I Don't Know What the Future Looks Like. Here Are Five Bets I'm Making Anyway.
Not predictions from authority. Bets from uncertainty — by a CPA trying to figure out what to do today.
I'm not a researcher. I'm not making predictions because I enjoy futurism. I'm trying to figure out what to do today.
When I moved to Canada in 2004, I was 19, alone, and I picked accounting because it was the most reliable path to a job for an immigrant kid with no network. That decision made sense for twenty years. I built a career on it — audit, FP&A, commercial banking. The CPA opened every door.
2004
Moved to Canada alone at 19. Picked accounting — safest path to a job.
2007 — 2025
EY → FP&A → Commercial banking. CPA opened every door. Two decades of compounding expertise.
2025 — now
AI changes everything. Will any of this matter in ten years?
Now I look at what AI is doing to knowledge work and I genuinely don't know if an accounting degree will mean anything when my son goes to university. I don't even know if universities will work the same way by then. Everything I understood about how professional careers are built is being questioned at once, and the people who should have answers — the institutions, the firms, the professional bodies — seem just as disoriented as everyone else.
This isn't an intellectual exercise. I'm sitting with real uncertainty about whether the skills I spent two decades accumulating are about to be commoditized. These five theses are my attempt to make some sense of what's coming so I have some idea how to prepare.
Here's what I think happens when the cost of building software drops to near-zero — which it already has.
01
HIGH
The system is the moat, not the products.
Any single product a domain expert builds can be cloned by another domain expert in weeks. The code isn't the hard part anymore. What's hard to replicate is the system that produces products — how you evaluate ideas, how you govern quality, how each build cycle makes the next one faster. I've been building that system for four months now. If I'm right, it's the only thing that compounds.
02
HIGH
Trust is the only differentiator when building is free.
Right now, "CPA who builds software" is an unusual story. Give it two years and it's a LinkedIn bio line that 50,000 people share. When everyone can ship code, features converge, prices race to the bottom, and UIs all start looking the same. The only question left is: do I trust the person behind this product? Do they actually know their domain? Are they honest about what works and what doesn't? That's what I'm betting on.
03
MEDIUM
Your day job is your R&D department.
There's a strong pull to quit the day job and build full-time. I think that's usually wrong, at least early on. My banking work gives me a live feed of real problems — real clients, real regulatory edge cases, real market dynamics. The moment I leave, that feed starts drying up. The dual life isn't the constraint everyone thinks it is. It's actually the advantage. At least until enough customers are using the product that their feedback replaces the day-job intelligence.
04
HIGH
The intelligence layer is the product, not the software.
The code for any of my products is a few thousand lines. Rebuildable in a weekend by anyone with Claude Code. The intelligence architecture — which data sources to trust, which edge cases actually matter, which regulatory interpretations are correct — that took months and draws on years of professional experience. I think build time should be roughly 80% intelligence design, 20% code. Most people I see are doing it the other way around.
05
HIGH
Fragmentation is the existential risk.
AI makes it so easy to start building that the temptation is to build everything. I have projects across trade compliance, intelligence briefings, cybersecurity, tax, investment analysis. The code scales. Customer relationships don't. Each product needs real users, real support, real feedback loops. I can write code for eight products. I cannot have the phone call that saves a churning customer for eight products at the same time. This is the thesis I'd push on hardest — because I'm living the tension right now.
The Tension
These theses pull in opposite directions.
I'll come back to these quarterly and be honest about what's tracking and what I got wrong. That's the deal — predictions are only worth anything if you're willing to revisit them publicly.