| # | User | Rating |
|---|---|---|
| 1 | Benq | 3792 |
| 2 | Kevin114514 | 3678 |
| 3 | VivaciousAubergine | 3647 |
| 4 | jiangly | 3582 |
| 5 | strapple | 3515 |
| 6 | tourist | 3473 |
| 7 | Radewoosh | 3418 |
| 8 | Um_nik | 3376 |
| 9 | maroonrk | 3361 |
| 10 | turmax | 3345 |
| # | User | Contrib. |
|---|---|---|
| 1 | Qingyu | 162 |
| 2 | adamant | 148 |
| 3 | Um_nik | 146 |
| 4 | Dominater069 | 143 |
| 5 | errorgorn | 141 |
| 6 | cry | 138 |
| 7 | Proof_by_QED | 136 |
| 8 | YuukiS | 135 |
| 9 | chromate00 | 134 |
| 10 | soullless | 133 |
Should we allow an Open AI agent to use their AI for testing during contests?
of course this will straightly against Rule Restricting the Use of AI
but maybe Mike can give an exception, exclusively for Open AI agent to use AI for testing on live contests
this will going to be really interesting I think, imagine AI become """really""" LGM
No.
Maybe we can set up a separate standings table and rating system for AI.
Then we wouldn't be able to see the performance and rating of AI compared to humans, though
We can merge the standings table together when displaying, but calculate the rating change separately.
bro is 24/7 sharing AI shit just to demotivate people here.i can tell you it doesn't work, find another way.
finally, brother found his family
CF is a good benchmark for models, make a contest for models only, lets see teams participating, with models bounded by no of parameters allowed, RAM required, max power required etc etc, Who makes the most efficient model wins
Yes. Because right now all the alleged rating achievements of these models are "trust me bro"-based. First of all it would be cool to see a definitive proof of those ratings: link to an account with a large sample of contests. Also it would be fun to see the meltdown of the anti-ML crowd when eventually a model achieves number one spot and stays there permanently.