Hacker News new | ask | show | jobs
user: OsamaJaber
created: 2025-10-19
karma: 227

submissions:

Compiles any HuggingFace model into a single persistent megakernel
2 points | 0 comments
Mega Kernels, Written by Agents
2 points | 0 comments
AutoMegaKernel: Compiling a LLM into a single CUDA kernel
3 points | 0 comments
AutoMegaKernel: Compile an LLM into one provably-correct CUDA megakernel
4 points | 0 comments
StreamIndex: Memory-bounded compressed sparse attention via streaming top-k
4 points | 0 comments
Show HN: AutoKernel, Auto GPU Kernel Optimization
2 points | 0 comments
DeepSeek V4's indexer dies at 65K. We got it to 1M on 6GB
5 points | 0 comments
AutoKernel: Autonomous GPU Kernel Optimization via Iterative Agent-Driven Search
4 points | 0 comments
0 points | 0 comments
DeepSeek V4's indexer OOMs at 65K context. We got it to 1M in 6G
8 points | 0 comments
Ouroboros: Dynamic Weight Generation for Recursive Transformers
2 points | 0 comments
0 points | 0 comments
0 points | 0 comments
0 points | 0 comments
0 points | 0 comments
Tide: Token-Informed Depth Execution for Per-Token Early Exit in LLM Inference
3 points | 1 comments
Own your AI. Optimized down to the kernel
1 points | 0 comments
0 points | 0 comments
Agents with "Hands"
7 points | 18 comments
Open-Source Agent Operating System
11 points | 3 comments
PicoLM: Run a 1B parameter LLM on a $10 board
4 points | 1 comments
The Floating Dock for Developers
2 points | 0 comments
0 points | 0 comments
0 points | 0 comments