Two Weeks on X, or, How I Accidentally Stress-Tested a Forty-Four Billion Dollar Revenue System While Trying to Post About God
I have been on X for two and a half weeks. Not in the way people say they've "been on" a platform, meaning they created an account in 2019 and occasionally like something their cousin posts about sourdough. I mean active. Threads, media, paid promotion, the full architecture. My handle transferred cleanly from @tlcsolomon to @timothysolomon and I treated the whole thing like what it is, which is a new channel in a revenue system, because that is what I do for a living and also what I do for fun, which is either a great sign or a very specific kind of personality disorder depending on how you feel about people who get excited about pipeline attribution models.
I should explain what I mean by "what I do for a living" because it matters for the rest of this story. I design, build, and hand over complete revenue operating systems. Not campaigns. Not dashboards. Not the quarterly deck where someone points at a graph and says "awareness is up" while the CFO quietly dies inside. I build the actual connective tissue between marketing spend and closed revenue, the part most companies pretend exists but doesn't, the part where you can trace a pound spent on a Tuesday in March to a deal that closed on a Thursday in June and say with a straight face that you know why. That's the day job. By night I'm a finitist mathematician and philosopher who writes papers about theosis and transcendental constants and why Cantor's conception of infinity is, to put it diplomatically, something I have opinions about.
The reason both of those things matter is that my entire X strategy, such as it was, came down to a single bet: that the overlap between those two identities is more interesting than either one alone. Not as a branding exercise. As a hypothesis about what actually works on a platform whose recommendation algorithm claims to reward depth and originality over noise.
Here is what happened.
The Hypothesis
I did not come to X to "build a personal brand." I came to X because I had never used it before and I wanted to know if the same paid-to-organic flywheel I build for clients would work on a channel I had zero history with. The experiment was simple. Promote high-signal content to a targeted audience, let the early engagement velocity seed the organic recommendation engine, and see if the algorithm would do the rest. This is not a novel strategy. This is how revenue systems work everywhere. Input, process, output. The only interesting variable was the content itself.
So I posted about God.
Not exclusively, but the threads that actually moved were the philosophical ones. A long reflection on prayer as "becoming loving awareness," which is the sort of phrase that sounds like it belongs on a candle until you actually sit with it for a while. The announcement of a paper I've been working on for years called "The Name of the Universe," which maps theosis and grace across Abrahamic traditions and argues, among other things, that pantheism is really just nominalism wearing a nicer hat. A short, sharp observation about the word "mean" and how "you're being mean" isn't an accusation of cruelty, it's the feeling of being averaged out by someone who should be treating you as irreducible.
These are not the sort of posts that growth hackers put in their "Top 10 Hooks That Convert" threads. They are long, sincere, occasionally dense, and they assume the reader is an adult who can hold two ideas at once. I promoted them with money because I believe in paying for distribution when you have something worth distributing, and because I wanted to see what the Grok-powered recommendation layer would do with them once they had enough early signal to register.
What it did was pick them up and run.
The Numbers, Briefly
My Account Analytics, which shows only organic metrics because X keeps ad impressions in a completely separate dashboard that does not talk to the consumer-facing one, looked like this: the flagship "Name of the Universe" post hit 14,000 organic impressions and 151 likes. The prayer reflection hit 11,000 and 119. Public view count on the main post crossed 22,000 and kept climbing. My follower count, which had been sitting at a polite and unremarkable 117, spiked by 19 in a single day and continued ticking upward. The daily impressions chart went from effectively zero across the first two weeks to a peak around 34,000 in one day, which is the kind of hockey stick you usually only see in pitch decks where someone is lying.
Nobody was lying. The ads seeded the velocity, the reposts created social proof, and the organic algorithm did exactly what it's supposed to do when it encounters content that generates genuine engagement from real humans who are genuinely engaged. The paid spend was the spark. The organic reach was the fire. The system worked.
I was quite pleased about this, in the way that a systems architect is pleased when a system does the thing, which is to say quietly and with an immediate desire to run the experiment again with different variables.
Then X's ad platform migrated.
The Clusterfuck
I want to be fair about this because I ship my own buggy websites and I know what it's like to push something live that isn't ready because the alternative is never pushing anything live at all. But there is a difference between "we shipped with some rough edges" and "we hard-coded the Stripe address element to default to the United States and then deployed it to a global user base during an active billing migration," and the difference is that the second one means I, a person in London, cannot add a payment card to my account.
Not "have difficulty adding." Cannot. The form requires a US state. There is no state dropdown for non-US addresses. The postcode validation fails because it's expecting a ZIP code. The whole thing just sits there, politely refusing to acknowledge that countries other than America exist, which is a fascinating design choice for a platform whose owner talks a lot about free speech and the global town square.
Because I can't touch my payment method, I can't edit my running campaigns. I can't pause them. I can't adjust targeting or budgets. The ads are in a kind of quantum state where they are simultaneously active and uncontrollable, delivering impressions that may or may not be costing me money, and the dashboard gives no clear answer about which. I posted a thread about this and described the resulting free views as "thanks, X, I hope," which captured the spirit of the situation quite well I think.
The promoted-article flow is also, to use a technical term, half-baked. The image and headline requirements don't play nicely with the article format, the promote button leads you through a path that doesn't seem to know whether you're promoting a post or an article or a prayer to the machine gods of engagement, and at one point I found myself genuinely suggesting in a public thread that X should maybe just decide ahead of time whether they want users to promote articles directly or not, because the current answer appears to be "we haven't thought about it."
I tagged @ads, @AdsSupport, @elonmusk, and @grok with screenshots and reproduction steps because that is what you do when you find production bugs in a revenue system. You document them, you escalate them, and you propose fixes. This is, again, literally my job. The fact that I was doing it for free on someone else's platform during my second week as a user is either a testament to my professional instincts or evidence that I need better hobbies.
Support has been quiet. I understand why. I'm a small account with small spend during a platform-wide migration that affects everyone. The engineers are presumably busy. But the irony is thick enough to spread on toast: the bugs I'm documenting are the same bugs preventing me from using the normal channels to report bugs. It's a perfect closed loop of dysfunction, which is actually quite elegant if you squint.
What the Mess Proved
Here is the thing that matters, the thing underneath the funny frustration and the broken Stripe forms and the threads where I'm half-laughing and half-doing free QA for a company worth more than most countries.
The system held.
My paid-to-organic flywheel, the one I designed in my head during the first week and executed during the second, kept working even while the platform was actively breaking around it. The ads, despite being frozen in their quantum state of uncontrollable delivery, still seeded enough velocity for the organic algorithm to do its job. The philosophical threads still resonated. The follower growth still happened. The Grok recommendation engine still picked up the "Name of the Universe" post and pushed it to 22,000 views and counting, which means the semantic layer that evaluates content for depth and originality and emotional resonance did exactly what Elon's team says it does, at least for one finitist mathematician in London who was posting about theosis while his ad dashboard was on fire.
That's the real result of this two-and-a-half-week experiment. Not the numbers, though the numbers are good. Not the bugs, though the bugs are spectacular. The result is that the growth strategy I've been building and selling for fifteen years, the one that treats every channel as a node in a larger revenue system with inputs and processes and outputs and feedback loops, works on X the same way it works everywhere else. Even when the infrastructure is actively cosplaying as a beta.
I have never used this platform before. I built a system on it in two weeks. The system outperformed the platform's own infrastructure. I'm choosing to find that encouraging rather than concerning, though I reserve the right to change my mind.
The Uncomfortable Part
There is a version of this story that's purely about marketing expertise, and if that's the version you want you can stop here and go visit timothysolomon.com where I will happily talk to you about revenue operating systems until one of us has to eat or sleep.
But the version I actually want to tell is about the content that worked, because the content that worked was not about marketing. It was about prayer. It was about theosis, which is the Eastern Orthodox concept of becoming one with the divine, and about how that concept maps across every Abrahamic tradition and most non-Abrahamic ones too if you pay attention. It was about spending most of my life looking at people who prayed and feeling a complicated mix of pity and envy and then realising, very slowly and with some embarrassment, that I had the whole thing backwards.
That is the content the algorithm picked up. That is what 22,000 people saw. Not "tired of marketing spend that never connects to actual revenue?" Not the RevOps service pitch, which did fine on its own but didn't touch the philosophical posts. The thing that resonated at scale, on a platform I'd never used, during a migration that was actively breaking, was a man being honest about God and mathematics and the strange place where those two things meet.
I don't know what to do with that information yet. I think it means something about what people are actually hungry for, which is sincerity and depth and the willingness to say "I had this completely wrong" in public. I think it means the Grok recommendation engine, whatever its flaws, is doing something right when it surfaces that kind of vulnerability over the usual noise. And I think it means my instinct to lead with the philosophical work rather than the professional pitch was correct, not because philosophy is a better "hook" but because it's a better representation of who I actually am, and platforms that claim to reward authenticity should, occasionally, be tested on that claim.
They passed, this time. The ad platform failed in about nine other ways, but the recommendation engine passed.
The Offer That Is Only Slightly a Joke
I would genuinely love to consult on the post-migration ad platform. Not because I think Elon Musk is going to read this and DM me, though if he does I promise to be both useful and entertaining, but because the problems I found are real RevOps problems and I solve real RevOps problems and the overlap is sitting right there being obvious.
The Stripe address hardcoding is a localisation issue that affects every non-US advertiser. The campaign control freeze is a billing-dependency architecture problem. The promoted-article flow is a product-design gap that needs someone to sit down and map the user journey from "I wrote something worth promoting" to "it is now promoted" without six branching paths and two dead ends. These are solvable. I've solved harder.
But I can't file a proper support ticket because the same bugs I'd be reporting are preventing me from accessing the support flow that requires a valid payment method. Which, again, elegant in its dysfunction. A perfect little ouroboros of platform debt.
So the offer sits here, in public, on the platform itself, which is probably the most appropriate place for it. I'm @timothysolomon. I build revenue operating systems. I accidentally stress-tested yours during your own migration and came out with better organic numbers than most accounts see in their first year. The least interesting thing about me is that I'm good at marketing. The most interesting thing is that I posted about the nature of prayer and 22,000 people showed up.
Both of those things are true at the same time, which is, if you think about it, the whole point.
Here's the addendum:
Update, Two Days Later
The payments came back. Then they didn't. Then the ads halted. Then they un-halted. The country localisation issues are still there, and several other bugs I won't catalogue here because at some point a list of platform defects stops being useful documentation and starts being a man yelling at infrastructure, which is not the energy I'm going for.
But something happened that I want to document properly because it matters more than the bugs.
I kept posting. I kept tagging, @grok, @ads, @AdsSupport, the usual suspects, not because I thought anyone owed me anything but because that's what you do when you find problems in a system and you want them fixed. I’d like to believe tagging Grok specifically might have helped surface the threads to the right people, or maybe it was the cumulative weight of all the pings and the consistency and the fact that the posts themselves were detailed enough to be useful rather than just frustrated. I genuinely don't know which variable did the work. That's the thing about multi-touch attribution, you never really know, you just know the outcome.
The outcome was that a senior account manager reached out to me. Not a bot, not a templated "we're looking into it," a real person with a real title who offered to book a call and said they would take care of it personally.
I accepted, obviously.
I want to be very clear about something here. I know they have their hands full. I know the migration affected thousands of advertisers and that my account is small and my ad spend is, in the context of a platform this size, genuinely tiny. The fact that someone took the time to reach out to me individually, to offer a call rather than a ticket number, to say "I'll handle this" rather than "we'll escalate this," is not something I take for granted. It is sincerely, uncomplicatedly appreciated.
There is a tendency in public platform criticism to treat any response from the company as either insufficient or suspicious, as though the only acceptable outcome is total capitulation or a job offer. I don't think that way. Someone saw the problem, someone reached out, someone offered to help. That's good. That's how it's supposed to work. I'm grateful, and I wanted to say so in the same public space where I raised the issues in the first place, because fairness runs in both directions.
The call is booked. The experiment continues. And the system, both mine and theirs, is still working.
Big shout out to @VermeulenLX (hope its okay I tagged you here) for reaching out and taking charge. I can’t tell you how much I appreciate it.