The Daily Index — Monday, March 17, 2025 · TLDR Index

HEADLINES & LAUNCHES

TLDR AI

TLDR AI

GOOGLE'S RESPONSE TO U.S. AI POLICY

Google outlines its vision for U.S. AI policy, advocating for investments in AI infrastructure, streamlined government adoption, and international pro-innovation standards to maintain leadership in the AI space.

blog.google · 6 min

TLDR AI

SAKANA CLAIMS ITS AI-GENERATED PAPER PASSED PEER REVIEW

Japanese AI startup Sakana used its AI Scientist-v2 to generate a paper accepted at an ICLR workshop but withdrew its post-initial review due to transparency concerns. Though Sakana highlights AI's potential, critics argue that human judgment was crucial and acceptance was facilitated by higher workshop acceptance rates. Skeptics remain wary, emphasizing AI's technical flaws and the need for established norms to prevent degrading the scientific review process.

techcrunch.com · 5 min

TLDR AI

TRANSFORMERS WITHOUT NORMALIZATION

Most Transformer models are normalized with a layer norm. This can be seen as equivalent to normalizing the model to a sphere. However, if you use carefully placed tanh functions, you can eliminate layer norm while maintaining stability. This is now equivalent to normalizing to a hypercube.

arxiv.org · 22 min

TLDR AI

SCALING LAWS FOR DILOCO

DeepMind has released a paper detailing scaling laws for the powerful cross data center training algorithm DiLoCo. These show how stable model training can be even with gradients synchronized across continents.

arxiv.org · 35 min

TLDR AI

SPEAKER IDENTIFICATION WITH WHISPER

WSI repurposes the Whisper ASR encoder for multilingual speaker identification using joint loss optimization. It outperforms Pyannote, ECAPA TDNN, and Xvector in recognizing speakers across diverse languages and environments.

arxiv.org · 9 min

TLDR AI

LUMA'S NEW PRE-TRAINING METHOD FOR MULTI-MODAL MODELS

Luma Chief Scientist Jiaming Song, who created the earliest accelerated algorithm for diffusion models, has unveiled Inductive Moment Matching (IMM)—a new multi-modal pre-training method that surpasses diffusion models with superior sample quality and 10x greater efficiency.

lumalabs.ai · 5 min

TLDR AI

THUNDERKITTENS ON BLACKWELL

ThunderKittens is a powerful and simple abstraction for writing efficient and performance CUDA kernels. This post explores how to use the framework for the newest Blackwell series of GPUs from Nvidia. The main difference is thinking in terms of data flow.

hazyresearch.stanford.edu · 17 min

TLDR AI

ARE OPEN MODELS REALLY OPEN?

Many AI models labeled as "open" come with restrictive licensing terms. Google's new Gemma 3 models and similar releases from Meta raise concerns about commercial limitations, potentially impacting smaller companies relying on these technologies.

techcrunch.com · 4 min

TLDR AI

NO PART OF AMAZON IS 'UNAFFECTED' BY AI, SAYS ITS HEAD OF AGI

Amazon's VP of Artificial General Intelligence, Vishal Sharma, has confirmed AI's integration across AWS, robotics, and Alexa, highlighting the company's extensive AI deployment.

techcrunch.com · 4 min

TLDR AI

PANORAMA CHALLENGE FOR CANCER DETECTION

The PANORAMA study is an international initiative evaluating AI models and radiologists in detecting pancreatic cancer in CECT scans.

links.tldrnewsletter.com · 2 min

TLDR AI

AI FIRMS FOLLOW DEEPSEEK'S LEAD, CREATE CHEAPER MODELS WITH “DISTILLATION”

Leading AI firms like OpenAI, Microsoft, and Meta are using "distillation" to create more cost-effective models by training smaller systems with a "teacher" LLM.

arstechnica.com · 5 min