Pico-Banana-400k

FAL Flashpack: High-throughput tensor loading for PyTorch

Krea Realtime 14B: an autoregressive real-time video model

Krea Realtime 14B: an open-source real-time video model

Docs for LLMs (Svelte)

OpenAI releases Sora 2

Stable Diffusion 3.5 Flash

PowerRetention: a drop-in replacement for FlashAttention in LLMs

Matthew Walker's "Why We Sleep" Is Riddled with Scientific and Factual Errors

How I Hire Programmers (2009)

Deleuzoguattarianism

Contributing to Complex Projects (2022)

My startup banking story (2023)

Krea announces a real-time video model

Fast frame-consistent video models as "world" models

Tell HN: any reasonably used DB will likely outlast the programs using it

Play with Generative Adversarial Networks (GANs) in the Browser

Data, objects, and how we're railroaded into poor design (2018)

I find journaling indispensable (2019)

Tell HN: Hacking AIs is similar to hacking humans

FLUX.1 Krea: post-trained text-to-image model from Black Forest Labs and Krea

Transformer Circuits: reverse-engineering transformers into graspable programs

Cartridges: Store long contexts in tiny caches with LLM self-study

Cartridges: Storing long contexts in tiny caches with self-study

Building Cursor

Seaweed APT2: a real-time, interactive, streaming video generation model

OmniPainter: Training-Free Stylized Text-to-Image Generation with Fast Inference

Text-to-LoRA: Hypernetwork that generates task-specific LLM adapters (LoRAs)

Good Writing

You can try Imagen 4 at Krea

Wacom drawing tablets track the name of every application you open (2020)