-Cantīde- is Healing 🧘‍♂️☯️🙏 direct preference optimization! arxiv.org/abs/2305.18290

📢New research on mechanistic architecture design and scaling laws. - We perform the largest scaling laws analysis (500+ models, up to 7B) of beyond Transformer architectures to date - For the first time, we show that architecture performance on a set of isolated token

Today, AlphaFold 3 came to life. A monumental leap toward the world of biomolecules. Predicting structures across proteins, nucleic acids, and small molecules. The world just changed - let me try to break it down the best I can. 🧵OPEN THE THREAD🧵

When you remove a line of code that looked useless

I used to think I was rational. Then I read Daniel Kahneman's Nobel Prize-winning work on human decision-making. He routinely asks 8 questions to expose cognitive traps you fall into daily. Test yourself with these questions (it's the ultimate BS detector for your brain):

Apple just released Depth Pro Sharp Monocular Metric Depth in Less Than a Second

app: huggingface.co/spaces/akhaliq…

Super interesting model just released by Apple: DepthPro, a monocular, fast, precise metric depth estimation model. It predicts absolute (metric) distance with no additional data. * Weights: huggingface.co/apple/DepthPro * Paper: huggingface.co/papers/2410.02… * Demo: hf.co/spaces/akhaliq…

While everyone was busy with ECCV, DepthPro, a foundation model for zero-shot metric monocular depth estimation in less than a second, has been released. Code and weights are up: arxiv.org/abs/2410.02073

Doesn't look like the super low-res depth maps from the iPhone Pro's LiDAR sensor would be too useful as supervision (or sanity check) for recent monocular depth estimators.

This could possibly be the greatest feat of engineering in human history

How Rotary Position Embedding Supercharges Modern LLMs New video! 🥳 youtube.com/watch?v=SMBkIm… Topics: 1⃣ Why is attention permutation equivalent 2⃣ Earlier attempts of using sinusoidal positional encoding 3⃣ Rotary Position Embedding (RoPE) for encoding relative position 4⃣

Very cool (and almost *frustrating*) optical illusion. If you pause this video at any time, the shapes disappear. They exist *only* temporally, not visually. This is from Branta Games on YouTube (full video linked below)

Excited to share our new pre-print arxiv.org/pdf/2502.01600 We train a digital agent that solves diverse day-to-day tasks from the AppWorld benchmark by interacting with its stateful environment using API calls. AppWorld is hard! The previous best open-weight agent (Llama 3 70B)

Sharp Monocular View Synthesis in Less Than a Second Lars Mescheder, Wei Dong, Shiwei Li, Xuyang Bai, Marcel Santos, Peiyun Hu, Bruno Lecouat, Mingmin Zhen, Amaël Delaunoy, Tian Fang, Yanghai Tsin, Stephan R. Richter, Vladlen Koltun tl;dr:feedforward 3DGS arxiv.org/abs/2512.10685

New paper from Apple - Sharp Monocular View Synthesis in Less than a Second Mescheder et al. @ Apple just released a very impressive paper (congrats! 🎉🥳). You give it an image and it generates a really great looking 3d Gaussian representation. Uses depth pro. It's really good.

People are under-estimating Apple in AI!

People are underestimating Apple in AI. I just ran Apple’s new SHARP model locally and watched my photos turn into 3D Gaussian splats in seconds, then stepped inside them on Vision Pro. This feels like the beginning of something special. You really have to try it.

Apple's ml-sharp is literal sorcery oh my god

favorite 3D generations from prompts in the past day. open sourced/free to create your own using Nano Banana + Apple SHARP-ML