I want to show it off as proof of my work. But the moment I do, everything connects
1. The reverse-portfolio problem
Say an anonymous project has grown to 12 languages and over 1,000 posts, with AI-assisted writing, translation, Markdown, GitHub and automation all running. As a track record to show web developers, that's pretty strong. But the moment you hand over the URL, you also hand over the JOIN key (the shared field that lets you match two databases) between "the anonymous writer" and "this real person." You want to show it because you grew it. You can't show it because you grew it too much.
2. Not a blog, a personality database
One post is just one opinion. Pile up enough of them and you can cut across all of it: what makes this person laugh, what they find annoying, how they weigh money against time, how they think about love, marriage and work, and what they get suspicious about and dig into further. What's stored isn't only text. It's a huge set of samples of input → judgment → reaction. It isn't the person, but it could be raw material for an AI that approximates "how would this writer respond?"
3. The dirty-joke search is far too good
Full-text search makes it worse. With a paper diary you'd have to flip through years of pages; with a search box, one word is enough. "What does this person think about this kind of thing?" → search → several hits → related words → older posts. Romance, sex, odd habits, pointless muttering to yourself: anyone can dig it all up self-service. It's way too searchable. It's like delivering your diary with a search engine and an index attached.
4. The combination is scarier than any single post
Privacy research keeps coming back to the same point: even with names removed, what matters is linkability, inference and re-identification against outside information. A 2025 paper in Scientific Reports assessed re-identification risk including how likely it is that sensitive information can be tied to a person and what the impact would be, and also looked at how exposed attributes already are in external data. A 2019 study in Nature Communications showed that richly detailed anonymized data can still be re-identified. The same goes for an anonymous blog. Work, trips, romance, events from a particular period may each be anonymous on their own, but once someone in real life knows who the author is, they become matching material. What matters isn't "did you write your name?" but "can all the posts be JOINed with real-world information?"
5. If they read it all and still say "marry me," that's a miracle
If a future partner reads the whole giant archive and still says "I want to be with you anyway," that's an absurdly strong form of due diligence on values. Normally it takes years to unlock someone's sense of money, how they live, how they see love, the things they're lazy about, their odd habits. Here it starts with "It's my long-term thinking log. It's full-text searchable." But the posts aren't the whole person. They leave out how you behave when living together, what you didn't write, and judgments that depend on the situation. It's a high-resolution user manual, not a marriage aptitude test.
6. You can separate how it works from who made it
For a development consultation, you don't need to show the real site. You can explain the tech by recreating the post count, language structure, generation flow, Markdown, GitHub integration and publishing pipeline with dummy posts. There's no need to pack a technical consultation and identity verification into the same packet.
7. Conclusion
Post A is anonymous, B is anonymous, C is anonymous. Yet JOIN A + B + C with information from people who know you in real life, and it may stop being anonymous. Once you tell someone "that's me," the link is almost impossible to undo. In a huge anonymous archive, the secret most worth protecting may not be any single post, but the JOIN key itself that connects your real-life identity to your anonymous one.
