LLMs, in its current form, has truly become a commodity.
@hardmaru
-
Llama 3.1 405B Launch: Open Weights, GPT-4o Comparable Performance
By
–
Llama 3.1 405B launches • Comparable to GPT-4o and Claude 3.5 Sonnet, according to the benchmarks
• The weights are publicly available
• 128k context even on the smaller models Discussion on r/ML -
Sakana AI Develops Foundation Models with High-Resolution Ukiyo-e Artwork
By
–
In collaboration with Ritsumeikan University Art Research Center (@RitsARC) and ROIS-DS Center for Open Data in the Humanities (@rois_codh), Sakana AI is proud to develop foundation models using 🇯🇵 national datasets containing the highest quality, high-resolution Ukiyo-e artwork. https://t.co/fyD8ay1bWb
— hardmaru (@hardmaru) 22 juillet 2024In collaboration with Ritsumeikan University Art Research Center (
@RitsARC
) and ROIS-DS Center for Open Data in the Humanities (
@rois_codh
), Sakana AI is proud to develop foundation models using national datasets containing the highest quality, high-resolution Ukiyo-e artwork. -
LLMs Face Hard Problems of Messy Reality Like Robotics
By
–
Enjoyed @AlexIrpan
’s “The Tragedies of Reality Are Coming for You” The hard problems of robotics are not unique to robotics. In fact, as LLMs start to actually work and become widespread, they encounter more problems from the “messy reality” of the world. -
MIT Press Releases Textbook on Artificial General Intelligence
By
–
Finally there is a textbook that explains AGI https://
mitpress.mit.edu/9780262549349/
artificial-general-intelligence/
… -
Neural Developmental Programs Research Advances from Grow-AI Team
By
–
Builds on the earlier line of work on Neural Developmental Programs (NDPs) from the Grow-AI team:https://t.co/smyJf6oBTk
— hardmaru (@hardmaru) 5 juillet 2024Builds on the earlier line of work on Neural Developmental Programs (NDPs) from the Grow-AI team:
-
Self-Assembling Neural Networks Growing Through Learning Experiences
By
–
Biological neural networks grow and develop over a lifetime of learning, and so should artificial neural networks! Here, they showed that a ‘self-assembling’ network can grow, and learn from experiences in different environments starting from randomly connected or empty networks. https://t.co/v6bcu10den
— hardmaru (@hardmaru) 5 juillet 2024Biological neural networks grow and develop over a lifetime of learning, and so should artificial neural networks! Here, they showed that a ‘self-assembling’ network can grow, and learn from experiences in different environments starting from randomly connected or empty networks.
-
AI Tour de France: A Proposed Concept
By
–
How about an AI Tour de France? 🚴♂️ pic.twitter.com/4DsnfQrEZE
— hardmaru (@hardmaru) 2 juillet 2024How about an AI Tour de France?
-
Slimebot Volleyball: 3D Multi-Agent RL Environment on SlimeVolleyGym
By
–
Slimebot Volleyball is a 3D version of Neural Slime Volleyball. The game is implemented on top of SlimeVolleyGym. The major difference is the additional z-axis, and 2-vs-2 multi-agent play, which makes training more challenging for RL algorithms like PPO.https://t.co/cyL5l8DI8M pic.twitter.com/vITaGNu0Ph
— hardmaru (@hardmaru) 24 juin 2024Slimebot Volleyball is a 3D version of Neural Slime Volleyball. The game is implemented on top of SlimeVolleyGym. The major difference is the additional z-axis, and 2-vs-2 multi-agent play, which makes training more challenging for RL algorithms like PPO. https://
github.com/jbakams/slimeb
ot-volleyball
… -
Research Paper on Psychology and Linguistics Discussion
By
–
Link to PDF: https://
gwern.net/doc/psychology
/linguistics/2024-fedorenko.pdf
… Discussion: https://
news.ycombinator.com/item?id=407561
76
…