thatwillnever .com

↓ a laboratory · run by one guy · est. 2026

a home for ideas people said would never work — until some of them did ↘

I run experiments in public. Half of them flop. A few of them hit. All of them get written down — what they cost, what they made, and what I learned either way.

see what's cooking → read the autopsies (the honest ones)
manifesto

(scribbled on a napkin, typed up later)

Every interesting thing I've ever built started with someone saying "that will never work."

So I named the lab after the sentence.

what this place IS:

  • → a working portfolio of bets
  • → real money, real hours, real answers
  • → small weekend experiments
  • → bigger, scarier ones too

what this place is NOT:

  • a course
  • a funnel
  • a get-rich anything
  • just a guy with ideas (that's me)

the deal: I'll show the work. Launches AND flops. Costs AND revenue. If a thing cost me $800 and made me $40 — that's what the autopsy will say.

if it works, great — I'll tell you how.
if it doesn't, better — I'll tell you why.

what's cooking

not yet post-mortem'd. jury's still out.

★ featured
LIVE current focus · launched 2026

yourSaaS.diy

chat-first SaaS-as-a-service. describe what you need — a doc, a dataset, a mini-app, or a full site — and an AI consultant builds it live in the chat. you preview the real deliverable before you pay. every Site ships with a free in-place editor (Squarespace charges $20/mo for less). productized wedding-website store is templatized end-to-end — intake form, photo upload, one-command build, hosted on a wildcard subdomain with SSL. Etsy + Pinterest pull traffic back into the live chat. no subscriptions, no logins, no "learn our platform."

thesis →
most people want the fish, not the fishing rod
tiers →
Doc $49 · Data $97 · Mini-App $179 · Site $499
stack →
Supabase + Stripe + Resend + Claude
status →
live + CRM-complete: admin dashboard, refund self-service, hosting fulfillment, 4-stage email nurture. wedding store templatized end-to-end. Stripe is LIVE — real money mode as of July 3. platform's green across the board; the bottleneck is traffic. Etsy + Pinterest launch is the whole ballgame now. north star: $1k/day in 90 days
visit yoursaas.diy →
DORMANT token + game

$DOORS

a community-driven real estate strategy game wrapped in a Solana token. each 21-day "season" is a fictional property deal — holders vote, hazard & fortune cards wreck plans, and it all plays out in public. formerly FlipCoin.

→ admin hub v2 designed. parked while yourSaaS.diy ships.
PAPER trading system

Prediction-Market Bot

a backtest-first prediction-market trading bot. started on Polymarket, pivoted to Kalshi in May. now runs a portfolio of a dozen tiny paper strategies — crypto momentum, weather ensembles, mean-reversion, lead-lag — each on its own little budget so I can kill the losers one at a time.

→ killed the live longshots June 8 after a hard lesson: paper fills every signal, real money only fills you on the losers. moved closer to the money — one near-money strategy is back trading live in micro size, hitting 66.7% across its first 18 live bets and basically tracking paper. thin margin still, but the closer-to-the-money thesis is holding up.
ON HOLD AI SaaS

MissedNOmore

an AI-powered missed-call recovery system for local service businesses. when a plumber misses your call, the AI calls back within 60 seconds, qualifies the lead, books the appointment, and sends a recap. originally called LeadCatch — renamed for brand reasons.

→ code ~90% recovered. paused while yourSaaS.diy ships.
COOKING freelance ops

Jimmy

the Business Swiss Army Knife — a React command center for Upwork freelancing. job scouting, proposal generation, pipeline tracking, delivery management. built for one user (me) on the theory that "built for one" sometimes becomes useful for many.

→ v3 shipped. waiting on Upwork Connects.
WRAPPING UP real estate

Open Orinoco LLC

a house-flipping LLC. two properties purchased and rehabbed in 2025, both sold March 2026. currently filing taxes and closing out the books.

→ properties sold. 2025 tax filing in progress.
DESIGNING commercial real estate

Flex Space Garage

a flex space development concept — 1,500 to 3,000 sq ft units built for contractors, tradespeople, and small service businesses. co-funded by investors who want long-term passive income from commercial real estate.

→ investor proposal v2 done (with the honest NNN cost math). broker calls + demand test next.
SHIPPED annual-cycle tool

Year End Books

a local Python web app that turns Chase bank statements, payroll exports, and a folder of receipts into year-end financial statements. one-file standalone version, drag-and-drop UI. built in a single April session for the 2025 C-Corp filing — comes back to life every spring.

→ v1.0 shipped. 2025 books closed. standalone version ready to hand to other solopreneurs.
INVESTIGATING recovery op

NOMY Disaster

a forensic recovery on an alleged crypto scam — NOMy Finance, rebranded from Cryptonomy. ~157,000 tokens locked up across vesting until Dec 2028. token is "live" with fake DEX liquidity, 99% insider concentration, TokenSniffer 0/100. two tracks: watch for a real liquidity event so I can sell, and file fraud reports with Action Fraud UK + FCA + KuCoin.

→ ~$3k already clawed back. pool's still only a few hundred bucks against a $50k sellable line, so the locked tokens stay locked. watching, not selling into that friction.
HUNTING distressed real estate

Multifamily Distress

deal-sourcing on a collapsed ~3,300-unit multifamily portfolio scattered across DFW and San Antonio. forensic-traced the whole ownership web, ruled out the dead ends, and narrowed it to four live targets — bank-owned, mid-foreclosure, and off-market.

→ four targets ranked. next: chase the servicers + watch the auction postings. no capital plan yet — that's the scary part.
NEXT always something

[?????]

there's always a next idea. this tile is a placeholder + a reminder that the list of "that will nevers" is basically infinite.

→ slot reserved

autopsies

the honest post-mortems. nothing polished.

every project eventually lands here. costs, revenue, what I learned, what I'd do different. no survivor bias, no retconned origin stories.

▸ Crypto Recovery Audit · closed May 2026 — the BTC was already gone
open autopsy
spent: $0 + 3 days · made: closure · net: —
ran for: 3 days (May 1–3, 2026)
the idea: I was pretty sure I had 5–10 BTC sitting in a wallet somewhere from 2012–2019. forensic investigation to find it before BTC went any higher.
what happened: searched 3 old hard drives, 2 Gmail accounts, every Blockchain.com wallet I could remember, and ran an on-chain Bitcoin trace. found ~2 BTC moving through my wallet across 4 addresses in Feb–Apr 2017.
why it died: the BTC went into "Wealth Rising" — a 2017 cryptocurrency MLM/Ponzi run by a guy named Gerhard and a guy named JohnMac. the platform collapsed mid-2017. exit-pivoted to "Global Wealth Assistance" in October. all gone. ~$2,200 at 2017 prices. ~$190K at 2026 prices.
what I learned: "I'm pretty sure I had bitcoin somewhere" memories from a decade ago are usually wrong about where it ended up. on-chain forensics are the easy part. accepting it went to a scam in 2017 is the harder part. also: 54 emails from wealthrising.net are still sitting in my Gmail like a museum exhibit.
would I run it again? yes — closure beats lingering "what if." would manage expectations earlier next time.
▸ Quick $5k · died in the research phase
open autopsy
spent: $0 · made: $0 · net: $0
ran for: 1 session (~2 hours)
the idea: an AI + money-making research sprint. find prediction markets, Claude-native opportunities, or passive income plays that could generate ~$5K fast.
what happened: web searches ran. a research doc started. then the session ended and the thread never got picked back up. no direction was chosen, no opportunity was evaluated, no plan was made.
why it died: it was a "what could I do to make money" question with no constraints, no customer, and no deadline. that's not a project — it's a mood. without a specific angle to chase, it dissolved into the ether.
what I learned: "make money fast with AI" is not a project brief. the Polymarket Bot came from the same impulse but survived because it had a specific thesis (whale signals are exploitable) and a specific budget ($300). the constraint made it real. Quick $5k had no constraint, so it stayed imaginary.
would I run it again? no — but the instinct was right. it just needed a sharper question.
▸ [template] · how future autopsies will read
peek inside
spent: $X · made: $Y · net: $Z
ran for: N days
what worked: …
what didn't: …
what I'd do different: …
would I run it again? …
follow along

next experiment, straight to your inbox

one email when it launches. one email when it ends. no drip funnels.

(prototype — form doesn't actually send yet)