Opportunity → Personal CFO Agent
Framework → The Growth Experimentation Framework
Tool → Claybird
Trend → Prediction Market Trading Volume
Quote → Ideas Need Execution
🔗 Houck’s Picks
My favorite finds of the week.
Fundraising
Growth
ICYMI
Opportunity: Personal CFO Agent
As your startup’s finances get more complex, every founder eventually hires a CFO.
Most founders also have relatively complex personal finances too (hopefully without a comparable burn rate though 😅). Multiple credit cards, crypto, stocks, etc.
So, why not do the same thing?
A human CFO is impractical for personal use, too expensive, and it would just be kind of weird unless you are mega loaded and have a big family office.
But what about an agentic option?
Existing personal budgeting tools are honestly pretty outdated.
Even well-regarded ones like Origin, Monarch, and Copilot aren’t AI-first and maintaining them likely takes more time than a busy person can give.
Common issues like connections with bank accounts breaking, or incorrect categorization, could be easily solved with an agent.
🧠 Framework: The Growth Experimentation Framework
Does your startup run true experiments?
An experiment is when you’re trying to learn something you legitimately do not know. It is not where you’re just trying to confirm you’re right about something.
Most of the little things we call experiments often fall into the latter bucket.
I love experiments because they expand what we know. Sometimes that leads to true clarity. Other times it leads to more questions to answer.
In both cases, there’s progress.
That’s why I like the growth experimentation framework.
You give yourself 30 days to learn something new (that’s sizable, and that you legitimately do not already have a strong guess about).
During those 30 days, follow these steps:
Define what you’re trying to learn and make an initial hypothesis. What question are you trying to answer?
Design the way you’re going to learn it. Generate a hypothesis, identify the change you’ll make to observe what happens, and how that will get you to an answer.
Run the experiment. Set your plan in action.
Analyze the results. Interpret what happened (without introducing bias).
Decide on next steps. Do you need to run it again, maybe with a tweak to the design? Or can you action on the learning? And what new questions did the results spark?
You might not always need 30 days. But don’t let these sprints take longer. Anything you can’t learn in 30 days can likely be broken down into smaller pieces before testing anyway.
🛠 Tool: Claybird
We’ll talk more about validation a bit later in this issue, but validating video concepts has historically been pretty hard.
To really know how people will react to a video, you have to actually make the video.
Expensive, time consuming, etc.
While Claybird’s AI prompt-to-video ads can likely still be outperformed by real recordings, you can use them to get good data on how concepts will perform relative to one another.
They’re backed by YC as well, so there’s some optimism that maybe the product will eventually reach true production readiness.
But, even in the meantime, it’s worth it.
📈 Trend: Prediction Market Trading Volume
You probably know people who are losing way too much money on betting platforms.
First, if you do, try to help them, gambling addiction is dangerous.
But second, where they’re losing that money has likely changed at least somewhat this year.
For years, FanDuel and DraftKings were the main games in town.
But since the election last year, and more recently in particular (after Polymarket and Kalshi both raised massive rounds and became unicorns), prediction markets are the rising option.
This seems unlikely to change. To be honest, it still feels quite early for them.
I’d go as far as to say these platforms can be 10x bigger than FanDuel ($31B valuation) and DraftKings ($16B market cap) because they allow bets to be made on such a wider range of things than sports.
Likely a bad thing for society, but also a space where there will be a lot of problems that need solving.
💬 Quote: Ideas Need Execution
Validating your ideas is just step one.
Validating helps fine tune your direction. Your strategy. Your understanding of the market.
But it doesn’t build your business.
AI lets you iterate faster than ever, but don’t let that be an excuse to get caught in a cycle of endless testing.
You’re never going to really know the path forward until you try to go out and grow the business.
Our worst fears lie in anticipation.
Or, in this case, validation.
Your ideas don’t need validation.
They need execution.
— Kay (@Dwriteway)
10:57 AM • Oct 6, 2025