🥽 Plastic Labs
Search
Dark mode
Light mode
Explorer
archive
ARCHIVED: YouSim Launches Identity Simulation on X
ARCHIVED: A Comprehensive Analysis of Design Patterns for REST API SDKs
ARCHIVED: Introducing Honcho's Dialectic API
ARCHIVED: Solving The Campfire Problem with Honcho
ARCHIVED: User State is State of the Art
ARCHIVED: Memories for All
ARCHIVED: Honcho: User Context Management for LLM Apps
ARCHIVED: Theory-of-Mind Is All You Need
ARCHIVED: Open-Sourcing Tutor-GPT
blog
Introducing Honcho Chat
Penny for Your Thoughts
Memory as Reasoning
Beyond the User-Assistant Paradigm: Introducing Peers
Launching Honcho: The Personal Identity Platform for AI
Agent Identity, Meta Narratives, and the End of Latent Thoughtcrimes
Xeno Grant: grants for autonomous agents
YouSim: Explore the Multiverse of Identity
careers
Working at Plastic
Summer Internships
notes
Identity is diachronic
YouSim Disclaimers
Cope is the canary, but context is key (for the end of software)
Context window size doesn't solve personalization
The model-able space of user identity is enormous
Humans like personalization
LLM Metacognition is inference about inference
On Intellectual Respect
Human-AI chat paradigm hamstrings the space of possibility
LLMs excel at theory of mind because they read
Loose theory of mind imputations are superior to verbatim response predictions
2023 recap
Honcho name lore
Machine learning is fixated on task performance
research
Introducing Neuromancer XR
SPIRAL: Letting LLMs Teach Themselves Through Self-Play
Can AI Models Predict What You'll Say Next? Developing Verifiable Social Rewards
Evaluating Steerability in Large Language Models
Achieving SOTA on OpenToM with DSPy
Violation of Expectation via Metacognitive Prompting Reduces Theory of Mind Prediction Error in Large Language Models
Tag: macro
1 item with this tag.
Jun 01, 2024
Cope is the canary, but context is key (for the end of software)