- Owns VSCode - Owns GitHub - Owns OpenAI (basically) - Owns npm - Pioneered in-editor AI with a 3-year lead - Drops the ball on editor autocompletion to indies Is Copilot ok

If your ai product doesn’t need evals. It’s probably fundamentally uninteresting and not useful.

I used runme and Simon Willison 's llm tool to help me unsubscribe from campaign emails. The AI in runme generated the CLI commands to accomplish the task. This used the llm tool to classify emails and extract the unsubscribe link.

I started a company! Introducing Melty, an open source AI code editor for 10x engineers. It's the first editor that understands what you’re doing from the terminal to GitHub and collaborates with you to write production-ready code.

Today Hugging Face accelerate 0.34.0 is now out, and it is a packed release! From `torchpippy` updates to resumable dataloader support, and revamped TransformerEngine support, there's a ton to cover 🧵

Inspired by this tweet… Tell me your environmental productivity hacks that use LLMs. How are you making your digital life a little better with a dash of LLM?

now that the weekend is over I must stress how extremely lame it is that bedrock doesn't supported quantized custom models just call it "claude but in aws" because there's no point in allowing the use of non-quantized custom models lmao

Seeing more “RAG is more effective than fine tuning” posts. It’s like saying dieting is more effective than exercise so don’t exercise. You can do both. In fact, if you are trying to optimize one in isolation it doesn’t make any sense. Your fine tuning data should have

i’ve been muting folks who consistently interpret others uncharitably and wow my feed’s in a better place ☀️

My GitHub Pages is stuck in this hell. I'm really close to migrating everything away. Have had a lot of hiccups lately

tiobe.com/tiobe-index/

Excited to announce that Unsloth AI is part of @YCombinator S24! My brother and I relocated to SF from Sydney, and it's been super hectic! Waymos, AI ads, and the people are the best! We’re so pumped to bring fine-tuning to an even larger audience and showcase how easy it can

RAG is increasingly going multi-modal, but document retrieval is tough, and layout gets in your way. But it shouldn't! Introducing 🪤RAGatouille's Vision-equipped, ColPali-powered sibling: 🐭Byaldi With just a few lines of code, search through documents, with no pre-processing.

Very excited to welcome Hamel Husain to the Answer.AI team! For his first day, we did a dev chat together. :D Here's our complete walkthru of an idiomatic FastHTML CRUD app: youtu.be/kfEpk6njb4s