They solved AI’s memory problem!

Attention Residuals by Kimi AI. Adaptive, continuous learning AI models. #ai #ainews #llm #airesearch #agi

Thanks to our sponsor Wondercraft. Use my code AI-SEARCH to get $25 OFF! https://www.wondercraft.ai/?via=ref

Original paper: https://arxiv.org/abs/2603.15031
Transformers explainer: https://youtu.be/U2hZFMVNSE0

0:00 Intro
0:27 AI’s amnesia problem
1:50 Design of deep AI models
3:19 Residual connections
6:50 The genius of current language models
9:03 Applying attention to residuals
13:05 Wondercraft
15:22 Infra problems
17:46 Compute results
18:50 Performance results
20:45 Wider or deeper
22:22 From static to adaptive

Newsletter: https://aisearch.substack.com/
Find AI tools & jobs: https://ai-search.io/
Support: https://ko-fi.com/aisearch

Here's my equipment, in case you're wondering:
Lenovo Thinkbook: https://amzn.to/4jWeKwH
Dell Precision 5690: https://www.dell.com/en-us/dt/ai-technologies/index.htm?utm_source=AISearchTools&utm_medium=youtube&utm_campaign=precisionai#tab0=0
GPU: Nvidia RTX 5000 Ada https://nvda.ws/3zfqGqS
Mic: Shure SM7B https://amzn.to/3DErjt1
Audio interface: Scarlett Solo https://amzn.to/3qELMeu