Project Euler Forum

wim59 wrote: Tue Aug 26, 2025 2:36 pm

echip wrote: Tue Aug 26, 2025 12:16 pm

wim59 wrote: Sun Aug 24, 2025 8:21 pm I guess the proof will be in the pudding...

...I think you're simple-minded.

As far as me being simple-minded: You may be right, but I choose to believe that toilet man has a conscience that we can appeal to, and to me his reply (“I will try to be more mindful for very new problems”) suggests that may be the case.

To avoid a lengthy debate that blurs the point, I'll just address the key points. You believe he has a conscience, but I don't. Why did you believe a few of his casual comments?

In the long history of Project Euler, you weren't the first one to ask him not to release the answers so early. Did he listen? Why would he listen to you, but not others? Are you special? Isn't it clear?

echip wrote: Tue Aug 26, 2025 3:56 pm ... Are you special?

No I am not special. I have only been doing PE for three-and-a-half years, and am not up to date on all its history. I'm just a guy trying...

My 2¢: Let's "Speak softly and carry a big stick". Toilet guy is now infamous among a fairly small set of PE insiders. The latest problem has the note "...this problem was spoiled...". It could have read "...this problem was spoiled by /in/libai ..." instead, and no matter how much of a conscience he has, I bet he wouldn't like to be infamous amongst all active PE users and hence try to avoid further escalation if he understands what the consequences might be.
If you think it makes sense, I am willing to convey the message to him. Of course, I cannot speak for the PE team, but I can make it clear to him that he is playing with fire.

Obergscheidle wrote: Wed Aug 27, 2025 10:43 am Toilet guy is now infamous among a fairly small set of PE insiders. The latest problem has the note "...this problem was spoiled...". It could have read "...this problem was spoiled by /in/libai ..."

Do you think he cares about his infamousness. The account "unLuckytoilet" only exists on ProjectEuler. As for the name he uses on his github repo,Li Bai,I bet it's a fake name.

Do you know what Li Bai means? Li Bai was a great poet from the Tang Dynasty of China (born in 701 AD and died in 762 AD).

https://en.wikipedia.org/wiki/Li_Bai

Come on, anyone who was educated in China knows who Li Bai is.

While many Chinese people share the same name, I bet this isn't a coincidence. He specifically chose the name Li Bai to pay tribute to this great Tang Dynasty poet.

He is a fake person, using a fake name. Everything else is fake. Only the ProjectEuler answers on his repo are real. He's not afraid of notoriety at all.

You are wrong echip.
Bai Li is the real name of this guy.
Search e.g. on Bai Li Toronto.
Toronto is the city where he went to university and got his PhD.
He was in high school in Waterloo (Canada).


I have been following him from 2009 remember
Perhaps you can stop talking nonsense now

echip wrote: Wed Aug 27, 2025 1:24 pm

Obergscheidle wrote: Wed Aug 27, 2025 10:43 am Toilet guy is now infamous among a fairly small set of PE insiders. The latest problem has the note "...this problem was spoiled...". It could have read "...this problem was spoiled by /in/libai ..."

Do you think he cares about his infamousness. The account "unLuckytoilet" only exists on ProjectEuler. As for the name he uses on his github repo,Li Bai,I bet it's a fake name.

Do you know what Li Bai means? Li Bai was a great poet from the Tang Dynasty of China (born in 701 AD and died in 762 AD).

https://en.wikipedia.org/wiki/Li_Bai

Come on, anyone who was educated in China knows who Li Bai is.

While many Chinese people share the same name, I bet this isn't a coincidence. He specifically chose the name Li Bai to pay tribute to this great Tang Dynasty poet.

He is a fake person, using a fake name. Everything else is fake. Only the ProjectEuler answers on his repo are real. He's not afraid of notoriety at all.

I'd like to mention that Chinese is not written in Roman characters. The map from Chinese character to Pinyin is many-to-one and therefore no valid inverse.

As much as it pains me to say it, I think Project Euler probably needs to rethink the fastest solver boards. While I do not at all condone people posting their solutions externally (far from it), it won't be long until LLMs will be able to solve the latest problem pretty much instantaneously (I'd be surprised if they can't solve it already to be honest).

Anyone who competes in Advent of Code as well will know that the leaderboards for that site are more or less meaningless. Sure, they show who the best prompt engineer is but anyone trying to understand the problem and hack together a solution under their own steam has zero chance of getting on the leaderboard.

So, to my mind the PE team has three options:
1. keep the PE leaderboards as they are, noting that being on there is not the achievement that it once was,
2. invest more and more resources into detecting cheat solutions, or
3. come up with a new way of ranking submissions (maybe transitioning to a code-based submission rather than solution-based submission, and/or publicly ranking by forum kudos or similar).

Update: After several emails, the maintainer of one of the spoiler repositories (Bai Li aka luckytoilet) agreed to wait until 50 solvers are on the leaderboard before publishing a spoiler from now on. Not quite what we wanted, but perhaps it is a reasonable compromise.

Many thanks to you, Obergscheidle.

Obergscheidle wrote: Fri Aug 29, 2025 6:01 pm Update: ...(Bai Li aka luckytoilet) agreed to wait until 50 solvers are on the leaderboard before publishing a spoiler from now on...

He did, in fact, add the following to his README page: "To help maintain the integrity of the leaderboard, new answers will be published only after at least 50 solvers have submitted solutions."

Interesting. According to a Project Euler administrator, I have been delisted for now. The reason given was LLM usage, not Googling or Project Euler Minus style. In the unofficial Discord I mentioned the structure of formal competitive programming platforms, but upon further reflection think there is something unappealing about forcing participants to submit C++ codes with a runtime limit on a judge server. One path forward is to strive towards significantly trickier tasks [out of the reach of LLMs], but this requires composing said tasks. I know some users who might intend to continue solving partially with LLM assistance from primary delisted accounts... and even insinuate that solves still imply some level of human problem solving ability. I do not know if limiting the leaderboard to 50 is wise. Also, his claim "Obviously, this is quite pointless, as Project Euler ranks can gain you nothing in the real world." is materially false and provably thus.

My unsolicited summation of the entire episode is: there is some timepass to be had with Project Euler problems, leave it at that!