Fivetran's Iceberg data lake implementation is basically a headless cloud data warehouse storage engine. We're looking for a principle engineer to lead development of it. Super cool opportunity for someone who loves DBMS! Open in many locations but here is the CA link:…

Thanks to Emil Eifrem, Bob van Luijt and others for talking about the future of graph databases and knowledge graph traversal in RAG pipelines! Lots more to talk about here but if you’re a company/dev experimenting with it would love to chat with you.

supervised.news/p/where-graph-…

I'm working with The New Stack to build out their contributor program. It's a fantastic site for programmers and technical writers, and I'd love to help to get you published. Read this post or DM! thenewstack.io/opinionated-yo…

On the one year-ish anniversary since the launch of Supervised sharing one of my first very important ideograms shared with Jon Russell and yashad

Just as a fun aside here, a year ago when I was thinking about starting Supervised I was thinking about *a lot* of things. One of them included going back and getting my masters in engineering.

I reached out on a whim to the academic department of a university I won't name just…

What an insane year it’s been

alright lets see what you got

Apple’s iOS 18 features they will highlight at WWDC are based on an in-house model. The talks with OpenAI and Google are for a chatbot/search component.

Quantized matmuls in the latest MLX are up to 40% faster thanks to Angelos Katharopoulos and Jagrit Digani

QLoRA fine-tuning Llama 3 70B on a single M2 Ultra stats:
- Batch size 4 with 16 LoRA layers
- 95 toks/sec
- Peak mem 41GB
- Avg power 120 W

I’ve never had an app crash on literally any of the PlayStations I’ve previously owned, and every OTHER app works and never crashes, and yet Disney+ crashes minimum once a day.

new model just dropped

does anyone else have a requirements.txt that you just keep adding to every time you run into a package you need and carry it over to literally every single env you work in

Llama-3 is now available for fine-tuning on Together AI! docs.together.ai/docs/fine-tuni…

Oh man I’m excited to play with this

Decided long ago when I launched to do my best to not spend my energy competing with places with way more resources or more specialized than myself so will let the experts do the teardowns and the Important Places get all the interviews. Tomorrow’s issue isn’t about llama 3 🙂.

🚀Exciting News: The Lance columnar format was a game-changer in efficiently managing AI/ML workloads. But hold onto your hats because Lance v2 is here, and it's going to blow your mind! 🤯

blog.lancedb.com/lance-v2/