Transforming AI Safety through Theoretical Computer Science
I recently shared my insights at the UK AI Safety Institute Workshop, blending over 25 years of theoretical computer science with AI alignment research. Here’s a quick overview:
- Background: A theoretical computer scientist, I’ve spent recent years at OpenAI and UT Austin, focusing on applying CS principles to AI safety.
- Research Opportunities:
- PhD Positions: Open for talented students interested in AI alignment or quantum computing. Apply by December 15 at UTCS.
- Postdoc Fellowships: Seeking individuals at the crossroads of CS theory and AI alignment, starting Fall 2026. Reach out by January 15.
- Key Challenges & Contributions:
- Advancements in empirical AI research highlight the need for theoretical underpinnings.
- Proposals like the Gumbel Softmax Scheme demonstrate practical applications (despite its adoption hurdles).
Let’s bridge the gap between theory and practice in AI! Interested in collaborating or sharing insights? Comment below or connect directly!
