Check out [The Gradient](https://thegradient.pub/)! Also, [Davis Summarizes Papers](https://dblalock.substack.com/) is awesome

those considering a move to sigmoid.social, some info: * It is run/administered by The Gradient (thegradient.pub). Disclaimer - I am part of the team. * It has about 5500 active users currently, including some notable folks like Karpathy

those who are curious, check out this blogpost from me: Transformers are Graph Neural Networks - [https://thegradient.pub/transformers-are-graph-neural-networks/](https://thegradient.pub/transformers-are-graph-neural-networks/) It explores the connection between Transformer models such as GPTs and other LLMs for Natural Language Processing

reasons. They emerge new abilities at scale - [https://arxiv.org/abs/2206.07682](https://arxiv.org/abs/2206.07682) They build internal world models - [https://thegradient.pub/othello/](https://thegradient.pub/othello/) They can be grounded to robotics -( i.e act as a robots brain) - [https://say-can.github.io/](https://say-can.github.io/), https://inner-monologue.github.io

Another, older post on this: https://thegradient.pub/state-of-ml-frameworks-2019-pytorch-dominates-research-tensorflow-dominates-industry/

important tasks that are sensitive to correctness. This is the basis behind this [Othello-GPT paper](https://thegradient.pub/othello/) - take an existing GPT-3 model and finetune it on valid Othello boards