7B model, 1M context window! Lots of results over long token windows; this looks very promising. https://
largeworldmodel.github.io RingAttention is what makes this works, summary generated by GPT: Ring Attention is a new method that reduces the memory requirements of Transformers. It
@jiquanngiam
-

7B Model Achieves 1M Context Window with RingAttention
By
–
-
AI Models: Bridging User Experience at Performance Limits
By
–
We’ve found that users of AI products tend to quickly find the boundaries of where the models reach their limits. A lot of product work is needed to bridge the experience when they inevitably get there — something that can only be done in the wild, not through fixed datasets.
-
React useContext: Developer Experience Insights
By
–
lol when I’m coding with react and useContext — it often felt this way (in a good way)
-

GPT Hack: Target Reading Ease Score 75
By
–
One of my favorite GPT hacks is to ask it to target a reading ease score of 75. GPT tends to produce flowery language, and this helps make the text it produces be much more readable instead. #ChatGPT
-
Lutra AI: English-Powered Automation Tool Saves Time Revenue
By
–
Love it. Check out @Lutra_AI – create automations that save you time (and possibly bring in more revenue), just from English — the AI creates it for you.
-
RAG on User Account Data for Personalized AI Systems
By
–
Slick! Do they also RAG on user specific account data to many it even more personalized?
-
Immersive Tech: Mental Distance Despite Physical Proximity
By
–
Part of me gets excited for the tech. Part of me feels concerned that we’ll be even more glued to screens / immersive devices; will we be even further apart mentally despite being physically there.
-
Lutra AI launches workflow automation tools with early access
By
–
This is awesome! We’re building tools at @Lutra_AI to make such workflows super easy to build. Happy to get you early access — DM me!