Let me be brutally honest: The Geodexes kind of sucks right now.
Not in a "humble brag" way. Not in a "we're being modest but actually crushing it" way. We genuinely, legitimately suck at several critical things that a job market infrastructure platform absolutely needs to be good at.
And I'm going to tell you exactly why that's true, what we're doing about it, and why sharing this matters.
The Current State of Suck
Here's our report card as of July 14, 2025:
User Experience: C-
We have a landing page. It captures emails. That's... about it. The actual platform? Still mostly Figma files and GitHub commits that don't quite connect yet.
Market Penetration: F
We've talked to 47 potential users. Actual users on the platform? Zero. Well, unless you count me testing our own signup flow 73 times.
Revenue: F (Obviously)
$0 MRR. We're not even at the "ramen profitable" stage. We're at the "thank God for savings accounts" stage.
Technical Infrastructure: C-
We can process resumes with AI. Sometimes. When the API doesn't timeout. And when Mercury retrograde isn't affecting our server response times.
Go-to-Market Strategy: D
Our strategy was "build it and they will come." Spoiler: They don't come. You have to go get them. Who knew?
Why This Level of Transparency?
You might wonder why I'm publicly admitting that our startup—which is supposed to revolutionize job market transparency—currently can't even properly display a user dashboard.
Three reasons:
1. We're Building Trust Infrastructure
The Geodexes is fundamentally about transparency in the job market. How can we ask employers to be transparent about their hiring processes if we're not transparent about our building process?
If we only shared our wins, we'd be just another startup selling hopium. The job market is broken partly because everyone's pretending everything is fine. We refuse to be part of that problem.
2. This Is Actually Valuable Data
Every founder goes through this phase. But most only talk about it in retrospectives after they've raised Series B. "Oh yeah, the early days were rough," they'll say, conveniently leaving out the panic attacks and the pivot meetings.
Real-time documentation of the struggle is rare. And it's exactly what other founders need to see.
3. Our Users Deserve to Know
If we're asking job seekers and employers to trust us with their career decisions, they should know exactly who they're trusting. Not the polished version. The real version.
The Specific Ways We Suck (With Data!)
Let's get granular. Here's what our actual metrics look like:
Technical Debt Already (Yes, Already)
Lines of Code Written: 14,000+ Lines of Code Thrown Away: 8,000+ Refactoring Sessions: 17 "This Time We'll Get It Right" Moments: 6
We've rebuilt our authentication system three times. THREE TIMES. For a platform with zero users.
User Research Paralysis
We've conducted 47 user interviews. Here's the breakdown:
- Job seekers who said "this would be amazing": 31
- Job seekers who actually signed up for updates: 12
- Job seekers who responded to follow-up emails: 3
- Job seekers actively engaged in helping us build: 1 (Thanks, Marcus!)
The feedback loop we dreamed of? More like a feedback dot.
The AI Integration Disaster
Remember when we said we'd use AI to parse resumes and provide instant feedback? Here's how that's going:
- GPT-4 API calls that worked perfectly: 40%
- Times the AI completely misunderstood the resume: 25%
- Times the AI decided someone's hobby was their job title: 15%
- Times the API just... didn't respond: 20%
One candidate's resume listed "Marathon runner" under hobbies. Our AI classified them as a "Professional Athlete in the Running Industry."
Marketing Attempts That Flopped
Our Twitter strategy:
- Thoughtful threads posted: 23
- Average engagement: 3 likes, 1 retweet (usually my co-founder)
- Viral moments: 0
- Times I've considered just posting memes instead: 47
Our LinkedIn presence:
- "Thought leadership" posts: 8
- Actual thoughts led: 0
- DMs from people selling me lead generation: 1,347
What We're Learning From Sucking
Here's the thing about sucking: it's incredibly educational.
Lesson 1: Nobody Cares About Your Features
We spent two weeks building a "smart resume parser." Users spent two seconds telling us they just want to know if a human will read their application.
Feature development: 80 hours
User value delivered: ~0 hours
Lesson 2: The Market is More Broken Than We Thought
We knew the job market was bad. We didn't know it was "people have completely given up hope" bad.
One user told us: "I don't need another platform. I need therapy for job search trauma."
That's when we realized we're not just building software. We're building hope infrastructure.
Lesson 3: Perfect is the Enemy of Shipped
We've been polishing features for an MVP that no one's using. It's like vacuuming a house that's still being built.
New rule: If it works 70% of the time, ship it. Fix it when someone actually complains.
Lesson 4: Building in Public is Scary but Necessary
This post? I've rewritten it 4 times. Because admitting failure publicly goes against every entrepreneurial instinct.
But every time we've been honest about struggles, something magical happens: people help. They share advice, introduce us to others, or just say "me too."
The Plan to Suck Less
We're not giving up. We're getting better. Here's how:
Next 30 Days: Operation "Actually Ship Something"
- Launch a basic job board. Not revolutionary. Just functional.
- Get 10 real users. Not sign-ups. USERS.
- Fix the AI only after real humans complain about it
- Stop building features nobody asked for
Next 90 Days: Operation "Prove the Concept"
- Get to 100 active job seekers
- Onboard 5 employers (even if they're friends)
- Generate $1 in revenue (yes, the bar is that low)
- Document every single failure and learning
Next 180 Days: Operation "Stop Sucking"
- Achieve product-market fit (or pivot)
- Build features users actually request
- Create sustainable growth channels
- Maybe, possibly, potentially not suck
Why This Matters
The job market is broken because everyone's pretending it's not. Employers pretend their 47-step interview process is "efficient." Job seekers pretend they're not dying inside after rejection #237. Platforms pretend their AI can solve human problems.
We're done pretending.
The Geodexes sucks right now because we're trying to solve a real problem, not a convenient one. We're building something that matters, not something that's easy.
And if that means publicly documenting our journey from "complete disaster" to "slightly less of a disaster" to "actually useful," then that's what we'll do.
Join Us in Sucking (Then Not Sucking)
If you're building something hard, you probably suck at it right now too. That's fine. Let's suck together, then get better together.
If you're a job seeker tired of platforms that pretend everything's fine, we're building for you. It won't be perfect. It might break. But it'll be honest.
If you're an employer who knows hiring is broken but doesn't know how to fix it, let's figure it out together.
Email me at kevin@thegeodexes.com. Tell me how your thing sucks. Let's compare notes.
Because here's the truth: Every successful company sucked before it succeeded. The difference is whether you're honest about the journey.
We choose honesty. Even when it sucks.
P.S. - This post took me 3 hours to write because I kept stopping to fix bugs in our platform. The irony is not lost on me.
P.P.S. - If you're a developer who's built authentication systems that don't need to be rebuilt every month, please DM me. I'm not too proud to beg.
Meta Notes for Publishing
Twitter Thread Version: Focus on specific failure metrics with humor LinkedIn Version: More professional tone, emphasize learning and growth Reddit Version: Extra technical details about what went wrong Follow-up Content: Weekly "Suck Less Updates" showing incremental progress