As a rule, I don’t trust centralized platforms, even if they have strange but reassuring mottos or convoluted non-public benefit ownership structures. But to participate in today’s tech-driven society, sometimes you have to hold your nose and pick a horse. For me, Google and Anthropic represent the lessor of two evils. That’s why I have a Pixel, not an iPhone, and it’s why, despite my deep reservations, I just uploaded a year’s worth of my personal financial transactions and an 89-page overview of my entire financial life into Claude AI.
Am I f*cking crazy?
Possibly. But let me explain. My wife and I have been engaged in a long overdue financial planning project. Paying attention to money is one of my least favorite past times. But just like root canals or arguing about politics, there comes a time in your life when you can no longer avoid the pain. Earlier this year we changed financial advisors, and we took that transition as a cue to do a complete audit of our financial life. That meant a review of nearly thirty years of investments in startups, real estate partnerships, venture funds, and friends’ restaurants, not to mention a dog’s breakfast of stocks, bonds, crypto, and retirement accounts.
Every so often, the CEO of one of those startups or the fund manager of a venture firm would reach out with news – “Huzzah, we sold the company!” Or “Man, I’m so sorry, but we’re capitulating to market forces, and your investment has turned to zero.” (I’ve had to do many such calls as a founder, and I know how challenging they can be).
In addition, every quarter the professional venture, real estate, or stock funds would send a quarterly report, usually brimming with charts, spreadsheets, disclosures, and extensive commentary. It’s information overload – no way I had the time, or honestly, the desire to spend hours analyzing and understanding all the data I was being sent.
So … I didn’t. I’d file them away into my “investments” folder, where they festered like a pile of unread New Yorkers. I’d tell myself that the money we invested in risky startups was lost as soon as I wired it. And the money we put into stocks or bonds? A long term play – no point in worrying about it in the present. I’m not a professional investor, I told myself. I’m not going to waste time pretending I am.
Our new financial advisor was eager to get a “whole picture” view of our holdings, and to be honest, so was I. I’m getting to the age where it’s important to understand what, if any, flexibility our investments might afford us. My benign neglect of our financial life meant we had more than fifty distinct accounts across at least a dozen brokerages and investment portals, though most of what mattered was concentrated in one place (in our case, Fidelity).
Still, getting our arms around everything meant doing a lot of work. And to get back to the point of this post, I did most of that work the old fashioned way – I attacked the problem with shoe leather, phone calls, and endless email queries. I didn’t wire up OpenClaw (or – shudder – Muse) and give it unfettered access to my digital life. Instead, I spent months running down more than 40 angel investments in often long-dead companies – did we take a write off, or do we still hold an interest in the firm that purchased their assets back in 2018? Is this real estate fund still active? How did we end up with 200 shares of Fervo Energy?! What the hell *is* Fervo Energy?
We tracked our progress in a Google Sheet (I know, very 2006), and when it was close to finished, we packaged it up with every possible statement we could muster from Fidelity, Coinbase, Wells Fargo, and various other platforms, and shipped it over to our new advisors. We also sent them a completed 22-page questionnaire detailing just about every dollar we spent over the past twelve months.
Now it was their job to make sense of it all.
Some weeks later, our advisors sent us a draft financial plan based on all that data. It arrived in the form of an 89-page presentation featuring slides with titles like “Wealth Simulation Analysis” and “Summary of Alternate Assumptions.” They asked for an hour to review it all with us. Would Thursday work?
I mean, an hour? No way would that be productive. Decades of bad habits almost led me to file the whole damn thing into my investments folder. I could just smile and nod my way through the meeting, right? I’d certainly done it many times before.
But then I thought about my relationship with AI. I find Claude to be most valuable when I need to turn an unmanageable amount of unresponsive data into something I can directly query. It can turn a mute, two-dimensional corpus into an artifact with which I can have a conversation. What if I just uploaded this monster deck into Claude, and start asking it questions?
Now, if you’ve been reading this site for any length of time, you know I’m deeply suspicious of business models that leverage the personal data of the masses for corporate profit. That’s why I chose Claude in the first place – the company has rejected advertising, which means it won’t be incented to use my prompt data as grist for a surveillance capitalism millstone.
Still, uploading something this personal gave me pause. Would Anthropic be training future models on this 89-page summary of my financial life? Turns out, the answer was yes – but it’s dead easy to tell them not to, which I promptly did. I then asked Claude what kind of information I should redact from a financial document like the one I was considering uploading, and once I deleted names, birthdays, and the like, I dragged my deck into the AI’s maw.
I then held my breath as it ingested forty years of my financial life.
What happened next was, well, pretty awesome. I asked Claude to summarize the document, then began a two-hour conversation spanning nearly every question I could imagine asking my advisors. To check the assumptions our advisors had made about our spending, I ended up uploading an anonymized copy of all our outgoing transactions for the past year. (I know!). By the end I had compiled more than a dozen queries, clarifications, and follow ups, all of which I sent to our advisors prior to our meeting. Each became an agenda item for our forthcoming meeting.
Now I was deeply invested, and just as importantly, completely briefed. That one hour meeting – which for years had been a rote performance on both sides of the call – turned into a robust dialog. We felt more in control, and our new advisors sensed they were working with a client who knew what they wanted.
A phrase comes to mind as I contemplate this story: “Information asymmetry.” For years my relationship to my own money was defined by informational asymmetry: Advisors, investors, fund managers, bankers, lawyers would send dense barrages of information at me, and I’d simply duck and cover. I’d cross my fingers and hope, in good faith, that things would work out. But our recent financial audit has taught me that used properly, AI can restore balance to those kinds of relationships. I no longer feel overwhelmed when I think of my money, or of interacting with the people charged with managing it. Instead, I feel empowered.
I imagine the same is true for people in similar situations: Patients facing complicated diagnoses, homeowners attempting to understand mortgage options, consumers deciding which new car to buy. For as long as there’s been money, there’s been informational asymmetry. Perhaps – assuming we get the inputs and guardrails right – AI can help correct those imbalances.
That’s kind of the point of all this technology, isn’t it?
—