The Daily Index — Thursday, March 13, 2025 · TLDR Index

HEADLINES & LAUNCHES

TLDR AI

TLDR AI

AI SCIENTIST'S FIRST PUBLICATION

The AI scientist from Sakana Labs has its first reviewed and accepted publication. It generated the idea, conducted experiments, and wrote a novel paper that was accepted to an ICLR workshop (with the full consent of the conference organizers).

sakana.ai · 8 min

TLDR AI

GOOGLE'S NEW OPEN WEIGHTS MODEL

Google has released the weights and a technical report for the Gemma 3 model. This model, which comes in four sizes, is approximately as performant as Gemini 1.5 Pro. It also has a strong understanding of 140+ languages. It seems to be close to the state-of-the-art dense model.

blog.google · 7 min

TLDR AI

GEMMA 3 TECHNICAL REPORT

Google has released the best dense models available. They score 9th in the arena, have phenomenal multilingual capabilities, and support more than 140 languages. Depending on the upcoming open launches, these may be the backbone of the community. They certainly are the best single GPU models and likely will be for some time.

storage.googleapis.com · 12 min

TLDR AI

PREDICTING FUTURE FEATURES IN DIFFUSION MODELS

TaylorSeer introduces a method to predict future timestep features in diffusion models using Taylor series expansion, significantly reducing errors in feature caching.

arxiv.org · 16 min

TLDR AI

SUPER-RESOLUTION WITH TOKEN AGGREGATION

CATANet enhances image super-resolution by aggregating long-range content-similar tokens.

arxiv.org · 12 min

TLDR AI

MULTIMODAL MOTION GENERATION

Motion Anything improves conditional motion generation by introducing Attention-based Mask Modeling for better spatial and temporal control.

steve-zeyu-zhang.github.io · 2 min

TLDR AI

DUAL-STREAM VIDEO INPAINTING

VideoPainter introduces a dual-stream architecture for video inpainting, significantly reducing learning complexity while improving background preservation and object generation. It also provides VPData and VPBench, the largest video inpainting dataset and benchmark to date.

yxbian23.github.io · 1 min

TLDR AI

OPENR1 UPDATE 3

The Hugging Face team has announced the next update for its open replication of the DeepSeek reasoning model. It found that small models can outperform the larger ones when tuned specifically for competitive coding.

huggingface.co · 8 min

TLDR AI

ROTATE LORA FOR WAN (HUGGING FACE HUB)

Wan is a generative image model that has recently been open sourced.

huggingface.co

TLDR AI

FLAT COLOR LORA (HUGGING FACE HUB)

Another LoRA for the Wan video model.

huggingface.co

TLDR AI

SMALL REASONING MODELS WITH TWO-STAGE RULE-BASED RL

By breaking the reasoning training process into two steps, this work shows that small models are still able to generalize across tasks and even modalities.

forjadeforest.github.io · 6 min