Finances
Replaying one retirement plan across every start year since 1928: the simulator, the agent pipeline behind it, and 98 lives ending broke or on $5.8 million.
4 parts
Part 1
Teaching an AI to run retirement experiments
retire-sim replays every historical start year since 1928, grades strategies on earliest safe retirement age, and lets Claude Code run the experiments.
11 min read
Part 2
My simulator watches retirement YouTube for me
A text-file queue, a transcript script and two Claude Code skills turn retirement videos into three experiment rows, a verdict and a channel scoreboard.
10 min read
Part 3
Same plan, same spending. Broke, or $5.8 million.
One textbook retirement plan replayed across 98 historical start years, from 1928 on. Four of those lives, and what sequence-of-returns risk really means.
9 min read
Part 4
Real taxes, less hearsay
My simulator's flat 5% tax knob was hearsay. A real tax engine measured 22%, the knob went to 15%, and asset location beat every Roth conversion video.
13 min read



