Awesome AI4AI
š
2
Search 223 papers on whether AI can reliably improve AI
kai
KaiWu123
AI & ML interests
llm coding
Recent Activity
updated a Space 19 days ago
KaiWu123/awesome-ai4ai updated a dataset 19 days ago
KaiWu123/awesome-ai4ai repliedto their post 20 days ago
Can AI improve AI? We read 223 papers to find out.
The literature is fragmented ā long-horizon agents, AI4AI, self-improvement, and RSI are four separate conversations that don't cite each other. So we organized all of it around one question: how far can an AI system reliably carry an improvement process from idea to verified result?
Two findings:
1. AI is now excellent at the *work* of improvement ā planning, coding, experimentation, repair. Humans still set the goals and decide what counts as progress. That asymmetry isn't closing.
2. The composition gap: strong performance on individual components rarely composes into reliable end-to-end improvement. A system can beat every component benchmark and still fail the full loop. Most "self-improving AI" claims live in this gap.
Search the catalog: https://huggingface.co/spaces/KaiWu123/awesome-ai4ai
Dataset: https://huggingface.co/datasets/KaiWu123/awesome-ai4ai
š Nice Layout Paper: https://github.com/KaiWU5/Awesome-AI4AI/blob/main/assets/AI4AI-Survey.pdf
š Site: https://kaiwu5.github.io/Awesome-AI4AI/
ā Repo: https://github.com/KaiWU5/Awesome-AI4AI
š Paper: https://www.preprints.org/manuscript/202608.2108/v1
Disagreement welcome ā especially on the composition gap.Organizations
None yet
updated a
Space 19 days ago
updated a
dataset 19 days ago
replied to their post 20 days ago
AI4AI ā AI4Human. lol. Now I just need it to do the dishes. š
reacted to their post with š„š 21 days ago
Post
2208
Can AI improve AI? We read 223 papers to find out.
The literature is fragmented ā long-horizon agents, AI4AI, self-improvement, and RSI are four separate conversations that don't cite each other. So we organized all of it around one question: how far can an AI system reliably carry an improvement process from idea to verified result?
Two findings:
1. AI is now excellent at the *work* of improvement ā planning, coding, experimentation, repair. Humans still set the goals and decide what counts as progress. That asymmetry isn't closing.
2. The composition gap: strong performance on individual components rarely composes into reliable end-to-end improvement. A system can beat every component benchmark and still fail the full loop. Most "self-improving AI" claims live in this gap.
Search the catalog: KaiWu123/awesome-ai4ai
Dataset: KaiWu123/awesome-ai4ai
š Nice Layout Paper: https://github.com/KaiWU5/Awesome-AI4AI/blob/main/assets/AI4AI-Survey.pdf
š Site: https://kaiwu5.github.io/Awesome-AI4AI/
ā Repo: https://github.com/KaiWU5/Awesome-AI4AI
š Paper: https://www.preprints.org/manuscript/202608.2108/v1
Disagreement welcome ā especially on the composition gap.
The literature is fragmented ā long-horizon agents, AI4AI, self-improvement, and RSI are four separate conversations that don't cite each other. So we organized all of it around one question: how far can an AI system reliably carry an improvement process from idea to verified result?
Two findings:
1. AI is now excellent at the *work* of improvement ā planning, coding, experimentation, repair. Humans still set the goals and decide what counts as progress. That asymmetry isn't closing.
2. The composition gap: strong performance on individual components rarely composes into reliable end-to-end improvement. A system can beat every component benchmark and still fail the full loop. Most "self-improving AI" claims live in this gap.
Search the catalog: KaiWu123/awesome-ai4ai
Dataset: KaiWu123/awesome-ai4ai
š Nice Layout Paper: https://github.com/KaiWU5/Awesome-AI4AI/blob/main/assets/AI4AI-Survey.pdf
š Site: https://kaiwu5.github.io/Awesome-AI4AI/
ā Repo: https://github.com/KaiWU5/Awesome-AI4AI
š Paper: https://www.preprints.org/manuscript/202608.2108/v1
Disagreement welcome ā especially on the composition gap.
- 2 replies
posted an update 21 days ago
Post
2208
Can AI improve AI? We read 223 papers to find out.
The literature is fragmented ā long-horizon agents, AI4AI, self-improvement, and RSI are four separate conversations that don't cite each other. So we organized all of it around one question: how far can an AI system reliably carry an improvement process from idea to verified result?
Two findings:
1. AI is now excellent at the *work* of improvement ā planning, coding, experimentation, repair. Humans still set the goals and decide what counts as progress. That asymmetry isn't closing.
2. The composition gap: strong performance on individual components rarely composes into reliable end-to-end improvement. A system can beat every component benchmark and still fail the full loop. Most "self-improving AI" claims live in this gap.
Search the catalog: KaiWu123/awesome-ai4ai
Dataset: KaiWu123/awesome-ai4ai
š Nice Layout Paper: https://github.com/KaiWU5/Awesome-AI4AI/blob/main/assets/AI4AI-Survey.pdf
š Site: https://kaiwu5.github.io/Awesome-AI4AI/
ā Repo: https://github.com/KaiWU5/Awesome-AI4AI
š Paper: https://www.preprints.org/manuscript/202608.2108/v1
Disagreement welcome ā especially on the composition gap.
The literature is fragmented ā long-horizon agents, AI4AI, self-improvement, and RSI are four separate conversations that don't cite each other. So we organized all of it around one question: how far can an AI system reliably carry an improvement process from idea to verified result?
Two findings:
1. AI is now excellent at the *work* of improvement ā planning, coding, experimentation, repair. Humans still set the goals and decide what counts as progress. That asymmetry isn't closing.
2. The composition gap: strong performance on individual components rarely composes into reliable end-to-end improvement. A system can beat every component benchmark and still fail the full loop. Most "self-improving AI" claims live in this gap.
Search the catalog: KaiWu123/awesome-ai4ai
Dataset: KaiWu123/awesome-ai4ai
š Nice Layout Paper: https://github.com/KaiWU5/Awesome-AI4AI/blob/main/assets/AI4AI-Survey.pdf
š Site: https://kaiwu5.github.io/Awesome-AI4AI/
ā Repo: https://github.com/KaiWU5/Awesome-AI4AI
š Paper: https://www.preprints.org/manuscript/202608.2108/v1
Disagreement welcome ā especially on the composition gap.
- 2 replies
published a
Space 21 days ago
published a
dataset 21 days ago
commented a
paper 6 months ago
commented 3
papers 7 months ago
MedXIAOHE: A Comprehensive Recipe for Building Medical MLLMs
Paper ā¢ 2602.12705 ā¢ Published ā¢ 64 ā¢
13
MedXIAOHE: A Comprehensive Recipe for Building Medical MLLMs
Paper ā¢ 2602.12705 ā¢ Published ā¢ 64 ā¢
13
MedXIAOHE: A Comprehensive Recipe for Building Medical MLLMs
Paper ā¢ 2602.12705 ā¢ Published ā¢ 64 ā¢
13
upvoted a paper 7 months ago
submitted a
paper to Daily Papers 7 months ago
authored a
paper 12 months ago
upvoted a paper 12 months ago
liked a
dataset over 1 year ago
updated a
model about 2 years ago
liked a
dataset over 2 years ago
liked a
dataset almost 3 years ago