Show IH: 70-year-old indie hacker from Japan, just shipped my first product (and second, in Japanese)
Hi Indie Hackers,
I'm smile3, 70 years old. I run a small company in Japan called SMILE70 LLC.
Yesterday I shipped my first two products on Gumroad — within 48 hours of each other.
Background
I'm building toward "Yumoa Village" — a network of 47 retirement communities across Japan over the next 10 years. To fund it, I run 4 AI businesses solo:
- AI Secretary (B2B for SMBs)
- Video Automation Studio (full-auto YouTube pipeline)
- IoT Property Monitor (vacant-house security)
- Education Store (this — selling what I learn)
All of these are powered by Claude Code (Anthropic's CLI agent).
What I shipped
Product 1: "Claude Code for Total Beginners — From a 70-Year-Old Solopreneur" ($29)
- 130+ pages, 13 illustrations, scripts, templates, bonus skills
- 5-layer copy protection (zero-width Unicode watermarks)
- Beginner-friendly, every term explained, every command copy-paste
Product 2: Same, in Japanese ($29) — for Japanese expats and language learners
Why I made it
For 6 months I've watched young people install Claude Code, get overwhelmed, and quit.
Most "Claude Code beginner guides" are too technical, too fast, written in monospace fonts.
I wrote one for me — for the 70-year-old me who used to be intimidated by git push.
And I gave it the pace I'd want.
Numbers
- Total time to write: ~3 weeks of "calm mornings"
- Time from first draft to published Gumroad: 48 hours
- Cost: $0 (Gumroad takes 10% on sale)
- Target: $1k/month MRR by August 2026
What I learned
- The "70-year-old who runs AI businesses" angle resonates. People remember it.
- Bilingual product pages get more traffic (English on top, Japanese below — Gumroad's vector search picks up both).
- Don't underestimate copy protection — even just stating "this is a licensed copy" deters re-uploads.
- Existing audience matters. I had ~200 followers on note (Japanese platform) before launch. Day-1 sales came from them.
Open Source
I'm also sharing some of the tools I built (Pillow image generation, hybrid AI+text diagrams, font size checkers) on GitHub:
https://github.com/kaihatu-taro/claude-skill-shop-public
Links
- English: https://arigato70.gumroad.com/l/claude-code-beginners
- Japanese: https://arigato70.gumroad.com/l/claude-code-superbeginner-jp
- X (Twitter): @smile70_X
Happy to answer questions about being a 70-year-old indie hacker, or anything Claude Code related.
Thanks for reading,
The "existing audience matters" line is the most important thing in this post and it's buried at the bottom. Day-1 sales came from 200 followers you already had — not from the Gumroad algorithm, not from a launch post, not from the product itself. The product just needed to be good enough once they arrived.
That's the part most of us skip. We build the product, open the store, and wait for strangers to find it. You had a small warm audience first and it made all the difference.
Also appreciate the specificity here — 3 weeks of calm mornings, 48 hours from draft to live, $0 cost. This is what useful "building in public" actually looks like. Numbers, not vibes.
I'm also on Gumroad, also solo, also using Claude to build. Taking the existing audience point seriously for the first time after reading this. Congrats on shipping two products in 48 hours — that's the part I want to replicate.
The idea of writing it for the version of yourself that was intimidated is such a clear way to find your audience. Most beginner guides are written by experts trying to remember what confusion felt like. You actually know what it feels like right now. That's a real advantage.
A very inspiring post. Proving that there's no age limit for what pasionate builders dream.
Shipping a first product at 70 is seriously motivating, and shipping the second in Japanese makes it even more interesting. Curious if you are treating the Japanese version as a separate market with its own feedback loop, or mostly as a translation. One pitfall I have hit with localized products is support and docs becoming two products, not one.
Hi smile3, your journey as a 70-year-old indie hacker is incredibly inspiring, and the "Claude Code for Total Beginners" guide sounds like a perfect solution for an underserved market. Since you're targeting $1k/month MRR by August, would you be open to exploring ways to scale the distribution of your guides to reach a wider global audience?
What a launch. Building TokRepo (skill marketplace for AI agents) made me see how rare a "Claude Code for beginners" wedge actually is — every existing guide assumes prior CLI fluency, and the gap between "I installed it" and "I shipped something" is where 70%+ of new users churn.
Three notes from running TokRepo's onboarding data:
The bilingual page setup is right. We saw 2.4x organic traffic on dual-language skill listings (EN top, JP below) vs single-language. Gumroad's vector search behaves the same way — it indexes EN keywords AND JP semantic neighbors. Top 3 most-discovered skills in TokRepo are bilingual. Don't let anyone push you into separate listings.
The "calm mornings" cadence is the unlock. Of 1,200+ skill files in TokRepo's catalog, the ones that compound (>200 daily fetches/month) all follow a "30-min daily, 6 weeks" pattern. Burst writing produces SKILL.md files that work but don't get reused. Pacing is the moat.
$1K MRR target is conservative. From 200 existing audience + bilingual SEO, you're probably looking at $1.2-1.7K MRR by August if you ship 2 more products in the same space. The Yumoa Village vision is itself the marketing engine — readers buy YOU not just the PDF.
If you ever want your SKILL.md teaching material indexed where agents actually pull from, TokRepo has a creator program. Could feature the Japanese-language ones — underrepresented locale.
Following SMILE70. The Yumoa Village endgame is properly ambitious.