Henry Saputra

145 Followers
61 Following
2K Posts

Working on intersection of system and machine learning

https://github.com/hsaputra

KernelSkill: A Multi-Agent Framework for GPU Kernel Optimization

#CUDA #LLM #Performance #Package

https://hgpu.org/?p=30665

KernelSkill: A Multi-Agent Framework for GPU Kernel Optimization

Improving GPU kernel efficiency is crucial for advancing AI systems. Recent work has explored leveraging large language models (LLMs) for GPU kernel generation and optimization. However, existing L…

hgpu.org
Breaking down complex systems, understanding the trade-offs, and making sense of how things actually work under the hood.
https://crackingwalnuts.com/about
CrackingWalnuts

A blog about system design, AI, data engineering, and deep technical insights.

CrackingWalnuts

AI Didn’t Simplify Software Engineering: It Just Made Bad Engineering Easier https://robenglander.com/writing/ai-did-not-simplify/

Look on the bright side in a few years there will be good demand for "artisanal" software engineers and specialized IT folks who can fix this AI slop generated code at 1000$ per hour.

AI Didn't Simplify Software Engineering: It Just Made Bad Engineering Easier

Digg’s open beta shuts down after just two months, blaming AI bot spam

https://www.theverge.com/tech/894803/digg-beta-shutdown-layoffs-ai

For anyone running a content driven site like a forum or blog, a new site or user generated content is now a bad idea. The users are long gone because these AI chatbots steal everything and keep users on their apps. So, I am not surprised they shut down. Even big newspapers are reducing staff, and everyone thinks LLMs will write news for them and I have the Taj Mahal to sell you ;)

Digg’s open beta shuts down after just two months, blaming AI bot spam

The Digg Beta relaunch lasted two months but now it’s shutting down again and laying off employees.

The Verge

AI agents running research on single-GPU nanochat training automatically

https://github.com/karpathy/autoresearch

GitHub - karpathy/autoresearch: AI agents running research on single-GPU nanochat training automatically

AI agents running research on single-GPU nanochat training automatically - karpathy/autoresearch

GitHub
Labor market impacts of AI: A new measure and early evidence \ Anthropic https://share.google/fhrN5JCUAGVtepwSv
Labor market impacts of AI: A new measure and early evidence

Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.

Gemini Embedding 2: Our first natively multimodal embedding model

https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-embedding-2/

Gemini Embedding 2: Our first natively multimodal embedding model

An overview of Gemini Embedding 2, our first fully multimodal embedding model that maps text, images, video, audio and documents into a single space.

Google

AlphaFold 3, Demystified: A Comprehensive Technical Breakdown of Its Architecture and Design

https://github.com/shenyichong/alphafold3-architecture-walkthrough/tree/main

After outages, Amazon to make senior engineers sign off on AI-assisted changes
AWS has suffered at least two incidents linked to the use of AI coding assistants.
https://arstechnica.com/ai/2026/03/after-outages-amazon-to-make-senior-engineers-sign-off-on-ai-assisted-changes/?utm_brand=arstechnica&utm_social-type=owned&utm_source=mastodon&utm_medium=social
Keep the Tokens Flowing: Lessons from 16 Open-Source RL Libraries
https://huggingface.co/blog/async-rl-training-landscape
Keep the Tokens Flowing: Lessons from 16 Open-Source RL Libraries

We’re on a journey to advance and democratize artificial intelligence through open source and open science.