Madrona Venture Labs’ Post

It's been a while since our last Rapid Fire, so we've doubled the fun by featuring two AI leaders this time. Featuring Joshua Starmer PhD, Founder and CEO of StatQuest and Susan Walsh, Founder of The Classification Guru Ltd, this session of Analytics Vidhya's Rapid Fire with AI Leaders is surely going to bring you double the insights! Join us for a lively conversation where we get rapid answers on topics such as: - Uses of LLMs and AI - Challenges in Data Cleaning - Transformers, Claude 3.5 sonnet, and more. 🎬 Watch the video and share your thoughts! #GenAI #RapidFireWithAILeaders #AnalyticsVidhya #DataClassification #StatQuest #LLMs #BAM

We asked Sandra Kublik: how do you see the evolution of RAG-based models, and where does Cohere stand within that context? TL;DW ↓ → Cohere is focused on solving big enterprise pain points, not chasing the AGI. Especially building scalable models. → The recently released Command R+ model is big enough to be competitive in terms of reasoning capabilities but small enough to be very cost-effective. → Although Cohere mostly covers text modality use cases, multimodal paths are also in their interests. — (link to the full interview in the comments) #generativeai #genai #llm #rag

Hints of human like #thought processes? Day 1 of #AGI? The ability to create #worldmodels? Ignacio de Gregorio Noblejas highlights a #development that might be day 1 in #AGI. It's about building better, smarter, and powerful #foundationalmodels. Imagine #foundationalmodels getting the ability to build #worldmodels, use #prioritization, create #mental short cuts & #heuristics, along with a level of #dynamic #self-refinement. Read on to learn more. Join the conversation and leave your comments below. Also for more AI insight like this follow Ignacio de Gregorio Noblejas

#FDAcleared #AIML list is still being processed. Here are the reasons we're still double-checking our revised findings. We also want to figure out how to capture the fast growing #ClinicalAI vendors who received a few clearances after July, like Qure.ai and Annalise.ai. (Congrats!) For those of you who know John Kalafut PhD, you understand why reason number 8 is hilarious. We're still considering having AI rewrite this for a 5th grader. Thoughts? We're also work-shopping some clever names for the article. "Cool Kids in Clinical AI" seems to be the favorite. Thoughts?

#DataOps is a real challenge and ever more important in the age of #LLMs and #generativeAI. Learn more about core capabilities in this article

🎉 Excited to share that GPT Root Cause Analysis V1.2 has now reached over 500 users! 🎉 A huge thank you to everyone who has tried it out and provided valuable feedback. 🙏 https://lnkd.in/ggY9ytB3 I also wanted to highlight another great tool for problem-solving: the Fishbone Diagram Creator. While I wasn't involved in its development, I highly recommend checking it out. It offers a unique and insightful approach to identifying root causes. Let's continue to explore innovative ways to tackle challenges together! #problemsolving #TRIZ #generativeAI #GPTStore #rootcauseanalysis #fishbonediagram #innovation

Here's a nice perspective piece from InfoWorld on combining large generalist LLMs and fine-tuned specialist LLMs for various #genAI applications, and the emerging use of "counsels of specialists" (i.e., SMoE models) to improve the accuracy, speed, and efficiency of #generativeAI systems. https://lnkd.in/gEN-WJ6w Author: Jignesh Patel Image: DALL-E #transformation #strategyconsulting

Heres an article on the pros & cons of using #GenAI in Call Centers. https://lnkd.in/gfTr-Q-D What is your thought about the author's insights. Share your feedback

We recently came across Swimm's comprehensive 2024 survey on code knowledge sharing and found it incredibly relevant to our work and discussions with devtool companies. We've summarized the key findings for you here, highlighting the significant impact of effective knowledge sharing on productivity and the growing role of AI tools in this area. Check out our summary here ⬇️ and find a link to the full survey in the comments (you just need to register for it). 🔗https://lnkd.in/d-MxXH-P #Develocity #SwimmSurvey #DeveloperTools #AICoding #Productivity #TechInnovation

tl;dr RAG is *not* pain-free, especially when moving from prototype to prod! Image: https://lnkd.in/dQ33g-3S - adapted from Barnett et al. (2024) References: + (Barnett et al., 2024) Seven Failure Points When Engineering a Retrieval Augmented Generation System - https://lnkd.in/d9_cNJsj PS: in case you're wondering, I found out about this post+article from a survey talk by LlamaIndex co-founder for the #MLOps community AI in Production event (https://lnkd.in/d-B8gzw6)