Example · PhD, Computer Science and AI
58% of the work a CS and AI PhD student is heading toward still needs a person.
AI now writes baselines and training code. Judging whether a result holds up stays with you.
Make my planA typical week at the middle pace. Your plan uses your own tasks.
Is a PhD in AI still worth it?
In a typical week at the middle pace, 25% of this work needs your judgment and 33% is work you do using AI. AI could do about 42%. Coding agents now do a large share of the work that fills a second- or third-year PhD week: the related-work sweep, the baseline reimplementation, the training script, the sweep config, the grading. That is uncomfortable, because those are the tasks that taught the last generation of students to be researchers, and research-engineer postings are written around them.
Read more
What sets a PhD apart at the end is the part that comes next.
What holds is the part your advisor means when they ask what the numbers mean. You pick a question worth a year. You design the experiment that would change your mind. You know a run is wrong before the plot says so, and you put your name on the claim. The next 13 weeks turn one of the hardest parts of the PhD, the baseline that never matches the paper, into a public record that you can tell a sound result from a broken one. Agents make that skill more valuable, and few labs have someone who owns it yet.
The week, task by task
Who does each hour of a typical 40-hour week.
See every task
Needs your judgment.
- Choosing the question and designing the experiment that could falsify it9%
- Presenting: talks, posters, reading group, conference conversations6%
- Leading sections and in-person office hours4%
- Mentoring undergraduates and newer students4%
- + 1 smaller task2%
AI does the slow parts. You still own it.
- Debugging runs that fail, diverge or don't match the paper9%
- Writing papers, rebuttals and camera-ready versions9%
- Analysis: plots, ablation tables, statistical checks7%
- Peer reviewing for conferences and workshops4%
- + 1 smaller task4%
AI could do most of it.
- Experiment code: training and eval pipelines, configs, sweeps14%
- Reimplementing and replicating baselines from papers12%
- Literature search, reading summaries, related-work drafts8%
- TA grading, auto-graders, routine course forum questions8%
Why some tasks still need a person
Choosing which question is worth a year of your life
It depends on what your lab can run, what reviewers in your subfield will believe, what the field is about to make obsolete, and what you can still care about in month nine. None of that is written down.
Designing the experiment that would prove you wrong
An agent runs the experiment you ask for. Asking for the one that could falsify your idea is the research skill, and a wrong design costs months before anyone notices.
Knowing a run is broken before the curve tells you
Data leakage, a seed that flatters the method and an eval set that overlaps pretraining data are all easy mistakes to make and costly to find.
Your name on the paper
Venue policies make authors responsible for everything in a submission, including what a tool wrote. That responsibility does not move to the tool.
Mentoring an undergrad through a first project
Undergrads come to you because you sat in their seat two years ago. That trust is how labs keep running, and it is part of how search committees judge you.
Good to know
What is hard about this field, and what you can do about it.
The hard truth
Frontier AI models are now mostly built in industry.
If you want to train the largest models, most of that work now happens at companies with huge compute budgets. Academic labs still lead on the ideas the field cites most.
Sources (2)
- Nearly 90% of notable AI models in 2024 came from industry, up from 60% in 2023. Stanford HAI, AI Index Report 2025, 2025-04
- Academia remains the top source of highly cited AI research. Stanford HAI, AI Index Report 2025, 2025-04
What careers come after a CS PhD?
Plan A is the best move in the field you are studying for.
Focus your studies
The main focus. Most weekly sessions build toward it.
- Each week
- 4 sessions
- Ready in
- 13 weeks to a submitted reproducibility study
Why it fits
Why you. If you have fought with baselines that won't match, you already know more about broken replications than people who skipped them.
The gap. You do not yet have public reproducibility work, a systematic record of agent failures or an evaluation paper.
The trade-off. Reproducibility papers count for less than new methods on many faculty committees, and they can annoy the original authors. Do it carefully, contact the authors, and keep it alongside your thesis work rather than in place of it.
2 other moves that fit
- Pick a thesis question that doesn't need huge compute budgetsReady in one semester to a pilot
- Apply beyond the few big AI labsReady in two summers
What's a plan B for AI PhD students?
Plan B is meant to be different from this job. It was picked because skills from this role carry over.
Get ready for a new kind of job
A second option. A newer job AI is creating that these skills already fit.
Designing the experiment that would prove you wrong is what model evaluation teams hire PhDs to do every week.
- Each week
- 1 session
- Outlook
- Hiring now
- Pay
- $150.4K to $385K a yearTwo postings: Apple (Oct 2026, base pay) and Perplexity (now closed).
2 other options
- Domain AI red teamerKnowing a run is broken before the curve shows it helps you find the failure modes a model's builders missed.
- AI policy analystFellowships want people who can explain technical limits clearly, and defending your work when someone pushes back in a talk is practice.
In your own plan you can switch to any of the 47 roles.
The 13 weeks
Small steps that add up. One plan for both: most sessions build Plan A, one a week builds Plan B.
By week 13 you will have
AFor Plan A
- A reproducibility study submitted to TMLR with a public repo
- A written audit of where a coding agent failed
- Three internship applications sent
BAnd for Plan B
- A regression table with a significance check
- A published eval write-up for your portfolio
- 1See where your time goes
- 2Pick a paper to replicate
- 3Let the agent try first
- 4Run your own replication
You finish withA table comparing you, the agent and the paper
- 5Explain the gap
- 6Thanksgiving: a lighter week
- 7Meet with your advisor
- 8Finals: time a routine task
You finish withTwo timings, with and without AI tools
- 9Finals: draft the methods
- 10Write the paper
- 11Release the code
- 12Get feedback
- 13Submit and apply
You finish withThree internship applications with your repo linked
Three courses from the plan
- NAIRR Pilot Start-Up projectNSF NAIRR Pilot · No fee listed
Yes, Week 2.
- TPU Research CloudGoogle Research · Free to accepted researchers
Yes, if your code runs on JAX or PyTorch/XLA.
- The Missing Semester (2026), Agentic Coding lectureMIT CSAIL · Free online; no fee listed
Yes, Week 1.
Checked 2026-10-13.
The fuller picture
Risks beyond AI: highFunding and visa-policy risk
Federal science money and immigration rules moved more in 2025 and 2026 than in the decade before. Roughly half of CS doctorates go to nonresidents, so visa rules shape this labor market for everyone in it, citizen or not. The faculty market is tight on its own, and most new assistant professors arrive through a postdoc or another school.
Where this could be wrong
- The task blend is ours. Theory students spend far less time on experiment code; systems and robotics students spend more on hardware. Your own Week 1 log corrects it.
- The 93.0% PaperBench figure is a provider self-report in a lighter setting. If independent runs land far lower, we have overstated how exposed replication work is.
- Taulbee leaves 27.6% of 2025 graduates' outcomes unknown. Unemployment among them could be higher than the 0.7% among known outcomes.
- METR's time-horizon page is no longer actively updated, and its top estimate has an interval from about 8.5 to 55 hours.
Also matched: phd student computer science, cs phd, computer science phd. Research checked 2026-10-13. Page updated .