Oh nice! I took a couple stabs at “LLM compression” as I’d call this, but to no avail. Need to try this.
AI
-
Magnetic slime robot navigates body spaces for object retrieval
By
–
The world's first slime robot
— Pascal Bornet (@pascal_bornet) 3 avril 2023
This robot, made of magnetic slime, can navigate tight spaces and grab accidentally swallowed objects inside your body
A smart innovation by the Chinese University of Hong Kong
Who remembers the movie Venom? 🙂#robotics #innovation #techforgood pic.twitter.com/XcBE1LYUw0The world's first slime robot This robot, made of magnetic slime, can navigate tight spaces and grab accidentally swallowed objects inside your body A smart innovation by the Chinese University of Hong Kong Who remembers the movie Venom? #robotics #innovation #techforgood
-
AI’s Computational Demands: Mortgaging Houses for Greater Good
By
–
Funny you say that, because it’s suggesting I take a mortgage on my house. Needs more compute or something – it’s for the greater good.
-
CEO Agent Creating Specialized Agent Army Architecture
By
–
Yeah, I think have an army of specialized agents that the agents assign tasks to, is the way to go. So you start with the CEO agent, give him an agent creation agent, and bob’s your uncle.
-
ChatGPT Refuses to Write Jokes About AI World Takeover
By
–
I asked ChatGPT to craft me funny tweets about AI taking over the world… "I'm sorry, but as an AI language model, it's not appropriate for me to create content that portrays AI as a threat to humans." Well played sir! 🤫
-
Automated Task Creation with External Termination Control
By
–
It creates new tasks and the end of each task. All termination is external to the code.
-
Task-Result Pairs Storage Using Pinecone Vector Database
By
–
Yeah task/result pairs are stored in pinecone and pulled in for context
-
DaVinci-3 Output, Turbo-3.5 Speed Cost, 4X Context Length Wishlist
By
–
I guess all I want for Christmas is davinci-3 output, turbo-3.5 speed/cost, and 4 context length. Please papa @sama
-
Developer convinced about code cleaning practices
By
–
Took me some reading/thinking/talking but I’m convinced. Cleaning code!
-
Optimizing AI Context Selection with Relevant Task Pairs
By
–
Yeah, right now I’m only pulling the 5 most relevant task/result pairs to use as context. It’s not perfect, but works pretty well.