I won multiple Sci-Tech Academy Awards for visual effects. I helped build the bullet-time system for The Matrix. I spent the last five years at Epic Games working on Unreal Engine and Fortnite and brought Darth Vader to life with GenAI.

None of that mattered on March 24, 2026.

That's the day I opened an email from HR and learned I was part of a 20% work force layoff. At 53 with a theoretical plan to retire within 3-5 years, with a condo in Miami Beach, a ski in/out property in Park City, a floating home on Sauvie Island, and a net worth scattered across 27 accounts — retirement funds I can't touch until 59 and a half, real estate equity I can't spend, and a checking account that wouldn't last a year.

The first thing I did was what every engineer does. I opened a spreadsheet.

The second thing I did was sign up for every retirement planning tool I could find.

They all lied to me.

Every retirement calculator I tried — Boldin, ProjectionLab, the free ones, the paid ones — makes the same assumption: you're still saving.

Steady income. Regular contributions. Decades of compounding ahead. Plug in your age, your savings rate, your expected return, and they'll tell you when you can retire. The math is clean. The charts go up and to the right.

But I wasn't saving anymore. I was spending down. My timeline wasn't 30 years — it was the bridge between now and 59 and a half, when retirement accounts unlock without penalties. Six years. Maybe seven.

And in that gap, the math is completely different.

When you're accumulating, a bad year means your contributions buy more shares at lower prices. Dollar-cost averaging works in your favor. Time heals.

When you're drawing down, a bad year means you're selling assets at a loss to cover expenses. Sequence-of-returns risk works against you. Time kills.

No calculator I found models this. They're all built for the first half of the journey, not the crisis in the middle.

Here's what I learned staring at my own finances in panic mode:

My net worth looked healthy on paper. It's not one number - it’s separate pools, each with different rules:

Liquid — Less than 10% of my net worth was in cash and brokerage accounts. This is what I could actually spend. A naive calculation — checking account divided by monthly burn — gave me about six months. That's the number I was staring at in the dark at 3am.

Retirement — About 30% was locked in IRAs and a 401(k). Can't touch it until 59 and a half without a 10% penalty. Epic's HR struck at 53 not 55 so the "Rule of 55" didn't apply. Unless I set up a SEPP/72(t) — Substantially Equal Periodic Payments — which lets me withdraw a fixed amount annually if I commit for five years. The IRS math for this is its own rabbit hole.

Real estate — Nearly 40% was equity across multiple properties. Looks great on a net worth statement. But equity isn't cash. The properties carried mortgages, and the carry costs were more than double the rental income. That gap was bleeding my liquid pool every month.

Private — The remaining sliver was in illiquid investments. Might be worth something someday. Might be worth zero. Can't model it, can't access it.

Every dashboard I checked showed me one number. A healthy-looking number. A number that made retirement calculators say I was "on track."

But my actual survival depended on the ratio between those pools — and the rate at which the liquid pool was draining into the gap between real estate costs and rental income.

No tool showed me this. So I built one.

When I finally ran the real numbers — pool-aware, with SEPP withdrawals modeled, with property decisions as scenarios instead of assumptions, with real-terms projections instead of nominal fairy tales — my runway wasn't six months.

It was fifteen.

Not because I found hidden money. Not because markets went up. Because the structure of the analysis changed. SEPP withdrawals from a split IRA unlocked retirement funds I thought were trapped. Modeling real estate as monthly P&L instead of a static asset revealed which properties were drowning me and which were carrying their weight. Running scenarios let me keep Miami Beach — the place I actually want to live — instead of panic-selling everything the way a spreadsheet would have told me to.

Fifteen months. With a clear path to making it work within normal life expectations.

That's what the right model does. Not optimism. Not hope. Just better math applied to the right structure. The kind of math that lets you make decisions about your life instead of having your checking account make them for you.

I'm a systems architect by training. I think in data models, APIs, and pipelines. So I didn't build a spreadsheet. I built an engine.

FIREMaster connects to Monarch Money (where all 27 of my accounts already sync), pulls the raw data, and models it the way I actually need:

Pool-aware wealth modeling. Separate pools — liquid, taxable, retirement, real estate, private — each with its own rules, growth rates, and accessibility timeline. Not one number. Separate trajectories.

Scenario planning. Not one projection — three, five, ten named scenarios with different assumptions. What if I sell a property? What if I pick up consulting work? What if the market drops 20% in year one? Each scenario gets its own projection in real terms, adjusted for inflation. And here's what no advisor or calculator will ever give you: scenarios that account for what you want, not just what the math says. Any spreadsheet will tell you to sell the Miami property immediately — it's the financially obvious move. But Miami is where my life is. FIREMaster lets me model "keep Miami, sell something else" and "keep Miami, pick up bridge income" and "sell Miami as a last resort at month 36" — and compare all three honestly. The math informs the decision. It doesn't make it for you. You do.

SEPP/72(t) optimization. The IRS lets you split your IRA into two accounts — one for SEPP withdrawals, one for growth. The split ratio changes everything. FIREMaster calculates the optimal split.

Real estate as a P&L, not an asset. Every property has carry costs, rental income, mortgage payments, insurance, and taxes. FIREMaster models each one as a monthly cash flow, not a static line on a balance sheet.

I built it with Claude Code — Anthropic's AI coding tool — in about two months. The total cost in API tokens was roughly $300 so far. The time investment was real: long days, evenings and weekends, the kind of building you do when the alternative is lying awake at 3am running numbers in your head. Panicking.

Here's the part I didn't plan.

I built FIREMaster as a dashboard — charts, projections, a config page where I enter assumptions. Standard web app. But the most valuable conversations I had about my finances weren't with the dashboard. They were with Claude Code.

Because Claude Code can see the entire codebase. It can query the backend APIs directly. It can pull my real account data, run a scenario with different assumptions, and explain the results in plain English — in the same terminal where I built the thing.

I'd ask: "If I sell one property and use the proceeds to pay down the mortgage on another, what happens to my runway?"

And Claude Code would call the projection engine, run both scenarios, and tell me: "Selling extends your liquid runway by 18 months and eliminates the carry cost gap. Terminal wealth at 82 is 25% higher than the keep-both scenario. The tradeoff: you lose the rental income from that property starting immediately."

No dashboard can do that. No predefined chart answers that question. It requires a model that understands the relationships between pools, costs, income sources, and time — and can reason across all of them.

This is what I now call the dual-interface architecture: the dashboard is the cockpit (you glance at it to see where you stand), and Claude Code is the co-pilot (you talk to it when you need to think through a decision).

It turns out the product isn't just the engine. It's the combination of the engine and the powerful AI model that can interrogate it.

And if you want to understand why I've never been more excited to use computers in 40 years, listen to a recent podcast from my favorite modern day philosopher Naval Ravikant - "A Return to Code". He says it better than I can. I just happen to be living it.

If you're reading this, you're probably in one of two situations:

You're in the crisis. Laid off in your 50s, staring at a net worth number that feels large but a runway that feels short. The gap between "what I'm worth" and "what I can spend" is terrifying. Every calculator tells you you're fine. Your checking account says otherwise.

You're adjacent to the crisis. Employed, but watching the layoffs roll through tech. You've done the math and it doesn't add up the way it used to. You want to stress-test your plan before the email from HR lands.

Either way, this newsletter is the playbook I'm building in real time. Not financial advice — I'm an engineer, not an advisor. But frameworks, analysis, and the actual tools I use to navigate the bridge between layoff and 59 and a half.

Every issue: one concept, one framework, one thing you can apply to your own situation.

Next issue: "The Co-Pilot I Didn’t Plan" — what happens when Claude Code can see all your financial data and reason about it.

If this resonated, forward it to someone who needs it. If you're building your own tools with Claude Code, I want to hear about it.

— George

George Borshukov is a multiple Sci-Tech Academy Award winner, former Special Projects lead at Epic Games, and the builder of FIREMaster. He writes biweekly about FIRE strategy, quantitative finance, and building in public with AI.