Counter arguments showing that it’s an easy-to-vary, unfalsifiable theory.
RESEARCH
-
IEEE Announces 2025 Fellow Class Recognition
By
–
Congratulations!!! IEEE Fellow Class of 2025 announcement. https://
ieee.org/content/dam/ie
ee-org/ieee/web/org/about/fellows/fellow-committee/2025-fellows-class-announcement.pdf
… -
Densing Law of LLMs: Capability Density and Training Quality
By
–
Densing Law of LLMs https://
arxiv.org/pdf/2412.04315
v2
… introducing the concept of “capability density” to evaluate the training quality of large language models (LLMs) and describe the trend of LLMs that considers both effectiveness and efficiency. -
IndiaAI Fellowship Empowers Tomorrow’s AI Leaders
By
–
Empowering tomorrow's AI leaders! BTech & MTech students are already transforming their AI ideas into reality with the support of IndiaAI Fellowship. Ready for the next big step? STAY TUNED! #IndiaAIFellowship #AIResearch #IndiaAI #TechForIndia #AIInnovation
-
The Beginning of Infinity: A Foundational Work Beyond Books
By
–
“The Beginning of Infinity” is a degree more than a book.
-
Google Quantum Computing Breakthrough Moves Closer to Reality
By
–
Quantum Computing Inches Closer to Reality After Another Google Breakthrough
-
Comparing π_0 Robot Training Data to Qwen LLM Training Scale
By
–
Thanks Kevin! Using your conversion rate, that works out to π_0 trained on 1 person-year of robot data vs. Qwen trained on 120,000 person-years of LLM data:
-
Google Research at NeurIPS 2024: ML Booth and Recruiter Meetings
By
–
Attending #NeurIPS2024? Drop by the Google Research booth (#133) starting tomorrow at 12PM, where you can chat with researchers, visit with recruiters, and learn all about the cutting-edge ML work being done at Google. Check out the booth activities at https://
goo.gle/neurips24 -
Generative Video WorldSim: Sora, Genie, VideoPoet Deep Dive
By
–
Generative WorldSim, Diffusion, Vision, Reinforcement Learning and Robotics Our longest episode ever! https://
latent.space/p/icml-2024-vi
deo-robots
… a deep dive into
– @OpenAI Sora (with @billpeeb
), – DeepMind Genie with @jparkerholder and @ashrewards – VideoPoet with @hyperparticle – -
Neural Attention Memory Models Boost Transformer Performance
By
–
Introducing An Evolved Universal Transformer Memory https://
sakana.ai/namm Neural Attention Memory Models (NAMMs) are a new kind of neural memory system for Transformers that not only boost their performance and efficiency but are also transferable to other foundation models,