Little Bear Labs

Looking forward to finding out what sessions wind up on the list for The Linux Foundation conference. Hoping to see some incisive discussion about the future of #AI. Anyone planning to submit a proposal — or have topics you’d like to see represented?  

Everyone needs a portfolio. (Yes, even developers. Bear with me.) 👇 The other day I searched for myself on Perplexity and I found something interesting… My phone number. Also, apparently I am a real estate agent. ¯\_(ツ)_/¯ Having some kind of digital footprint is inevitable, but unless you’re the one creating content, you’re at the mercy of whatever the scrapers can find on you 😹. If you don’t create your own portfolio website, your identity is made up of the detritus left by all your online activity, and often scrambled up with people who share your name. If you want to control your individual professional identity, having a website, newsletter — something that you *own* — is increasingly important. But a lot of people in tech don’t create that digital identity; maybe they don’t see why it matters, or maybe it doesn’t feel feasible. Before the engineers reading this run away... okay, you might be picturing a website full of code screenshots — which mean nothing to anyone other than other devs. But your work creates something real in the world, and without a portfolio, it’s tough for people to understand the what, why, and how-awesome of it. Realizing you need a portfolio website is one thing; making one is another. It’s hard, time consuming, and difficult to get the design right. And unless you’re a content strategist in your spare time, organizing the information is itself pretty daunting. That sounds like a lot of problems — and it is. But I’m excited about the solution — which I’ve been having a lot of fun working on.  

Really fun and informative live demo of the Mosaic AI gateway from Kasey Uhlenhuth! (And cookies…. yum 🍪) https://lnkd.in/grWvBaVY (demo at 0:58-1:14). This was just one of the zillion cool things at the Databricks Data +AI Summit 2024. Looking forward to watching more of the on-demand sessions! #DataAISummit #AI #GenerativeAI 

Have been seeing a lot of good use cases for RAG lately. Nice bit of perspective from Stack Overflow on the whys and hows of a good chunking strategy! #LLMs "#AI #GenerativeAI

Ever get halfway into a conversation with someone and realize you're talking about two totally different things? I was gushing about Groq the other day, and it took several minutes for us to realize I mean the fastest LLM inference platform & chip (not the xAI chatbot). I don’t have much to say about the latter at the moment, but I have a *lot* of enthusiasm for the former...

*Yikes* Scalability is one of the biggest challenges in using generative AI at the enterprise level. Great article from Mozilla on just one of the many (and often hard-to-predict) risks. 

Fascinating look from Stack Overflow at how devs *really* feel about #AI coding assistants. 

Really looking forward to this convo! Open source and AI are two of our favorite topics—and the question of how they intersect is more pressing than ever. #AI #OpenSource #OpenSourceFriday GitHub

Musho AI just keeps getting faster! 🚀

Exploring various providers for open source LLM models. A few that impressed: Fireworks AI - I had a feeling this would be big. With the recent announcement they power Stable Diffusion 3 and same day access to things like Mistral's new model it's clear they are. Lepton AI - A more recent discovery of mine. They add function calling to all their open source models with Structural Decoding. Anyscale - Very easy to use fine tuning UI with same format as openai