Ep-1 - Training LLMs with Unsloth.ai Co-Founder
What is RAG, explain GPUs like I'm five & more
What a week (I'll be saying this every week..) Here's the best bits.
The very first episode of Unsupervised Learning has just dropped, and it’s a conversation with Unsloth.ai Co-Founder Daniel
Or check out Simplecast
Resources and Cool Things
Awesome Local AI- This is an open source repo of all local ai projects, curated from friends at Jan. From inferences to open models- contribute or browse.
Folks I’m Reading
Charlie Guo at Ignorance AI- Charlie doesn’t hedge, his newsletter aims to reduce the AI overwhelm and he does so with a level-headedness that I appreciate in my inbox.
Devanash’s AI Simplified- Devanash is another who I’ve come to value as a voice of reason in the space on both Linkedin and Substack (Q&A coming soon!🎉)
Cobus Greyling- want your research papers in visual format? Cobus does this like no one else.
What we’ve learned
This week, we talked about what goes into an LLM, specifically word vectors.
We also touched on the power of RAG (retrieval augmented generation) and why it’s important for mitigating hallucinations.
Enable 3rd party cookies or use another browser
This incredible article from Ian Kelk (video included)
Around the globe 🌎
Episodes coming soon-
Developer empathy- with Noah (Host of NVIDIA AI podcast)
What we can learn from llm leaderboards- with Shriyash (Martian Co-founder)
Defining autonomy for SMB’s with AI- with João (CrewAI founder)
”How to be open”- with James (Posthog CEO)
Thanks for spending time learning about AI with me!
As always, feedback is a gift.
renee@unsupervisedlearning.co
Or course the simplecast upload failed - thanks for letting me know folks! One moment