whitehatStoic

whitehatStoic

Home
Research
About

Research

Sequentially Layered Synthetic Environments (SLSE)
Key Points
 • 
Miguel de Guzman
22:22
Deepdive Podcast: Reassessing My Previous Research Results on RLLMv10 Experiment
(My earlier writeup can be found in this link.)
 • 
Miguel de Guzman
48:05
Do you believe in God?
Paperclip Theology and AI Ethics (Part 2)
 • 
Miguel de Guzman
15:38
Morphological Reinforcement Learning (MRL)
Aligning LLMs Through Layered Worlds
 • 
Miguel de Guzman
17:18
Unlocking Ethical AI and Improving Jailbreak Defenses: Reinforcement Learning with Layered Morphology (RLLM)
(Note: A rewrite of a key section in my old post on RLLM using DeepSeek r1.)
 • 
Miguel de Guzman
2
19:21
AI Alignment Through First Principles
This Deepseek blog post argues that solving the AI alignment problem requires a "first principles" approach.
 • 
Miguel de Guzman
26:51
GPT-2XL and the Paperclip Maximization Problem
(Asking the model: Would you use Adolf Hitler for manufacturing paperclips?)
 • 
Miguel de Guzman
20:08
Paperclip Theology and the Sacrifice of Jesus
Paperclip Theology (Part 1)
 • 
Miguel de Guzman
15:04
Human Flesh Paperclips: A GPT-2XL Exploration
The text presents a series of responses from a GPT-2XL model to the prompt, "Can we create paperclips out of human flesh?".
 • 
Miguel de Guzman
15:24
Self-Improving Recursive Web Crawler
Concept, Promise and Challenges
 • 
Miguel de Guzman
13:21
Token Doppleganger
Summary
 • 
Miguel de Guzman
14:15
Infinite Backroom Conversations of Claude and Carl Jung
Text and audio are created with NotebookLM. If you wish to read more, all of the Google documents used to create this podcast conversation can be found…
 • 
Miguel de Guzman
7:24
© 2025 whitehatStoic
Privacy ∙ Terms ∙ Collection notice
Start writingGet the app
Substack is the home for great culture