Ep-1 - Training LLMs with Unsloth.ai Co-Founder

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

Thanks for reading 🤘 (un)supervised learning! Subscribe for free to receive new posts and support my work.

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.

@unsupervisedlearningpodI like to imagine hallucinating large language models as cheeky Aussie schoolkids.  Are you building on open source AI project and fact a chat? link in bio 🤌 ps- check out withmartian.com for their LLM inference provider leaderboard for agnostic reporting on the best llms for your use case. not spon - see their GH #llm #deepmind #mistral #ai #openai

Tiktok failed to load.

Enable 3rd party cookies or use another browser

This incredible article from Ian Kelk (video included)

Around the globe 🌎

FOSDEM 2024 event

State of OS 2024 (UK)

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

Thanks for reading 🤘 (un)supervised learning! Subscribe for free to receive new posts and support my work.