12 years ago I was a failed computer science student, wasting my time on drugs. Having failed so many classes I did not see any future at all, and I was considering killing myself to get out of the anxiety and stress.
My confidence regarding programming etc was 0.
I decided to apply for a junior developer job. I got an interview and to prepare for an interview I found this site, Project Euler. I did ten or so tasks.
The interview started out pretty bad, they asked me some technical questions which I did not give good answers to, and I saw that they where not impressed. Then they wanted me to solve two programming problems on a white board. Imagine the relief I felt when both of these questions where from the ones that I solved on Project Euler a couple of days before! I nailed them and the interviewers where clearly impressed. In the end they hired me and motivated it with that although I lack a lot of theory I am obviously a very good coder haha.
Anyway, that was what I needed, when I got this job I quit the drugs and got my act together. 12 years later I live a confortable life as a freelancer and have even managed to build my own SaaS with paying customers! Thank you Project Euler.
Good for you! Inspiring stories like yours are positive externalities that will no longer be extant in the same way when everyone’s minds are just thin clients on top of LLMs!
I'd love to but I can't be too specific due to the personal stuff I shared in the original comment.
It's a system that helps services companies within a certain industry to digitalize all their paperwork, report to the national goverment agencies etc. They do a lot of manual work which can be digitalized easily.
I found this opportunity by just randomly throwing out in a big community that I build software and is looking for ideas, and some guy answered that ended up being my business partner for 3 years now. We are not rich from it but earn like $2000 each a month after tax which is quite a lot for us since we live in a country where healthcare, schools, parental leave etc are covered by taxes. And we dont need to put more than a few hours a month on support. I have put in basically all my spare time for 2 years to get to this point though, the biggest reward is not the money but the process of sitting through the nights being completely in the zone and building this stuff knowing that it will be great :D
I got an email six years later (in November 2024) telling me they accepted the problem:
Greetings!
Please excuse the email. It's quite possible that you are no longer active at Project Euler, but we are currently trying to retrospectively acknowledge contributors of problems.
You were identified as being involved with Square subsets [619 m] in Sep-17.
If you would like to be added as a verified contributor, please reply to this email and provide the username you currently use at projecteuler.net. Please note that contributor usernames will not appear publicly, rather they will be used internally to unlock new contributor awards we are introducing.
I was curious if LLMs are good for this problem. ChatGPT-5.1-Thinking one-shotted a correct Python script without any library use (https://pastecode.io/s/jg6ggxpm).
Claude Opus failed to solve after trying for a while.
First of all, don't do this. No one cares. Secondly obviously every Euler solution is going to appear in the training data many times over. It's no surprise at all that an LLM can regurgitate data that was given to it.
This is one of my favourite problems, I still remember that it has a very real edge case even though I solved it more than 10 years ago. Thank you for the problem!
I'm guessing if you only calculate based on the digits, the probability is going to be slightly different than the real one, because you only have a finite number of plates you can choose from.
That sounds like a combinatorial problem...alphabets from AAA to ZZZ, numbers from 000 to 999.
That means one of the total sum of possible car plates is 26^3.
Since we want to find pairs (x, y) that x + y = 1000. That means the total sum would also add up sum([1 for x in range(1000) for y in range(1000) if x + y == 1000])/2 since there is a symmetry.
But wait, find the expected number of plates he needs to see for a win. So maybe we need to borrow something from statistics (Possion/chi-squared distribution) or queueing theory...?
Edit: ah I saw the solution, it is a Markov chain.
Interesting- I ask a license plate question (when will california run out of plates in its current serialization format, based on a couple of plates observed in two different years). It's a much simpler question, though (just linear extrapolation).
Ha! I did a lot of these around 2007-2008. I solved the first 100 or so ones. Solving these were more fun than any other programming problems site I've seen. I quit it when I realized I was spending 4+ hours per problem - I couldn't justify the time.
Beyond a certain point, most required some knowledge of elementary number theory.
Same for me, the first 100 are doable, after that they take more and more time and require deeper mathematical knowledge. Which is nice as well, but at some point I started to wonder if it wouldn't be better if I spent all that time on learning some more practical knowledge.
Same timeline for me! This was my great procrastination side quest during grad school. I learned haskell and a bit of ocaml working on these. Functional languages are extremely well suited to these kinds of problems.
I agree with you. About a decade ago, I got really tired of coding interview prep. I somehow ended up getting into Project Euler and worked through the first 100 problems. It was a great way to learn a bit of number theory and regain my interest in programming.
This website was a part of my education as a computer scientist and seeing it here again I'm curious for the full story of how this site was made. Who made it, what do they do now, is it part of a broader project they have?
The most fun on this site is solving a problem and then having your mind blown by solutions in Apl/j/k and trying to guess what they mean without knowing anything about those languages
The biggest thing I learned from PE was that neither elegant theory nor brute force had a monopoly on successful optimization strategies. It's been something I've carried with me ever since and has over and over again proven its value.
I've solved about a hundred PE problems in Livecode, maybe 40 in Python, and about 20 in J. I highly recommend giving it a try in a language you don't know, it's fun! Especially with something as obscure as J.
I have done Project Euler very sporadically since high school. To date I've only done problems 1-54. Despite this, I think the website was a big contributor to my love of programming. I found it before knowing about leetcode and I think it prepared me well. I think I owe Project Euler a decent %-age of my eventual entry into Software Engineering as a field.
What's really neat is that most of the problems are so compact that you can just read about the full problem, then spend hours and days thinking about possible solutions.
Readit News
I decided to apply for a junior developer job. I got an interview and to prepare for an interview I found this site, Project Euler. I did ten or so tasks.
The interview started out pretty bad, they asked me some technical questions which I did not give good answers to, and I saw that they where not impressed. Then they wanted me to solve two programming problems on a white board. Imagine the relief I felt when both of these questions where from the ones that I solved on Project Euler a couple of days before! I nailed them and the interviewers where clearly impressed. In the end they hired me and motivated it with that although I lack a lot of theory I am obviously a very good coder haha.
Anyway, that was what I needed, when I got this job I quit the drugs and got my act together. 12 years later I live a confortable life as a freelancer and have even managed to build my own SaaS with paying customers! Thank you Project Euler.
Not one interviewer had heard of it.
Glad it worked out for you!
It's a system that helps services companies within a certain industry to digitalize all their paperwork, report to the national goverment agencies etc. They do a lot of manual work which can be digitalized easily.
I found this opportunity by just randomly throwing out in a big community that I build software and is looking for ideas, and some guy answered that ended up being my business partner for 3 years now. We are not rich from it but earn like $2000 each a month after tax which is quite a lot for us since we live in a country where healthcare, schools, parental leave etc are covered by taxes. And we dont need to put more than a few hours a month on support. I have put in basically all my spare time for 2 years to get to this point though, the biggest reward is not the money but the process of sitting through the nights being completely in the zone and building this stuff knowing that it will be great :D
The problem was based on 2013 Putnam Exam problem A2 (https://kskedlaya.org/putnam-archive/2013.pdf) which I took at the end of undergrad.
I got an email six years later (in November 2024) telling me they accepted the problem:
Greetings!
Please excuse the email. It's quite possible that you are no longer active at Project Euler, but we are currently trying to retrospectively acknowledge contributors of problems.
You were identified as being involved with Square subsets [619 m] in Sep-17.
If you would like to be added as a verified contributor, please reply to this email and provide the username you currently use at projecteuler.net. Please note that contributor usernames will not appear publicly, rather they will be used internally to unlock new contributor awards we are introducing.
Regards,
Project Euler Team
Claude Opus failed to solve after trying for a while.
I'm so happy to have spent twenty years of my life learning math and solving problems on Project Euler and elsewhere.
That means one of the total sum of possible car plates is 26^3.
Since we want to find pairs (x, y) that x + y = 1000. That means the total sum would also add up sum([1 for x in range(1000) for y in range(1000) if x + y == 1000])/2 since there is a symmetry.
But wait, find the expected number of plates he needs to see for a win. So maybe we need to borrow something from statistics (Possion/chi-squared distribution) or queueing theory...?
Edit: ah I saw the solution, it is a Markov chain.
Deleted Comment
Beyond a certain point, most required some knowledge of elementary number theory.
There is only very basic info here: https://en.wikipedia.org/wiki/Project_Euler
This article is excellent. It mentions the creator, Colin Hughes, and the story behind the project, but it boils down to a passion project for him.
That said, raw brute force often did far better than you'd like to admit.
A real gem of a resource.
Also a great way to learn a new programming language is to try project Euler problems with it.
I created a PDF version that I keep on my Remarkable for puzzling: https://github.com/pveierland/project_euler_offline