A new video in Karpathy "Intro to LLMs for general audience" series dropped. This is a world gift. Thank you for all high quality videos, Andrej!!!
RESEARCH
-
Understanding Generalist Definition in Robotics Discussion
By
–
BTW, I appreciate this exchange between @LerrelPinto and @KarlPertsch : what is meant by “generalist” is very nuanced and important for robotics.
-
Non-invasive glucose tracking via Raman spectroscopy breakthrough
By
–
Taking non-invasive glucose tracking to the next level, using Raman spectroscopy, without any needles
-

Deep Dive into LLMs: Complete Training Stack Overview
By
–
New 3h31m video on YouTube:
"Deep Dive into LLMs like ChatGPT" This is a general audience deep dive into the Large Language Model (LLM) AI technology that powers ChatGPT and related products. It is covers the full training stack of how the models are developed, along with mental -
Gemini 2.0 Series Advances Cost, Performance, and Agentic Capabilities
By
–
Exciting progress as we relentlessly drive quality up & cost down. The Gemini 2.0 series are the leading models for cost & performance. w/today’s releases, everyone can tap into their powerful reasoning/multimodal capabilities, which also lay the foundations for our agentic work https://t.co/eRSoedLuTU
— Demis Hassabis (@demishassabis) 5 février 2025Exciting progress as we relentlessly drive quality up & cost down. The Gemini 2.0 series are the leading models for cost & performance. w/today’s releases, everyone can tap into their powerful reasoning/multimodal capabilities, which also lay the foundations for our agentic work
-

Dr. Fei-Fei Li Wins QE Prize for ImageNet Contributions
By
–
Congratulations to our Co-Director @drfeifei for winning this year's @QEPrize
! Her seminal work on ImageNet has laid the foundations for the technological innovation that's shaping our world today. Another well-deserved honor! #QEPrize2025 https://
bit.ly/4gwqSC7 -
DeepSeek r1 Chain-of-Thought Traces Now Visible Inline
By
–
See what your AI models are thinking as they go: CoT traces are now visible inline for all reasoning models.
— Aaron Ng (@localghost) 5 février 2025
Here’s DeepSeek r1 thinking through a problem. pic.twitter.com/u282dULJKtSee what your AI models are thinking as they go: CoT traces are now visible inline for all reasoning models. Here’s DeepSeek r1 thinking through a problem.
-
Meaningful scientific discussions with Claude
By
–
At the same time I am having really meaningful scientific discussions with Claude.
-

Stanford HAI Scholars Guide Policymakers on AI Policy Globally
By
–
From the U.S. to Southeast Asia, @StanfordHAI scholars are helping policymakers navigate AI’s impact with evidence-based insights and tailored training programs. Learn more about our AI policy initiatives: https://
stanford.io/4hNOwLr -
How Transformer LLMs Work: Deep Dive Course
By
–
Announcing How Transformer LLMs Work, created with @JayAlammar and @MaartenGr, co-authors of the beautifully illustrated book, “Hands-On Large Language Models.”
— Andrew Ng (@AndrewYNg) 5 février 2025
This course offers a deep dive into the inner workings of the transformer architecture that powers large language… pic.twitter.com/zUv9dnVvSSAnnouncing How Transformer LLMs Work, created with @JayAlammar and @MaartenGr
, co-authors of the beautifully illustrated book, “Hands-On Large Language Models.” This course offers a deep dive into the inner workings of the transformer architecture that powers large language
