Psyho (@fakepsyho) 's Twitter Profile
Psyho

@fakepsyho

Game Designer; Problem Solver; past: OpenAI (Dota), Pro Competitive Programmer, Poker

ID: 593047872

calendar_today28-05-2012 18:48:46

1,1K Tweet

2,2K Followers

336 Following

Psyho (@fakepsyho) 's Twitter Profile Photo

LLM wrapper that iteratively improves solutions for problems that can be automatically evaluated (like optimization problems). Weird release: used in-house, no serious benchmarks (like contests), no mention of costs, very vague paper. Just showing off with nothing useful shared

Psyho (@fakepsyho) 's Twitter Profile Photo

Given the rapid improvement of coding agents, sponsors in competitive programming will likely disappear within 1-2 years Even with slow adoption and minimal progress in AI, there will be simply too many developers and too few jobs (except in highly specialized niches)

Psyho (@fakepsyho) 's Twitter Profile Photo

AWTF questionnaires are published: info.atcoder.jp/entry/awtf-Heu… I think my good TCO performance comes from the fact that none of the finals' problems ever required beam search. Not sure if was too obvious about that😅

Psyho (@fakepsyho) 's Twitter Profile Photo

Since I can't fall asleep and had no time to properly research that. In OPTUNA, which sampling algorithm is the most suited for AHC?

Psyho (@fakepsyho) 's Twitter Profile Photo

Humanity has prevailed (for now!) I'm completely exhausted. I figured, I had 10h of sleep in the last 3 days and I'm barely alive. I'll post more about the contest when I get some rest. (To be clear, those are provisional results, but my lead should be big enough)

Humanity has prevailed (for now!)

I'm completely exhausted. I figured, I had 10h of sleep in the last 3 days and I'm barely alive.

I'll post more about the contest when I get some rest. 

(To be clear, those are provisional results, but my lead should be big enough)
Psyho (@fakepsyho) 's Twitter Profile Photo

I'm sorry, but did you just put limited number of lives in a puzzle game in 2025? I love the concept and I think it's a step in the right direction, but maybe you should work with experienced puzzle designers (actual experts in the field) instead of relying on crowdsourcing?