Career Plans

TL;DR

As a third-year student at MIT, I plan to accelerate completion of my bachelor’s degree in AI and Decision Making by December 2025—graduating a semester early to immediately begin work in AI safety or policy. My timeline predictions have shortened, and I’m now focused on mitigating AI existential risk within a 5-year AGI horizon. I’m leaning toward work that bridges technical capability and policy influence, particularly through developing evaluations (similar to METR) or creating demonstrations that help policymakers understand existential AI risks (in the vein of CivAI). While I have strong interests in writing, strategy, and communications, I’m also considering technical roles and organizations like CAIP that operate at this intersection.

---

Career Plans

01 Timelines

My AGI timeline estimates have recently shortened considerably. I now see a significant probability of AGI development within 5 years, driven by:

• Rapid scaling of large language models
• Ensemble architectures (like Devin)
• Breakthroughs in prompting and in-context learning techniques
• AI systems capable of AI R&D

I believe ASI would likely follow shortly after AGI, creating an urgent need for safety measures and policy interventions before these capabilities materialize. After participating in AI forecasting competitions, I’ve continuously updated toward shorter timelines based on recent progress.

02 Relevance of Timelines to Plans

These accelerated timelines are directly shaping my educational and career decisions, including my choice to graduate early. They emphasize the need for immediate, practical interventions in AI governance and safety. With such compressed timelines, I’m prioritizing work that can have near-term impact on how AI systems are developed, evaluated, and regulated.

03 Career Interests

My current assessment of high-impact areas includes:

• [30%] Technical work influencing policy (developing capability/alignment evaluations, creating demonstrations for policymakers)
• [25%] Think tank research (organizations like CAIP that connect technical insights with policy)
• [20%] Strategic communications (helping bridge technical concepts with policy needs)
• [15%] Control mechanisms (technical approaches to AI governance)
• [10%] “AI for AI Safety” (using AI systems to address AI safety challenges)

I’ve become more bearish on interpretability research given shortened timelines, believing control mechanisms and evaluation frameworks offer more tractable paths forward.

Skills assessment: While I enjoy writing, strategy, and communications, I’m still calibrating my comparative advantage between policy and technical roles. My background is stronger in applied work than theoretical research, suggesting I might provide more immediate value in applied safety or policy roles rather than fundamental research positions.

04 Undergraduate Plans

I’m accelerating my CS degree with an AI and Decision Making concentration to graduate by December 2025. My approach includes:

• Prioritizing coursework that builds directly relevant skills for AI safety and policy
• Strategically satisfying degree requirements while minimizing time spent on less relevant areas
• Building a strong network within both technical and policy communities at MIT
• Pursuing research and projects that help me assess my fit for different post-graduation paths
• Continuing involvement with MIT AI Alignment to expand professional connections

05 Post-Graduation Plans

Given my shortened timelines, I’m focusing on opportunities that address near-term AI risks:

• Think tanks: Organizations like CAIP that connect technical work with policy impact
• Evaluation-focused organizations: Developing robust testing frameworks (similar to METR)
• Public education and demonstration projects: Helping policymakers understand AI risks (similar to CivAI)

Rather than considering graduate school as I previously had, I now plan to transition directly into roles where I can contribute to mitigating x-risk within compressed AGI timelines. I believe the urgency of the situation calls for immediate involvement in the field rather than additional academic preparation.