Off the top of my head (and with no endorsement of related politics/ideology):
Paul Graham:
- That web-based apps were the future of software; the product that made him rich the first time around was Viaweb, the first ever web application, built in about 1995 when everyone else was selling thick-client Windows desktop apps in shrink-wrap boxes (http://www.paulgraham.com/road.html)
- That startups using niche languages like (in 2004) Python and Ruby would do better than those using mainstream platforms like Java and MS .Net (http://www.paulgraham.com/pypar.html)
- That startups were becoming dramatically cheaper/easier to start, and therefore could be started by fewer, younger people, didn't need "business people" early on, could get much further without needing to sell controlling stakes to VCs, and could effectively be mass-produced (http://www.paulgraham.com/webstartups.html)
- That a team of founders who funded themselves by selling novelty breakfast cereal would build an amazing company, even if their idea seemed to suck (http://www.paulgraham.com/airbnb.html)
- That a new, tech-focused news discussion forum, in a world that already had Slashdot and Reddit, could become popular and could remain popular for many years by employing specific mechanisms to raise the standards of content and discussion (http://paulgraham.com/hackernews.html)
Nassim Taleb:
- He explained the fundamental weaknesses in the economic system that would lead to crashes like 2007-08, and the present turmoil.
- Was predicting that a pandemic was the world's greatest risk at least 7 years ago (don't have time to sift through all the material but plenty is out there if you care to search for it)
Peter Thiel:
- Identified in 1993 that Silicon Valley offered more compelling career opportunities than NYC legal/securities work
- With Max Levchin, predicted that a money transfer platform would be one of the most important pieces of infrastructure, so launched PayPal, and built the team that enabled them to win the market
- Made first investment in Facebook, predicting that it would be a key piece of internet infrastructure due to its focus on identity, and that Zuckerberg - then aged 20 - would be able to execute
- Identified that big data analytics could be useful/attractive for government intelligence agencies, founded Palantir
- Was one of the first to suggest, in 2011, that the U.S. higher education market was in a bubble (https://techcrunch.com/2011/04/10/peter-thiel-were-in-a-bubb...). (Granted this hasn't been proven yet but there are signs the current crisis may be the needle that brings on the burst).
- Understood that a well-executed operation could take down Gawker
- Saw that Trump could win in 2016, and would be worth his while to endorse/support
===
No doubt they've all made predictions that have turned out to be wrong or mistimed. I follow Graham and Taleb on Twitter and often see them make comments I disagree with or doubt.
But it's not really; he was posing a question, which helped him to make further decisions about how to proceed as he built YC.
YC itself has funded many more companies that failed than succeeded, but that's fundamental to the business model - http://www.paulgraham.com/swan.html
What matters is that they're able to identify investment opportunities with a very big upside potential relative to the downside risk, and back them even if almost everyone else believes they're insignificant/wrong/stupid.
Thiel's investment fund Clarium Capital has famously lost plenty of money, but other investments of his have performed spectacularly well; that's how investing works.
Bottom line: all three of them have become vastly wealthy by repeatedly making bets that others didn't notice or thought were stupid.
You say this is measured by the number of "true non-obvious predictions" - can you list those for each of these folks?
I'd also think a batting average would be useful, instead of just a count of "correct predictions".