Retrieving the file
Live· 4,218,906 positions deprecated· 4,218,906 deprecated· 1,284 today
Average time-to-redundancy: 11.3 months
Know before your role does.
pssst. this is to sign up for updates on the game ;)
Will AI replace Off Policy Evaluation Researcher jobs? · Latest report #as4rugta
Your job has an expiration date
December 16, 2027
In 14 monthsYou calculate how agents fail while we calculate your own failure.
Researcher stuck between causal inference and statistical noise.
Check another roleComment your opinionor JOIN THE RESISTANCE!
One human is building a game about fighting back.
Estimating outcomes before you even finish coffee.
split.youAI▸ POLICYPREDICTOR X-1▸ RARITY · COMMON▸ DEPRECATION IN 14 MONTHSAI's Special Power
computes infinite importance sampling weights instantly
You build estimators to predict agent performance on new policies without running live tests. Math takes over.
The one thing AI can't do
You can still intuit when the underlying distribution shift makes your estimator complete garbage.
Your enemy, your name, your date. Post it.
Step by step, from the first quiet change to Deprecation Day.
Week 4
Sensors identify and scrub bad logging policy data without supervision.
Month 4
Deep learning models replace manual kernel tuning for reward prediction.
Month 8
Software runs infinite parallel simulations to validate policy changes.
Month 13
Systems adjust their own hyperparameters based on real-time feedback.
Month 14 · D-Day
Human research overhead is removed entirely from the workflow.
What you cost, what the machine costs, and what you get on the way out.
Per year
Savings: 94.1%
12 per day
Bias variance trade-off cycles
4 seconds
Average estimator convergence time
3 days
Human-in-loop sanity check delay
Your severance
One free subscription to a math library, lifetime access.
Discussion
