Three Days Before Launch, I Let My Own Tool Tear Me Apart
The database holds.
The server is configured with a three-minute timeout.
But my laptop fan has started making a strange whistling noise, and three days before launching Obliqo on Product Hunt, I have the almost mathematical certainty that something is going to break.
Before pressing that button, I wanted to write down three thoughts from the inside of this very small room.
1. The market does not care who you are
The market does not care where you come from.
It does not care how many nights you spent writing code until your eyes burned. It does not care how much the project matters to you. It does not care how carefully you built the thing.
It only asks one brutal question:
Does this solve a problem someone actually feels?
That is hard to accept when you come from research, writing, culture, or any kind of long-form work. A project can take years to build and still disappear in six months because nobody understands it, nobody finds it, or nobody is willing to pay for it.
That does not make the work meaningless.
But it does mean the work is not enough.
2. The solitude of the solo founder is real
A solo founder cannot handle everything forever.
You can pretend you can, but eventually the system tells the truth. The frontend slips. The backend needs attention. The copy is unclear. The launch page is half-good. The distribution plan is weak. The support emails arrive. The logs start whispering.
And still, accepting help is not simple.
Not all help is alignment. Not every collaborator sees the same map. Some people want a task. Some want a percentage. Some want proximity to the product without carrying the weight of it.
Finding real companions — people who understand not only the tool, but the reason it exists — is much harder than I expected.
So for now, I walk alone.
Not because it is heroic.
Because it is the only honest configuration I have.
3. My limit as a maker is embarrassingly simple
I do not know how to sell.
I can spend nights balancing the behavior of multiple AI agents, reading server logs, rewriting prompts, adjusting timeouts, or tracing why a response takes three seconds longer than it should.
But when I have to explain the product clearly, quickly, and without hiding behind theory, I hit a wall.
And that is a serious problem.
Because you can build something useful, original, even necessary — and still create nothing more than beautiful ghost code rotting on a cheap server if you cannot put it in front of the right people.
Obliqo’s paradox
Before publishing this post, I did the obvious thing.
I ran it through Obliqo — the pre-publish critique engine I built to find weak points in writing before the public does.
It gave me a BENDS verdict.
And it hit exactly where it should have:
You say the market is cynical and does not care about rigor, honesty, or depth. Yet you are launching a tool built on rigor, intellectual honesty, and critical friction. You expect people to pay to have their writing challenged. In practice, you are betting on a non-cynical market while claiming the opposite.
That is the contradiction.
And it is real.
Obliqo exists because I still believe there are people who do not just want smoother writing. They want stronger thinking. They want to know where their argument bends before someone else breaks it in public.
Maybe that is naive.
Maybe it is the whole bet.
I do not want polite encouragement. I built a tool that feeds on constructive criticism, so compliments are not very useful right now.
The Product Hunt preview page is here:
https://www.producthunt.com/posts/obliqo/maker-invite?code=QN2kzN
Tear it down.
Tell me where the messaging does not hold. Tell me where it sounds like generic AI marketing. Tell me what makes no sense. Tell me what you would not understand as a first-time visitor.
I am here to take the punches.
We will see if it holds.
The pre-launch dogfooding panic is something else. There's a specific
discomfort in switching from builder mode to user mode after months —
suddenly you're noticing every rough edge you'd quietly normalized.
What surprised you most? The worst ones are usually the things you'd
already "accepted" internally and just... forgot to revisit.
went to the Product Hunt preview. the headline is where the clarity problem lives. it describes the mechanism but not the outcome. a pre-publish critique engine tells me how it works. something like 'find where your argument breaks before your audience does' tells me why I want it. the former sounds like a tool. the latter sounds like insurance against embarrassment which is a much stronger motivation for the person who cares enough about their writing to pay for criticism
Whenever you feel lonely, just remember what pushed you to start your own business. For me, the regret of not taking action is far worse than any failure.
point 1 is the hardest one to internalize because the market indifference feels personal when you built the thing yourself. the distance between knowing it intellectually and accepting it emotionally is where most founders stall before launch.
Yes — exactly.
That gap between understanding it intellectually and accepting it emotionally is brutal.
I think that’s where a lot of solo founders either harden, disappear, or finally learn to listen better.
Thanks for naming it so clearly.
Impressive! What stack did you use?
Thanks rachidex!
SvelteKit, Appwrite, n8n, OpenRouter, single VPS. The differentiator isn't the stack, it's the four-agent architecture orchestration.
Section 3 is where I'd push back. "I don't know how to sell" isn't the gap — this post sold me on Obliqo harder than any feature list would. Your selling instinct switches on when you're naming a contradiction and switches off the moment you try to "market." So stop marketing. That BENDS verdict on your own launch is the strongest landing-page copy you've got — lead with the contradiction instead of hiding it.
This is probably the most useful punch I could receive.
You’re right: when I try to “market,” I get weaker. When I name the contradiction honestly, the product becomes clearer.
“Lead with the contradiction” is going straight into my notes.
Thank you.
You asked for honest feedback so here it is. The Product Hunt page headline needs to answer one question immediately: who is this for and what specific problem does it solve? Right now it leads with what the tool does conceptually rather than the pain the buyer already feels. The strongest version of your positioning probably lives in what you wrote here: people who want to know where their argument bends before someone else breaks it in public. That is specific, that is a real feeling, and that is a reason to pay. The landing page copy should open with that.
Landing page and messaging is what I work on. Happy to take a closer look if useful
That line is very strong.
You’re right: “where your argument bends before someone else breaks it in public” gets much closer to the real fear behind the product.
I had to compress it for the Product Hunt limit, but I’m testing this version now:
“Find where your argument bends before it breaks in public.”
I’ll keep “private pressure-testing for critical writing” as the clearer positioning line.
Thank you — this really sharpened the page for me.
You asked for punches, and your product is literally a pressure-test, so I'll review the page the way Obliqo would review a draft. Honest, not nice.
What holds: the top of the page is strong. "AI That Pushes Back" and "find the weak point before the world does" is a sharp, clear promise, and "most writing tools help you finish a draft" sets up the contrast in one line. I knew what this was and why I'd want it inside five seconds. Don't touch that part.
Where it bends, hard: the method section. "The Friction Orchestra method, grounded in Cognitive Rhythm Theory, a formal model of structured cognitive friction published on Zenodo." You were worried about sounding like generic AI marketing. This is the opposite problem, and a riskier one. It's a wall of capitalised, proprietary theory a first-time visitor has no way into. To your exact audience, skeptical people who publish under their own name, a grand private framework is the first thing they pressure-test, and they bounce off it. "Published on Zenodo" also works against you with anyone who knows Zenodo is an open upload repository, not peer review, so it reads as borrowed authority. And you reach for that authority in the same breath as "no academic affiliation." Pick one posture. I'd pull the theory naming off the landing page and let the output prove the rigor.
Second bend: the four voices are written beautifully but abstractly. "Questions the premise, not the prose. The uncomfortable angle" reads well, but I can't picture what it actually returns. The fact-check voice is the exception, it's concrete ("checks your facts against the web, with sources") and instantly clear. Show one real example of what a voice says about one real sentence, and the whole concept clicks. Right now I have to take the poetry on faith.
Small one: "creators, consultants, and anyone who publishes under their own name" is three audiences in a trench coat. Who is the one person you're building for first? The page would hit harder aimed squarely at them.
Technical PS, and one of these actually matters. Your "Example Run" demo (the draft v3 one) grows its own height as it animates, which pushes everything below it down and then snaps back on each loop. Scroll past it to read and the page keeps yanking you, which is a rough thing to do on a product about careful reading. Lighthouse rates layout shift at basically zero, because it only watches the first few seconds of load, but a real reader feels this the whole way down the page. The fix is to give that demo a fixed height so the text animates inside a stable frame instead of reflowing the page. The smaller stuff: a console error on load, two title tags (the tab and Google use the weak "Obliqo" one), and headings that skip levels. The duplicate title is a thirty-second fix worth doing before the Product Hunt traffic lands.
One question, the real one: if you had to ship with only plain language plus a single concrete example, and delete either the Zenodo/theory framing or the four-voices poetry, which one goes?
Respect this approach
tbh, i'm also currently preparing an MVP launch too and the hardest part is figuring out whether a problem is the product or just uncertain abt the launch.
What was the most painful thing your tool exposed?
Thanks for the question — it was actually useful.
I tried to answer it inside Ashive because I wanted to see how your product handles the offer/positioning problem in practice, not just reply in the abstract.
The onboarding felt good. The first question was sharp and relevant: “what exactly are you selling and what outcome does it create for the buyer?” That is probably the right place to start for a pre-revenue product.
One piece of feedback: I hit the login wall right after that first serious answer. I understand why — sessions, saved context, costs, abuse prevention, etc. Obliqo has the same kind of constraint, even if I offer 5 free runs per month after signup.
But as a user, I felt I was just starting to understand the value when the gate appeared.
“Save and continue your session” is clear, but maybe a bit functional. I might have felt more pull from something closer to the outcome, like:
“Save this context and get your offer diagnosis.”
or
“Continue to your first positioning diagnosis.”
That said, the question itself was strong. It also made me reflect on Obliqo: the signup is justified, but the user needs to understand the value of the critique before the account step feels natural.
So thanks — your product already gave me a useful mirror.
Thanks for reading.
The thing I’d really like feedback on is not the code, but the positioning:
does “pre-publish critique engine” make sense immediately, or does it still sound too abstract?
Also: would you personally use something that challenges your draft before publishing, or does that feel like too much friction?
Possibly.
The reason I'd be careful is that I don't think those are necessarily two separate questions.
The decision sitting underneath them may be more important than either answer on its own.
That's not something I'd try to work through casually in a thread.
I think you’re right.
At launch, “is this a product problem?” and “is this a positioning problem?” are not always separate questions.
If people don’t understand the promise, that may look like marketing failure — but it can also mean the product surface is not translating the value clearly enough.
That’s why I’m trying to test the language as hard as the code.
For Obliqo, the dangerous abstraction is “pre-publish critique engine.”
The concrete version is probably closer to:
“Before you publish, find the weak point someone else will attack.”
That’s the gap I’m trying to close.
I’d be careful with one thing.
The interesting part isn’t whether the wording makes sense.
It’s whether users understand the value before you explain it.
That gap usually decides what needs to change first.
Good point — thanks.
You’re right: the real test is not whether the wording makes sense after explanation, but whether the value is understood before I explain it.
That gap probably tells me what needs to change first: the product surface, the positioning, or both.
Possibly.
I think the harder part usually isn’t the wording itself, but what decision the wording is trying to solve for.
That’s where things start to diverge.