STUDIO TORIUMI / SHADOW
People say the truly excellent hide from public view. Investigated properly, it's the other way around. All ten people below left traces in openly accessible places while they were still unknown — arXiv, a personal homepage, GitHub. Not hidden. Simply not where people think to look. The main ranking is written in Japanese; this page is a complete English introduction to it, not a machine translation.
At the time of each achievement, nearly all of them were unknown.
Proved the Poincaré conjecture, still the only Millennium Prize problem ever solved. He never submitted to a journal — three preprints on arXiv in 2002–2003, that was all. He declined the Fields Medal (a first) and the million-dollar prize, and left mathematics.
TRACE — three arXiv papers, readable by anyone right now. The fact that a proof of the century never passed through peer review is, in itself, the thesis of this list.
In 2013, at age 58, an unknown lecturer — who had worked in accounting and delivery after his PhD — proved that infinitely many prime pairs differ by less than 70 million: the first finite bound toward the twin prime conjecture. Annals of Mathematics accepted the paper in three weeks, extraordinary speed for a top journal.
TRACE — arXiv and the Annals. Obscurity was a fact of his résumé, not of his ability. The referees saw it in three weeks.
Heard about the Conway knot problem — unsolved for over 50 years — at a conference in 2018. Solved it in about a week, as a graduate student, and reportedly didn't tell anyone for a while because she didn't think it was a big deal.
TRACE — arXiv:1808.02923, "The Conway knot is not slice." One preprint ended fifty years.
Possibly the most productive programmer alive. FFmpeg (which processes a large share of the world's video), QEMU (a foundation of virtualization), and Tiny C Compiler — each started alone. He also found a new pi formula and set a computation record on a single ordinary PC, not a supercomputer.
TRACE — bellard.org, a plain homepage with a plain list. Almost no social media. This is the purest "shadow" shape on the list.
In August 2012, released four papers, about 600 pages, claiming a proof of the ABC conjecture (Inter-universal Teichmüller theory). He published them on his own Kyoto University webpage, not in a journal. International consensus on correctness still does not exist — this page states that plainly, and includes him anyway, because the list collects work placed outside conventional channels, not settled results.
TRACE — his own page at Kyoto University. Six hundred pages that shook number theory, sitting in one place for over a decade.
Won the 2022 Fields Medal for resolving long-standing conjectures in combinatorics using tools from algebraic geometry. His path is remarkable: he wanted to be a poet, dropped out of high school, and came to mathematics late, with unremarkable grades — the strongest counterexample to "you must be a prodigy to become a mathematician."
TRACE — arXiv. A break in the résumé is not a break in the ability.
Cryptographer behind Curve25519 and ChaCha20 — quite possibly part of the connection you're using to read this page. He also sued the US government over crypto export controls and won, helping establish that code is speech. Today's free cryptography stands on that case.
TRACE — cr.yp.to, a personal site that looks like the 1990s and contains the present state of the art.
Solved optimal sphere packing in 8 dimensions — a problem that took 400 years even in 3 dimensions — with a proof famous for its brevity and beauty, then co-solved the 24-dimensional case within weeks. Fields Medal 2022. Outside mathematics she was almost unknown until the award.
TRACE — arXiv. The proof appeared as a preprint the moment it existed; anyone could read it from day one.
Built the foundations of distributed systems: Paxos, logical clocks — and LaTeX, on which a large share of the world's papers are written, though few of those authors could pick him out of a crowd. Turing Award 2013; he stayed mostly out of sight, building the specification language TLA+.
TRACE — his Microsoft Research page, where his papers are listed with annotations written by himself — a remarkably rare format.
After winning the Fields Medal, he publicly announced errors in his own earlier papers and turned to the limits of human proof-checking. That work became Univalent Foundations — an ongoing project to rebuild mathematics in a form machines can verify. He died in 2017, at 51.
TRACE — arXiv and his IAS pages. The only entry on this list that begins from "I was wrong."
Not one of them was hiding. When building this list I also believed that "the real ones sit in private networks." Tracing all ten showed that this is false:
That's why this list required zero leaks. Everything traces back to public primary sources. If something truly exists only in private, that is because its owner decided not to make it public — and that is not ours to dig into.
This is the practical part. If you want to find exceptional people before they're famous, watch these five places:
One tip: don't search for "great people." Search for "strange papers nobody has cited yet." Heavily cited work has already been found. Only the unseen places leave room to be early.
Ranking is by achievement relative to fame, not by ability — ability itself can't be compared, and this page doesn't try.
Living people are included. Only publicly documented achievements and publicly posted material are referenced; nothing about private lives or whereabouts.