AI can easily recognize objects, but can it understand how it functions? Stanford researchers developed a new computer vision model that recognizes the real-world utility of objects in images:
MACHINE LEARNING
-
Programmers Must Learn AI Tools or Face Career Risks
By
–
The real danger is for programmers who say "but I don't want to be a manager, so I'm not going to figure out how to use this stuff on principle!"
-
AI Computer Vision Applications for Wildlife Conservation Efforts
By
–
Ask us your questions about how AI & computer vision can help w/conservation efforts. We’re featuring a few in an upcoming explainer w/MIT PhD student Justin Kay. For more on his work: https://
justinkay.github.io -
Teaching AI Models New Skills: The R-Counting Example
By
–
Last night I taught nanochat d32 how to count 'r' in strawberry (or similar variations). I thought this would be a good/fun example of how to add capabilities to nanochat and I wrote up a full guide here: https://
github.com/karpathy/nanoc
hat/discussions/164
… This is done via a new synthetic task -
Recoded ChatGPT from Scratch: Understanding Generative AI and Attention
By
–
If you're looking for something to watch this weekend. I recoded ChatGPT from scratch on YouTube. It's super simplified. The goal is really for every person who watches the video to understand how Generative AI actually works, and especially the attention mechanisms. Go
-
AI Facial Verification Platform National Deployment Call
By
–
One day to go! Does your AI solution perform automated duplicate detection and one-to-many identity verification using advanced facial verification technologies? This is your chance to move from pilot to national platform. Submit your application by October 25th, 2025 !
-
Quality over quantity: improving training datasets for better AI models
By
–
Great idea for a metric to further improve what datasets the models train on. It likely leads to an answer that is not web-scale crawling… Less data is often better, better data takes less.
-
Nostalgia for Activation Function Variations in Neural Networks
By
–
Brings back old days of ReLU, MELU, SELU, ELU, GELU, LeakyReLU, PELU, thousands more variations.
-
Determinism and Predictability in Complex Material Patterns
By
–
The universe is almost fully deterministic, because the more material a pattern gets, the more it has to be predictable across its futures.
-
Visual System as Generative Model: Sensory Constraints and Hallucinations
By
–
The visual system is a generative model that gets punished when it contradicts sensory patterns. In the absence of sensory patterns, it generates freely: during dreams and out of body experiences, it invents hallucinations.