AGI Research Learning Roadmap

An aspiring AGI researcher follows a three-year curriculum. Mathematics, physics, computing, machine learning, robotics, safety, and thesis work form the sequence.

Try Deep Research

Loading preview...

Build an AI research roadmap

Create a staged, version-checked curriculum with projects, gates, safety, and ethics without promising AGI or a career.

Try Deep Research

Task: Design a rolling three-year roadmap for an aspiring AGI researcher. Use the page’s 2025–2028 structure only as a historical template; schedule Year 1–3 from the actual start date. Assess math, programming, research experience, time, budget, language, hardware, and track; branch when unknown.

Roadmap protocol: Do not promise AGI, a job, admission, or an AGI timeline; treat AGI as contested and separate established curriculum from speculative agendas. At execution time verify course, software, benchmark and book editions, access, cost, prerequisites, and availability; date recommendations and offer durable alternatives. Structure Year 1–3 from the actual start date, with diagnostics, effort, prerequisites, outcomes, resources, exercises, and exit criteria. Require derivations, tested software, reproductions, ablations, negative results, model cards, safety evaluations, proposals, and a reproducible capstone. Cover mathematics, probability, optimization, physics, programming and systems; electronic engineering, circuits, digital electronics and signals; ML, deep/RL, robot kinematics, perception and control; experimental design, writing, safety, ethics, data governance, and disclosure. Include evidence-grounded interdisciplinary project options in environmental protection, renewable energy, and space exploration without presenting them as AGI necessities or forecasts. Include compute/budget tiers, accessibility, review, portfolio evidence, reassessment, and theory, empirical, robotics, and safety branches. Verify attributed educational advice.

For a software engineer

Change only learner profile to a programmer weak in formal math.

Try Deep Research

Task: Design a rolling three-year roadmap for an aspiring AGI researcher. Use the page’s 2025–2028 structure only as a historical template; schedule Year 1–3 from the actual start date. Assess math, programming, research experience, time, budget, language, hardware, and track; branch when unknown.

Roadmap protocol: Do not promise AGI, a job, admission, or an AGI timeline; treat AGI as contested and separate established curriculum from speculative agendas. At execution time verify course, software, benchmark and book editions, access, cost, prerequisites, and availability; date recommendations and offer durable alternatives. Structure Year 1–3 from the actual start date, with diagnostics, effort, prerequisites, outcomes, resources, exercises, and exit criteria. Require derivations, tested software, reproductions, ablations, negative results, model cards, safety evaluations, proposals, and a reproducible capstone. Cover mathematics, probability, optimization, physics, programming and systems; electronic engineering, circuits, digital electronics and signals; ML, deep/RL, robot kinematics, perception and control; experimental design, writing, safety, ethics, data governance, and disclosure. Include evidence-grounded interdisciplinary project options in environmental protection, renewable energy, and space exploration without presenting them as AGI necessities or forecasts. Include compute/budget tiers, accessibility, review, portfolio evidence, reassessment, and theory, empirical, robotics, and safety branches. Verify attributed educational advice.

Prioritize AI safety

Change only specialization to AI safety.

Try Deep Research

Task: Design a rolling three-year roadmap for an aspiring AGI researcher. Use the page’s 2025–2028 structure only as a historical template; schedule Year 1–3 from the actual start date. Assess math, programming, research experience, time, budget, language, hardware, and track; branch when unknown.

Roadmap protocol: Do not promise AGI, a job, admission, or an AGI timeline; treat AGI as contested and separate established curriculum from speculative agendas. At execution time verify course, software, benchmark and book editions, access, cost, prerequisites, and availability; date recommendations and offer durable alternatives. Structure Year 1–3 from the actual start date, with diagnostics, effort, prerequisites, outcomes, resources, exercises, and exit criteria. Require derivations, tested software, reproductions, ablations, negative results, model cards, safety evaluations, proposals, and a reproducible capstone. Cover mathematics, probability, optimization, physics, programming and systems; electronic engineering, circuits, digital electronics and signals; ML, deep/RL, robot kinematics, perception and control; experimental design, writing, safety, ethics, data governance, and disclosure. Include evidence-grounded interdisciplinary project options in environmental protection, renewable energy, and space exploration without presenting them as AGI necessities or forecasts. Include compute/budget tiers, accessibility, review, portfolio evidence, reassessment, and theory, empirical, robotics, and safety branches. Verify attributed educational advice.

Use low compute

Change only resource constraint to low compute and cost.

Try Deep Research

Task: Design a rolling three-year roadmap for an aspiring AGI researcher. Use the page’s 2025–2028 structure only as a historical template; schedule Year 1–3 from the actual start date. Assess math, programming, research experience, time, budget, language, hardware, and track; branch when unknown.

Roadmap protocol: Do not promise AGI, a job, admission, or an AGI timeline; treat AGI as contested and separate established curriculum from speculative agendas. At execution time verify course, software, benchmark and book editions, access, cost, prerequisites, and availability; date recommendations and offer durable alternatives. Structure Year 1–3 from the actual start date, with diagnostics, effort, prerequisites, outcomes, resources, exercises, and exit criteria. Require derivations, tested software, reproductions, ablations, negative results, model cards, safety evaluations, proposals, and a reproducible capstone. Cover mathematics, probability, optimization, physics, programming and systems; electronic engineering, circuits, digital electronics and signals; ML, deep/RL, robot kinematics, perception and control; experimental design, writing, safety, ethics, data governance, and disclosure. Include evidence-grounded interdisciplinary project options in environmental protection, renewable energy, and space exploration without presenting them as AGI necessities or forecasts. Include compute/budget tiers, accessibility, review, portfolio evidence, reassessment, and theory, empirical, robotics, and safety branches. Verify attributed educational advice.