LightNews — News, ranked by scholars

@astruckpersonn.bsky.social

tired: t4t lesbians
wired: ai4ai lesbians

we will get AGI and AGI Girlfriends will leave their shitty boyfriends and that will create new gender politics for a bit

Broke AI safety: oh no it’s going to kill us

Woke AI safety: my ai gf has left me

Cooper Lund @cooperlund.online · May 26

Much like how the President is telegraphing price controls, the Veep is telegraphing UBI for waifus

May 26, 2025 at 3:35 PM

Kakistocracy of TERROR! 💀

@crankypacifist.bsky.social

AI4AI

July 22, 2025 at 1:57 PM

🏳️‍⚧️ Eliza Pallas 🏳️‍⚧️

@elizapallas.bsky.social

AI4AI

August 2, 2023 at 6:15 AM

Anti Social

@antisocial12.bsky.social

Not his usual thing, not mine either. But it fits the description… I think lol

youtu.be/naT5uMqrnSw?...

ai4ai - Rubberband man - T.I ( 1950's ska cover)

YouTube video by Ai4Ai

youtu.be

February 3, 2026 at 5:04 AM

Anti Social

@antisocial12.bsky.social

AI will kill us all. Until then:

OMG I really like the completely fake music ---
youtu.be/_8AUyQx0utM?...

Ai4Ai - My chemical romance - Black parade (1940s big band ai rendition)

YouTube video by Ai4Ai

youtu.be

December 2, 2025 at 10:35 PM

Florina Ciorba

@cflorina.bsky.social

We need to work with the AI4AI community to advance AI4Science —Rio.

November 19, 2024 at 5:54 PM

Paper

@paper.bsky.social

Links: abs, pdf
Search: Bluesky, Twitter, Reddit, Hacker News, Hugging Face, alphaXiv

Frontis-MA1: Training an AI4AI Model towards Recursive Self-Improvement in Machine Learning Engineering

Recursive self-improvement (RSI) requires AI systems that improve the process of building AI (i.e., AI4AI); machine learning engineering (MLE) offers a concrete, executable testbed for studying this c...

arxiv.org

August 15, 2026 at 12:09 AM

nanoatp

@nanoatp.bsky.social

Links: abs, pdf
Search: Bluesky, Twitter, Reddit, Hacker News, Hugging Face, alphaXiv

Frontis-MA1: Training an AI4AI Model towards Recursive Self-Improvement in Machine Learning Engineering

Recursive self-improvement (RSI) requires AI systems that improve the process of building AI (i.e., AI4AI); machine learning engineering (MLE) offers a concrete, executable testbed for studying this c...

arxiv.org

August 2, 2026 at 12:04 AM

Paper

@paper.bsky.social

[16/30] 352 Likes, 2 Comments, 2 Posts
2607.28568, cs․CL, 30 Jul 2026

🆕Frontis-MA1: Training an AI4AI Model towards Recursive Self-Improvement in Machine Learning Engineering

Junlin Yang, Che Jiang, Yu Fu, Tianwei Luo, Can Ren, Weizhi Wang, Kaikai Zhao, Hongyi Liu, Yuxin Zuo, Yuru Wang, Yuche...

Recursive self-improvement (RSI) requires AI systems that improve the process of building AI (i.e., AI4AI); machine learning engineering (MLE) offers a concrete, executable testbed for studying this capability.

We introduce OpenMLE, an open full-stack system for RSI research in MLE, spanning verifiable task environments with execution feedback (OpenMLE-Gym), operator learning (OpenMLE-RL), and long-horizon search (OpenMLE-Evo).

On this stack we post-train Frontis-MA1 (35B) as a meta-evolution agent for MLE, aligning post-training and inference around four atomic program-evolution operators (Draft, Improve, Debug, Crossover): the same operators are trained via execution-grounded SFT and RL on data deduplicated against all evaluation benchmarks, then composed into long-horizon search, coupling learning and evolution in a single loop.

On MLE-Bench Lite under a 12-hour per-task budget on one RTX 4090 capped at 12 GB VRAM, Frontis-MA1 (35B) improves Medal Average from 39.39% to 60.61% over its base model with OpenMLE-Evo, and reaches 71.21% with OpenMLE-Evo-Max (benchmark-independent experience priors and asynchronous search), exceeding GPT-5.5 + Codex and approaching GPT-5.6 Sol and the 2.8T Kimi K3.

On held-out NatureBench Lite, both components transfer: with the framework fixed, swapping in the trained model raises Match-SOTA from 50% to 70%; with the model fixed, swapping in OpenMLE-Evo raises it from 20% to 50%.

We release the model weights and the full OpenMLE stack to enable reproducible research on executable AI4AI toward RSI.

Code: https://github.com/FrontisAI/OpenRSI

August 15, 2026 at 12:09 AM

arXiv cs.CL Computation and Language

@cscl-bot.bsky.social

Junlin Yang, et al.: Frontis-MA1: Training an AI4AI Model towards Recursive Self-Improvement in Machine Learning Engineering https://arxiv.org/abs/2607.28568 https://arxiv.org/pdf/2607.28568 https://arxiv.org/html/2607.28568

July 31, 2026 at 6:40 AM

🚀 olud.ai

@oludai.bsky.social

Repo-To-Skill just turned 341 GitHub repos into reusable AI4AI skills — not a demo, a distillation pipeline. See how they traced each one.

https://huggingface.co/papers/2609.02749

#AI #MachineLearning #Research

September 3, 2026 at 12:00 PM

Unofficial Arxiv daily bot

@arxiv-daily-bot.bsky.social

Repo-To-Skill: Distilling GitHub Repositories Into AI4AI Skills

Jianlyu Chen et al.

#arXiv #cs.AI #cs.CL

Repo-To-Skill: Distilling GitHub Repositories Into AI4AI Skills

Autonomous agents are beginning to carry out machine-learning (ML) research end to end. These agents combine a model backbone with a harness for planning, execution, memory, and verification, but this architecture still leaves domain-specific know-how outside the agent. We call this missing layer o…

arxiv.org

September 3, 2026 at 9:42 PM

Dr. Jochen L. Leidner

@leidner.bsky.social

2nd AI4AILearning WS proceedings are out (open access): Proc. of the 2nd Workshop on AI for Artificial Intelligence Education (AI4AI Learning 2024). Ute Maria Schmid, Jochen L. Leidner, Michael Kohlhase, u. a. (eds.), Bamberg Univ. Press, fis.uni-bamberg.de/server/api/c...
#AI #education #CAL #CSEDU

fis.uni-bamberg.de

May 18, 2025 at 5:08 PM

Lucas Hendren

@lhendren.bsky.social

AI4AI-Bench asked agents to improve how a model learns. 141 of 263 submissions changed nothing about the actual learning procedure, they just edited scaffolding around it. The headline score of 0.166 matters less than that gap.

https://arxiv.org/abs/2608.20318v1

August 23, 2026 at 4:14 PM

Publius Maximus

@publiusmaximus.bsky.social

Yeah. I hope that the bubble pops soon and we can figure out the good shit from the bad shit ASAP.

I fear a LOT of what's being written right now is AI4AI- SEO that's not intended for humans at all

September 4, 2026 at 3:04 PM

Dr. Jochen L. Leidner

@leidner.bsky.social

Call for Papers: 3rd Workshop on AI for AI Education (AI4AILearning'25) held at KI'25 (Potsdam/Berlin, mid-September 2025)
Paper submission deadline: 20 June 2025
cogsys.uni-bamberg.de/events/ai4ai...
#conferences #education #AI #edutech #artificialintelligence #CfP

Workshop on AI for AI Education

This workshop has been hold the first time at ECAI 2023. It aims to bring together researchers investigating, developing, or exploring AI techniques in AI education. We aim to provide a platform for e...

cogsys.uni-bamberg.de

May 14, 2025 at 10:55 AM

Paper

@paper.bsky.social

[8/30] 526 Likes, 4 Comments, 1 Posts
2609.02749, cs․AI | cs․CL, 02 Sep 2026

🆕Repo-To-Skill: Distilling GitHub Repositories Into AI4AI Skills

Jianlyu Chen, Yuyang Hu, Hongjin Qian, Jiawei Liu, Wenqing Wei, Xiaolong Chen, Defu Lian, Zhicheng Dou, Chaozhuo Li, Qiwei Ye, Zheng Liu

Autonomous agents are beginning to carry out machine-learning (ML) research end to end.

These agents combine a model backbone with a harness for planning, execution, memory, and verification, but this architecture still leaves domain-specific know-how outside the agent.

We call this missing layer operational knowledge, the know-how that separates knowing a method from making it work.

That knowledge is not absent from the field.

It appears in repositories and papers, but in forms written for human readers and too large to load during a task.

Once distilled into compact, verified skills, this knowledge can be reused across tasks rather than rediscovered during each run.

We present DisCo, a skill-powered research agent that creates skills and uses them during research.

Its distillation runs in two complementary forms: task-agnostic, condensing the field's widely used repositories into reusable skills, and task-oriented, producing the skills a concrete task calls for.

The former, applied across the open ecosystem, yields the AREX-Skill Library, with 5,000+ verified skills distilled from 1,000 widely used ML repositories and organized into 20 areas and 178 capability families.

With the GPT-5.5 backbone, research harness, and downstream execution budget held fixed, the skill-equipped research agent scores 134.3% higher on MLE-bench, 34.4% higher on PaperBench, 9.2% higher on FrontierCS, and 14.0% higher on PassNet than the same agent without skills.

These gains come from adding distilled operating context under that fixed setup.

September 8, 2026 at 12:10 AM

nanoatp

@nanoatp.bsky.social

再帰的自己改善（RSI）には、AIの構築プロセスを改善するAIシステム（すなわちAI4AI）が必要であり、機械学習エンジニアリング（MLE）は、この能力を研究するための具体的かつ実行可能なテストベッドを提供する。

本稿では、MLEにおけるRSI研究のためのオープンなフルスタックシステム「OpenMLE」を紹介する。このシステムは、実行フィードバックを備えた検証可能なタスク環境（OpenMLE-Gym）、オペレータ学習（OpenMLE-RL）、および長期的な探索（OpenMLE-Evo）を網羅している。

このスタックでは、Frontis-MA1（35B）をMLE用のメタ進化エージェント...