Introducing SVG Arena! SVGs are XML-based text files that... encode geometry, and they're surprisingly tough for state-of-the-art LLMs to get right Here's how bad models (still) are at drawing a pelican on a bicycle - creds to Simon Willison Leaderboard coming soon! In the meantime, here are some recent generationsshow more

Grok models are highly advanced at search and retrieval,... securing 2 of the top 5 spots on the official Search Arena leaderboardshow more

Major AI breakthrough: Diffusion Large Language Models are here!... They're 10x faster and 10x cheaper than traditional LLMs. Here's everything you need to know:show more

Grok models are incredibly advanced at search and now... hold two spots among the top five on the Search Arena leaderboardshow more

The Toothless model is almost done. I'm also working... on the Stormfly model at the moment to get them both ready. They will both be released hopefully soon for download. In the meantime have a render of the Toothless model based on @/_SaltySerpent art.show more

Bits are coming soon to Magic Eden on Bitcoin... 🟧 Here's some alpha:show more

2025 was a landmark year for local machine learning.... We're looking back at how Transformers.js and WebGPU enabled state-of-the-art models right in the browser. See how to build the future of the web with @HuggingFace's Xenova →show more

Beta invites for Bulma v1 are out! Hoping to... launch it officially this spring. In the meantime, here's a preview of the new website:show more

Islamabad’s first Pelican pedestrian crossings are now in place... on Constitution Avenue, right in front of the Secretariat. More such initiatives coming soon by Capital Development Authority - CDA, Islamabadshow more

In the constant effort to give you all better... communication and visibility on XDefiant, we are introducing a new effort called Community Loop. In advance of the Y1S1.5 patch, here are some of the changes you can expect to drop next in XDefiant.show more

Look to the stars! The cosmos are vast but... the fight’s right here in the Stellar Planetarium arena, coming in TFT Patch 17.5.show more

Happy Thanksgiving! Totems are a new switchable grid coming... soon. Totems can retract and extend from the floor at will from switches, making for some tricky puzzles and clever obstacles! Also, Puffballs are new fungal foes that get projectiles stuck right onto them.show more

"Imagine what the Iranians are going through now, those... who are supposed to get bombed. They're in suspense, wondering if the Americans are coming and what kind of bombs will fall... It's good that they're feeling that way." Shai Levi, military correspondent for Mako (Channel 12) enjoys the fear of ordinary Iranians following the recent insane statements by Trump. Aired about 50 minutes ago.show more

“There are only older palace women here” It hurts... to think about the reason why so many of them are still here.. We’re getting closer and closer to the episodes focusing on the former emperor and it’s gonna be a tough watchshow more

D2 is a diagram scripting language that turns plain... text into architecture diagrams. - CLI tool that renders `.d2` files to SVGs with live-reload on save. - Supports multiple layout engines like ELK and customizable themes. - Exports to SVG, PNG, PDF, and other formats. - Includes a web playground and language tooling for editors.show more

Drawing to 3d and animation. Here's how I did... this: @Vizcom_ai to make a more "rendered" version of the drawing. Leonardo ai image to image for the "3d" pass. Magnific ai for upscale/detail. Runway Gen-2 for animation. And yes, the soul of the drawing is kinda lost in translation 😬, but imagine once we're able to get that across.. I'm very impressed with how well the process carries over the various elements of the drawing. #ai #artshow more

Cruz: If you are a cop in California, an... electrician in California, it's really tough to make ends meet and there's a reason the people of California are fleeing. They are coming to Texas.show more

A wet and windy day ahead for some to... start the weekend Here are some of the details in the #4cast 👇show more

We’re excited to introduce Text-to-LoRA: a Hypernetwork that generates... task-specific LLM adapters (LoRAs) based on a text description of the task. Catch our presentation at #ICML2025! Paper: Code: Biological systems are capable of rapid adaptation, given limited sensory cues. For example, our human visual system can quickly adapt and tune its light sensitivity to our surroundings. While modern LLMs exhibit a wide variety of capabilities and knowledge, they remain rigid when adding task-specific capabilities. Traditionally, customizing these models requires gathering large datasets and performing often expensive, time-consuming fine-tuning for specific applications. To bypass these limitations, Text-to-LoRA (T2L) meta-learns a “hypernetwork” that takes in a text description of a desired task, as a prompt, and generates a task-specific LoRA that performs well on the task. In our experiments, we show that T2L can encode hundreds of existing LoRA adapters. While the compression is lossy, T2L maintains the performance of task-specifically tuned LoRA adapters. We also show that T2L can even generalize to unseen tasks given a natural language description of the tasks. Importantly, Text-to-LoRA is parameter-efficient. It generates LoRAs in a single, inexpensive step, based solely on a simple text description of the task. This approach is a step towards dramatically lowering the technical and computational barriers, allowing non-technical users to specialize foundation models using plain language, rather than needing deep technical expertise or large compute resources.show more