RPA #MachineLearning #ArtificialIntelligence #ML #MI cc: @karpathy @alvinfoo @HaroldSinnott -

Post-Digital Society: Digital Technologies as Fundamental Enablers
By
–
A post-digital society is one in which digital technologies are fully integrated into daily life and are no longer seen as novel or disruptive but as fundamental enablers of human interaction, business, and governance. Microblog by @antgrasso #DigitalTransformation #PostDigital
-

Demystifying Long Chain-of-Thought Reasoning in LLMs
By
–
Demystifying Long Chain-of-Thought Reasoning in LLMs Yeo et al.: https://
arxiv.org/abs/2502.03373 #ArtificialIntelligence #DeepLearning #MachineLearning -

Advanced Foundation Models and AI Predictions for 2025
By
–
How advanced foundation models will expand what #AI can do (and other predictions for 2025)
by Martina Lauchengco @VentureBeat Learn more: https://
buff.ly/3VYRtjT #ArtificialIntelligence #MachineLearning #MI #Tech cc: @terenceleungsf @alvinfoo @bernardmarr -
Robot Safety Through Guardrail Constraints in DINO-WM
By
–
Robot safety through guardrail constraints in the representation space of DINO-WM https://t.co/NY0ogAnhyE
— Yann LeCun (@ylecun) 6 février 2025Robot safety through guardrail constraints in the representation space of DINO-WM
-

Reinforcement Learning AI Models Demand Infinite Computational Resources
By
–
And that’s before we flip into a version of training AI models through RL that has an infinite hunger for compute.
-
Synthetic Data Generation Research Advances in AI
By
–
We see synthetic data generation being used a bunch – cool new research in this area
-
OpenAI’s o3-powered Deep Research mode revolutionizes AI agent era
By
–
OpenAI’s surprise new o3-powered ‘Deep Research’ mode shows the power of the #AI agent era
by @carlfranzen @VentureBeat Learn more: https://
buff.ly/3WIAEtG #MachineLearning #ArtificialIntelligence #ML #MI #Tech #Technology cc: @yuhelenyu @miketamir @JimMarous -
Reasoning Models vs General LLMs Post-Training Techniques
By
–
Thanks for clarifying. E.g. in my view, "Pre-training, post-training, post-inference," would be too general because that's also the same flow for LLMs that are not specialized reasoning models. In a sense all reasoning techniques are post-training techniques because they all are