Just links

EsoLang-Bench: Evaluating Genuine Reasoning in Large Language Models via Esoteric Programming Languages https://arxiv.org/abs/2603.09678

LATENT Learning Athletic Humanoid Tennis Skills from Imperfect Human Motion Data https://zzk273.github.io/LATENT/

Preventing Learning Stagnation in PPO by Scaling to 1 Million Parallel Environments https://arxiv.org/abs/2603.06009

Tech boss uses AI and ChatGPT to create cancer vaccine for his dying dog https://www.theaustralian.com.au/business/technology/tech-boss-uses-ai-and-chatgpt-to-create-cancer-vaccine-for-his-dying-dog/news-story/292a21bcbe93efa17810bfcfcdfadbf7

Efficient Exploration at Scale https://arxiv.org/abs/2603.17378

Offline Materials Optimization with CliqueFlowmer https://arxiv.org/abs/2603.06082

Academic Fraud Inclination Metric https://www.alexalemi.com/arxiv-metric/index.html

idk if there is at least 1 person who 1. reads this channel 2. will be interested 3. doesn't know about it already but I was asked to share this info, so here goes: There will be a competition called Midnight Code Cup. Main features: • Covers a much wider range of programming-related activities: algorithmic problems (not even sure about this one), optimization problems, pvp bots competitions, hardware challenges, esoteric languages, you name it • Can use whatever you want. Yes, even ChatGPT (the...

прикольно дизайнить процессоры агентами Design Conductor: An agent autonomously builds a 1.5 GHz Linux-capable RISC-V CPU https://arxiv.org/abs/2603.08716 https://www.alphaxiv.org/ru/overview/2603.08716

Research note: Many SWE-bench-Passing PRs Would Not Be Merged into Main https://metr.org/notes/2026-03-10-many-swe-bench-passing-prs-would-not-be-merged-into-main/