Hugging Face is always on top of it
@hardmaru
-
Rethinking Neural Network Abstractions Beyond Traditional Models
By
–
There has been previous attempts like OpenWorm from many years ago, but progress hasn’t been spectacular. Maybe we need to rethink the way we abstract the concept of “neurons” and “weights” and their interactions in our own artificial neural network models. https://t.co/H74qcjgP3i
— hardmaru (@hardmaru) 2 juillet 2023There has been previous attempts like OpenWorm from many years ago, but progress hasn’t been spectacular. Maybe we need to rethink the way we abstract the concept of “neurons” and “weights” and their interactions in our own artificial neural network models.
-
Why Can’t We Simulate Mapped Brain Connectomes Computationally?
By
–
There are many projects that have mapped whole-brain connectome of various creatures, from C. elegans (302 neurons) to Drosophila fruit fly (~100k neurons). With our current compute, why can’t we create accurate computer simulations of these creatures in a virtual 3D environment? https://t.co/PXCsnIaYAJ
— hardmaru (@hardmaru) 2 juillet 2023There are many projects that have mapped whole-brain connectome of various creatures, from C. elegans (302 neurons) to Drosophila fruit fly (~100k neurons). With our current compute, why can’t we create accurate computer simulations of these creatures in a virtual 3D environment?
-
Bootstrap over VC: Why Constraints Drive Innovation
By
–
This is so important. That's why I'm generally weary of creating a business that requires a ton of VC money (with a plan that involves burning it all on GPUs). Prefer instead to bootstrap and build the most with minimal resources instead. Constraints lead to innovation.
-
Podcast Discussion on LLMs: Open Source, Legal, and Existential AI Risks
By
–
I’m usually not a listener of hours-long podcasts, but this one with @pmarca presents a great discussion on the future of LLMs and open-source, legal, commercial, ethical, existential “risks”, and geopolitical issues surrounding AI.
-
Commitment to Continue Working with Hugging Face
By
–
I'll definitely continue to work with @huggingface whatever I end up doing next 🙂
-
Impact of Machine Learning on the US Economy in the Decade
By
–
True. Though I don't understand why he won't acknowledge the impact machine learning will have on the US economy in the next decade. Even if one doesn't believe in the "AGI hype", the impact of ML to the economy in the next 10 years is a trend that I wouldn't want to bet against.
-
AI Companies Ban Model Outputs for Training While Using Online Content
By
–
OpenAI, Google & Anthropic ban the use of the generated output content from their AI models to train other AI models, under their terms-of-service. However, they’ve been using other online content for their own model training. They can’t have it both ways.
-
SlimPajama-627B: Large Deduplicated Open-Source LLM Dataset
By
–
SlimPajama-627B: the largest extensively deduplicated, multi1corpora, open-source dataset for training large language models. Sometimes less is more! https://
reddit.com/r/MachineLearn
ing/comments/1467jvm/np_introducing_slimpajama627b_the_largest/
…