Francis Clase
AI Researcher

3 years of research in superintelligence, AI safety, and advanced AI systems. Exploring the theoretical foundations and practical implications of artificial general intelligence.

Research Overview

I focus on understanding the dynamics of advanced AI systems, from theoretical models of superintelligence to practical questions of AI safety and alignment. My work bridges foundational theory with emerging empirical evidence from modern AI development.

3+
Years of Research
Multiple
Research Areas
Independent
Researcher

Publications

Research contributions to AI safety and superintelligence theory.

The Intelligence Explosion: From Singular Event to Complex System Dynamics

Francis Clase

Independent Research • 2025

A comprehensive analysis synthesizing 60 years of intelligence explosion theory, from I.J. Good's foundational 1965 hypothesis through modern critiques and alternative models.

Featured

Unexplored Frontiers in AI Superintelligence Research

Francis Clase

Research Survey • 2025

Systematic identification of theoretical gaps and practical opportunities in superintelligence research. Covers documentation gaps, interdisciplinary connections, and accessible research projects.

AI Safety Frameworks for Advanced Systems

Francis Clase

In Progress • 2025

Developing practical safety evaluation frameworks for AI systems approaching human-level capabilities across multiple domains.

Draft

Empirical Analysis of Large Language Model Capabilities

Francis Clase

Technical Report • 2024

Systematic evaluation of reasoning capabilities and emergent behaviors in current large language models.

Value Alignment in Multi-Agent AI Systems

Francis Clase

Working Paper • 2024

Exploring coordination mechanisms and value preservation in systems with multiple interacting AI agents.

Latest Research Highlights

Recent developments and ongoing work in AI safety and superintelligence theory.

FEATURED RESEARCH

Intelligence Explosion Dynamics

Comprehensive analysis of 60 years of theory from Good to modern critiques.

IN PROGRESS

AI Safety Frameworks

Developing practical evaluation frameworks for advanced AI systems.

ANALYSIS

LLM Capabilities Study

Systematic evaluation of reasoning and emergent behaviors.

Research Specializations

Core areas of focus in artificial intelligence research.

Superintelligence Theory

Analyzing intelligence explosion dynamics, recursive self-improvement, and takeoff scenarios.

AI Safety & Alignment

Exploring control problems, value alignment, and safety measures for advanced AI systems.

Emerging AI Systems

Studying current AI developments and their implications for future superintelligent systems.

Explore Research Topics

From ensuring AI systems align with human values to optimizing performance at scale, modern AI research spans multiple interconnected disciplines. Below are simplified introductions to key research areas.

AI Safety & Alignment

How do we ensure AI systems do what we want them to do? This field focuses on making AI systems that understand human values, follow instructions reliably, and remain safe as they become more capable.

Key concepts: Reward modeling, value alignment, honesty

Model Training

Training AI models requires massive amounts of data and computation. Researchers study how models learn, how to make training more efficient, and how model capabilities scale with increased resources.

Key concepts: Pre-training, scaling laws, reinforcement learning

Performance Engineering

Making AI systems run faster and more efficiently. This involves optimizing hardware usage, reducing memory requirements, and ensuring models can handle real-world workloads at scale.

Key concepts: GPU optimization, inference speed, resource efficiency

Interpretability

Understanding how AI systems make decisions. When a model produces an output, can we explain why? This research helps us peek inside the “black box” of neural networks.

Key concepts: Model transparency, feature analysis, decision explanations

ML Infrastructure

Building the systems that power AI research. This includes specialized hardware, distributed training systems, and the software infrastructure needed for large-scale machine learning.

Key concepts: TPUs, distributed systems, ML accelerators

Multimodal & Specialized AI

AI that can understand and work with multiple types of information - text, images, audio, and more. Also includes specialized applications like biology research and scientific discovery.

Key concepts: Cross-modal learning, specialized domains, scientific AI

Browse 25+ topics across introductory, intermediate, and advanced levels

Research Collaboration

Interested in discussing AI safety, superintelligence theory, or potential research collaborations? Connect with Francis Clase for academic exchanges and research insights.