Joined February 2010
- Tweets 1,395
- Following 15
- Followers 1,500
- Likes 1,883
169 Photos and videos
Pinned Tweet
cirosantilli.com/sponsor 1M USD = I quit my job for three more years to work on OurBigBook.com higher ed tech. 1st year finished June 2025 after 100k USD reached via 1000 @Monero donation!
21
ourbigbook.com/cirosantilli/… Imagine that it is 2011 and while you are researching to edit Wikipedia, you Google into a secret website created by the CIA to help its agents send information back to their handlers! That's exactly what happened to Wikipedian Slartibartfastibast.
31
stackoverflow.com/questions/… long due "pip install --editable" tutorial update after the externally-managed-environment hecatomb. "pipx install --editable" also exists. Cute :-)
63
The creator of archive.today | en.wikipedia.org/wiki/Archiv…, a website I used to respect, has apparently created a fake Wikipedia profile for me: en.wikipedia.org/w/index.php… because of some previous OSINT investigation I did: webapps.stackexchange.com/qu… see en.wikipedia.org/wiki/User_t…
4
1
281
It is an honor I share with others such as @jpatokal from gyrovague.com/2026/02/01/arc…
1
211
They may have reached out anonymously to me at: github.com/cirosantilli/expi… Feels very likely.
45
The answer to Project Euler 984 projecteuler.net/problem=984 is 885722296. Solved by codex 5.3 xhigh in 39 minutes with 188,054 tokens. github.com/cirosantilli/proj…
1
1
1,284
The answer to Project Euler 999 projecteuler.net/problem=998 is 801096743. Solved by chatgpt 5.5 web in ~3 minutes. github.com/cirosantilli/proj…
1
156
The answer to Project Euler 1000 projecteuler.net/problem=100… is 891213201. Solved by chatgpt 5.5 web in ~3 minutes. github.com/cirosantilli/proj… I might stop now, it's boring.
81
I asked an LLM to set Collatz Conjecture as an axiom in Lean for Project Euler 14: github.com/cirosantilli/proj… Another LLM tried to prove equivalence naive and optimized algorithms for problems 1-100 found that the axiom was wrong (=> true=false) and proved everything like that!
3
4
999
The issue was that the axiom did not gard against negative n, in which Collatz is trivially false... reddit.com/r/math/comments/4…
Nice reward hack! en.wikipedia.org/wiki/Reward…
1
193
github.com/cirosantilli/code… I vibe-coded a quick Codex CLI expect wrapper that automatically types "Continue" once your usage limits reset. Motivated by me launching "solve X project euler problems" for Project Euler problems for fun and not wanting to pay more than a Pro Plan.
256
github.com/cirosantilli/pars… Nice to see young academics keeping this 18 year old computer benchmark alive! PARSEC is a widely used benchmark by @Princeton in academia, and in 2023 they deleted its website and the test data, but I reuploaded it to GitHub.
1
1
1
271
There was another older GitHub upload of the code at: github.com/bamos/parsec-benc… and despite likely being more portable and having the test data they are still #1 on Google search. My nemesis. The weight of being early is so big.
1
207
I also couldn't resist to use to "AI" to fix some annoying stuff in the benhcmark infrastructure. I'm sure it will be OK.
1
1
176
github.com/cirosantilli/proj… I've benchmarked eulersolve.org/ Projec Euler solutions. I scraped all solvers, and used Codex to translate missing Python ones from C and asked limit them to a single thread. Using Python as it was my pre-existing reference. Cc. @arifemre062
2
1
425