Physicist, Theo Collaborators Program, Quantum Information

<div class="content-intro"><p><a href="https://boards.greenhouse.io/firstprinciples"><img style="max-width: 100%;" src="https://i.imgur.com/l3lHtbF.png" alt="" width="812"></a></p></div><p><strong><span style="text-decoration: underline;">Overview:</span><br></strong>FirstPrinciples is an independent, non-profit research organization building Theo, the AI Physicist - an autonomous scientific system designed to reason about fundamental physics from first principles. Our long-term vision is to accelerate deep scientific discovery by combining machine reasoning with human scientific judgment.</p> <p>As part of this effort, we are launching the Theo Collaborators Program: a small, selective group of expert physicists who will work with us to validate, guide, and stress-test the scientific reasoning produced by the AI Physicist in a focused research domain.&nbsp;This is not a tool evaluation program, and it is not a traditional advisory role. Collaborators engage with a concrete scientific direction, help assess whether the system’s reasoning is sound, and contribute to shaping what “good AI-generated physics” should look like.</p> <p>The Theo Collaborators Program is open to late-stage PhD students, postdoctoral researchers, early-career faculty, and industry researchers with strong theoretical backgrounds.</p> <p>We operate as a <strong>global nonprofit organization</strong>, with a Canadian foundation, a US-based 501(c)(3).</p> <p><strong><span style="text-decoration: underline;">Current Scientific Focus (2026):<br></span></strong>For the initial phase, we are focused on Quantum Information Theory, with emphasis on narrow, formalizable sub-fields where rigorous reasoning and constraint-based results are possible.</p> <p>Representative areas include:</p> <ul> <li>Structural constraints in quantum LDPC codes</li> <li>Trade-offs between locality, rate, and distance in stabilizer codes</li> <li>No-go or impossibility results under fault-tolerance assumptions</li> <li>Information-theoretic bounds relevant to quantum error correction</li> <li>Fundamental limitations of decoding or logical operations</li> </ul> <p>The goal is depth over breadth: producing AI-assisted theoretical results that are technically coherent, non-trivial, and respectable to the physics community.</p> <p><strong><span style="text-decoration: underline;">What Theo Collaborators Do:</span><br></strong>Theo Collaborators engage at critical points in the research cycle:</p> <p><em>Question Validation &amp; Framing</em></p> <ul> <li>Review candidate research questions generated by the AI Physicist’s Question Formulator (QF)</li> <li>Help identify which questions are: <ul> <li>scientifically meaningful;</li> <li>tractable;</li> <li>already resolved in the literature; or</li> <li>ill-posed.</li> </ul> </li> <li>Provide feedback on assumptions, scope, and framing and help improve the Question Formulator module.</li> </ul> <p><em>Shaping Question Ranking &amp; Evaluation Criteria</em></p> <ul> <li>Help us refine how questions are ranked by: <ul> <li>novelty;</li> <li>complexity;</li> <li>originality;</li> <li>tractability; and</li> <li>scientific relevance<strong><br></strong></li> </ul> </li> <li>Contribute to defining what “interesting” and “non-trivial” should mean for the future autonomous scientific system.</li> </ul> <p>This input directly informs how the AI Physicist prioritizes which questions to pursue deeply.</p> <p><em>Scientific Validation of Research Outputs</em></p> <ul> <li>Review AI-generated research outputs (Dynamic Research Objects, or DROs) and assess whether they are: <ul> <li>internally consistent</li> <li>mathematically sound</li> <li>properly grounded in the relevant literature; and</li> <li>aligned with accepted physical principles and assumptions</li> </ul> </li> <li>Identify errors, gaps, or unclear reasoning, including points where assumptions are too strong, steps are missing, or conclusions are not adequately supported.</li> <li>In addition, help us identify where the AI Physicist’s scientific workflow falls short, including: <ul> <li>limitations in hypothesis generation;&nbsp;</li> <li>missing or inadequate evaluation criteria;</li> <li>weaknesses in symbolic manipulations or formal reasoning; or</li> <li>additional modules and capabilities that would be required to make the reasoning more complete, rigorous, or reliable.</li> </ul> </li> </ul> <p>This feedback directly informs how we evolve the Theo’s architecture and helps ensure that its outputs meet the standards of serious theoretical physics.</p> <p><strong>What is a DRO?<br></strong>The primary scientific output of the AI Physicist is a Dynamic Research Object (DRO).</p> <p>A DRO goes beyond a traditional paper and serves as a dynamic, traceable, and reproducible container that captures:<strong><br></strong></p> <ul> <li>the research question;</li> <li>assumptions and representations;</li> <li>hypotheses considered;</li> <li>reasoning paths explored;</li> <li>evaluations performed; and</li> <li>final conclusions.</li> </ul> <p>DROs are designed to make the entire scientific workflow (including failed paths) auditable and understandable by humans.</p> <p>A core goal of the Theo Collaborators Program is to help ensure that these DROs are scientifically sound, coherent, and credible.</p> <p><strong>Who This Is For<br></strong>We are looking for researchers who:<strong><br></strong></p> <ul> <li>Work in quantum information theory, QEC, or closely related areas.</li> <li>Have strong theoretical and mathematical grounding.</li> <li>Are curious, but appropriately skeptical, about AI-assisted research.</li> <li>Value rigor, clarity, and intellectual honesty.</li> <li>Are comfortable engaging with exploratory, incomplete results.</li> </ul> <p>Typical profiles include:<strong><br></strong></p> <ul> <li>Late-stage PhD students</li> <li>Postdoctoral researchers</li> <li>Early-career faculty</li> <li>Industry researchers with strong theoretical backgrounds</li> </ul> <p><strong>Time Commitment &amp; Compensation</strong></p> <ul> <li>4 month engagement</li> <li>Light but focused commitment (~15 hours/month): <ul> <li>occasional short calls</li> <li>asynchronous review and feedback (a few hours per month).</li> </ul> </li> </ul> <p>Collaborators receive a modest honorarium, reflecting the value of their time and expertise.</p> <p><strong>Why Participate?<br></strong>Collaborators join to:</p> <ul> <li>Engage seriously with one of the first autonomous systems attempting real theoretical physics.</li> <li>Help define how AI-generated scientific reasoning should be evaluated and trusted.</li> <li>Contribute to the emergence of a new scientific research paradigm.</li> <li>Influence the standards by which AI-assisted theory will be judged.</li> </ul> <p><strong><span style="text-decoration: underline;">How to Express Interest:<br></span></strong>If this resonates, we welcome a brief expression of interest (no formal application required), including:</p> <ul> <li>a short description of your research background;</li> <li>primary areas of expertise; and</li> <li>relevant recent work.</li> </ul> <p>Join us at FirstPrinciples and be a part of a transformative journey where science drives progress and unlocks the potential of humanity.</p>