Jose M.

Meta is revealing their secrets 🤫 Ricardo Silveira Cabral, Director of Generative AI Engineering at Meta, will dive deeper into Llama 3 than ever before at AIAI Berlin on September 12. Llama 3 has been trained on two recently announced custom-built 24K GPU clusters on over 15T token of data….. A training dataset 7x larger than that used for Llama 2, including 4x more code. In the session, Ricardo will share more on how Llama 3 got to where it is and where it will go from here. Trust us, you don't want to miss this. Get your pass: https://lnkd.in/eMXmKT57 #llama3 #genai #berlin #ai

4

"Context is king!" In a recent insightful discussion with Edo Liberty, Founder and CEO of Pinecone, alongside Derrick Harris and Satish Talluri from Andreessen Horowitz, the importance of how vector databases and #RAG can deliver the most relevant context for AI products. At Perigon, we understand that to extract the maximum value from LLMs, providing the right context is essential. This isn’t just about recognizing the subject matter—which many conversational AI applications can manage—it’s also about ensuring temporal relevance. Without it, even the most sophisticated AI can miss the mark, delivering outdated or irrelevant information. Why does this matter? Because in a world where data is constantly evolving, the accuracy and relevance of the information provided by AI must be on par with what we would find if we did the exploration ourselves for it to be trustworthy. I invite you to experience this firsthand through Perigon’s chat feature on our homepage. Powered by our API, it not only captures subject matter accurately but also ensures the information is timely, tapping into real-time, dynamic data from across the web. Check it out and see the difference real context makes: www.goperigon.com #AI #ConversationalAI

13

1 Comment

🚀 Looking to streamline fine-tuning for models like Llama 3, Mistral, and Mixtral? Our latest blog shows how Anyscale simplifies LLMs fine-tuning with Ray: 🔑 Key takeaways: • Scalable fine-tuning with LLM-Forge • Model Deployment with RayLLM Read more: https://lnkd.in/gQwD4A2U

34

In 2019, John Carmack, one of the most brilliant software engineers of our generation, left his job as Oculus CTO to work on Artificial General Intelligence. Ilya Sutskever, the Chief Scientist at OpenAI who is widely considered as the technical mind behind their LLM breakthrough, sent Carmack a list of 30 research papers and said "If you really learn all of these, you’ll know 90% of what matters today." Here's the full list with the actual research papers (Shared in the first comment). Do you want to be a world-class AI scientist? You know what to do.

10

2 Comments

Can you type at 500 words per second? Guess not but Groq, a well-funded Silicon Valley startup, has been developing a novel processor architecture optimized for the matrix multiplication operations that are the computational heart of deep learning. Lets nerd talk: that’s 800 tokens per second on LLaMA 3. Groq's architecture is a significant departure from the designs used by Nvidia and other established chip makers. Instead of adapting general-purpose processors for AI, Groq has built its Tensor Streaming Processor to accelerate the specific computational patterns of deep learning. Here’s another trillion dollar valuation coming up at some point! In the picture below, it does look like the DeLorean so the llama is travelling at 88 mph #SpeedOfThought • #GrogInnovation • #AIProcessorRevolution • #DeepLearningTech • #TensorStreaming • #TechDisruption • #SiliconValleyStartup • #FutureOfAl • #MachineLearningSpeed • #TrillionDollarTech

3

1 Comment

NVIDIA just became the most valuable company in the world Watch Lucas Atkins from Arcee break down the future and financials of NVIDIA from an LLMs perspective. This particular LLM was continuously pretrained (CPT'd) by Arcee on ~50B tokens from LLama3-70B-base then merged by Charles Goddard with LLama3-70B-Instruct. A lot of folks wonder why we are so focused on CPT + Merge at Arcee. This video breaks down exactly why you should be domain adapting your LLMs - especially if you have proprietary text corpora to infer against. I.E. anyone who has a RAG index. https://lnkd.in/g4tqM-x2

45

4 Comments

Nile now has a native integration with LlamaIndex This makes it easy to build AI-native multi-tenant apps with Postgres and LlamaIndex. If you are building AI-native B2B apps with Postgres, this should make it easy. pip install llama-index-vector-stores-nile

4

This week Jam hosted 15 DevRel for an off-the-record dinner Here's 7 quick bites about what we learned about The Art of DevRel: 1. The best DevRels work closely with engineering and product. It’s not just about sharing what they're building with devs, it’s also about sharing dev feedback with product. When you get that feedback loop right, that’s how you get magical developer experiences. 2. DevRel should help the entire engineering team do DevRel Like how HR enables any manager to hire, DevRel should enable any engineer to write blogs, create tutorials, give talks, etc. Swyx’s “DevRel as a Service” post was big topic of discussion— will link it in the comments. 3. No 2 companies do devrel the same. No 2 people do devrel the same way There is a lot of art here that’s not science. It’s about finding the right fit person for your exact company Like how there is founder market fit, there is a devrel company fit too 4. Reporting up was a huge point of discussion The best devrels over-communicate internally in the company what is happening on their side Communication creates trust, and trust creates freedom for devrel to do things that don’t scale that builds real community (and results). 5. DevRel is always on; you are always representing the company Community never takes time off, devs around the world are building nights, weekends and holidays You have to love helping people build - truly - to love devrel (and it’s so inspiring to meet people who do) 6. The best DevRel build genuine relationships They become friends with community members and devs The best community and relationships aren’t measurable But can have immeasurable impact on the trajectory of a product/company 7. Last one: Are developers on LinkedIn? Consensus: yes - yes, they are.

20

3 Comments

Attention SoCal engineering leaders! ELC is hosting two free events in November: 📍 Santa Monica / LA – Nov 14: Don’t miss an exclusive live interview with renowned speaker and author Michael Feathers. Get a preview of insights from his upcoming book, AI Assisted Programming, and get a clear picture of how AI can (and can’t) enhance your team’s software practices. This conversation will cover concrete strategies and will steer clear of the usual AI hyperbole. 📍 San Diego – Nov 21: Join us for a roundtable on Practical Measurement & Tracking of Engineering Metrics. Share ideas with peers and separate what truly drives results—and what incentivizes the wrong things. 🔗 Registration links are in the comments below. Attend to network, learn, and grow!

10

5 Comments

What I’ve learned in working with a 10X’er engineering team — A high performing engineering team operates like an R&D lab that focuses on inventing and bringing new products to market. This is in contrast to teams that are focused on building features from a conveyor-belt like backlog. Here’s the DNA of a highly talented team: 1. Great engineers are passionate and curious about inventing new technology. 2. They get a kick on solving valuable problems that no one has solved. 3. Owning a product and having customers use it and see value. 4. Small teams of 10xers out-invent and out-build much larger teams. 5. Talented engineers are some of the most creative people I know; They take an almost playful and curious approach when building technology. Here’s what to be careful on: 1. Too much process around Scrum, JIRA, story-points, and backlogs. 2. Don’t outsource inventing new technology; keep it with the core team. 3. Understand why and how the new invention will create value for customers. 4. Don’t be an overlay or nice-to-have app, solve a hard problem in a new category. 5. Don’t just connect data with API’s. Work at the value layer — making data actionable.

21

7 Comments

I just witnessed a VIP demo of Blitzy. Mind Blowing 🤯 If you are building enterprise software and want to get 1000X better, you should reach out to Brian/Pardeshi Link Ventures

57

1 Comment

Back in 2010 or so I was obsessed with game engines like Unity and Unreal. Sure, they had been around for a while before that, but for me, at that time, these engines unlocked the magical world of game development by taking away the pain of graphics programming, user input, platform deployment (Android Studio, looking at you) and so much more. In 2012 I think I made one game a month for 12 months. Yes, they were all terrible but that's not the point. The point is that it was fun and easy to make games. Remember the mantra, "make games not engines"? The same logic applies to building CAD applications. Developing an application that lets users edit geometry in a robust, easy, and scalable way is very difficult. If you come from game dev, you might think meshes are the way to go. That will seem like great idea until your users want to boolean two meshes that came from who-knows-where. ❌ If you've got lots of extra dollars to spare and you come from the world of engineering, you might consider licensing one of the desktop CAD kernels. Good luck to you! Have fun learning the math behind BREPs for the next couple years let alone shipping that application to paying customers to recover your upfront costs. ❌ At Metafold 3D our vision is to change this. In this talk, our CEO Elissa Ross demonstrates how we developed a personalized medicine application that any clinician could use in less than a week. ✅ Get started today with our #opensource SDKs: https://lnkd.in/gkETXB5k

20

1 Comment

Very useful ... timely, succinct yet comprehensive.

1

🚀 Discover the Future of AI: Claude 3.5 Sonnect vs. GPT-4o Are you ready for a showdown? The GenAI competition just got hotter with the arrival of Anthropic Claude 3.5 Sonnect. This new contender has boldly stepped up to the plate, outperforming OpenAI GPT-4o across multiple tasks. But what sets Claude apart? How fast and/or accurate is it? 🔍 Epsilla (YC S23)’s RAG Research Paper Search App: Your Ultimate Guide Imagine having a side-by-side comparison of a chatbot leveraging each technology covering 361 research papers since January 1, 2024. Epsilla’s cutting-edge app reveals the secrets hidden within these scholarly works. Claude’s lightning-fast answers or GPT-4o’s tried-and-true approach clash head-on? Who emerges victorious? You can see the results from the post below and conclude yourself. 📲 Try It Out: Explore the RAG Research Paper Search App Don’t miss this opportunity to witness the battle of the AIs. Dive into the world of knowledge graphs, speed demons, and groundbreaking research. Click the link here: https://lnkd.in/gxsN6dQp and experience the future today! 🌟 BTW, you can easily switch LLM technologies using Epsilla's no-code modular RAG-as-a-service platform, allowing greater flexibility and velocity to build your own chatbots / smart search apps within minutes at high quality, completely grounded by your own private/public knowledge of choice. Check it out at: https://epsilla.com . #genai #llm #rag #chatbot #knowledgegraph #competition #innovation

9

“... native solutions built entirely around vectors will provide the ‘speed, memory safety, and scale’ needed as vector data explodes.”  “... we do advanced vector search in the best way possible.” At Epsilla (YC S23), we share the same vision and approach. Thus we have been leading the charge with our unique innovations in a 10x faster, cheaper, and better vector search engine, as the built-in foundation for our no-code RAG as a Service Platform. You can experience the accuracy and relevancy from our advanced technologies via a number of high quality AI-powered smart search and chatbot applications below: 1. Smart Search on 180+ RAG Research Papers (https://lnkd.in/eMPdVae5): This smart search app leverages our RAG platform and vector database technology to help user navigate through the massive 180+ papers with ease and keep up with the latest and greatest advanced research on this important topic. . 2. AI Assistant Chatbot for Knowledge Graphs (https://lnkd.in/eeBTFe_Q): Inspired by insights from Mike Dillinger's posts, this AI assistant chatbot leverages our RAG platform and vector database technology to generate highly relevant, nuanced and contextually aware responses to help user stay ahead of the game on topics like LLM, knowledge graph, and GenAI. 3. AI Assistant Chatbot for Taylor Swift (https://lnkd.in/e_m6dB4P): Powered by our RAG platform and vector database technology, this interactive AI-powered chatbot can answer any questions about Taylor Swift, one of the most impactful artists and successful business women in the world. These high quality AI applications underscore the fact that  Epsilla's vector database technology is not just keeping up with the demands of generative AI, but also setting new standards for what is possible. In addition, our no-code RAG platform allows AI app builders to develop and deploy high quality apps like these within hours. Join us as we continue to innovate and lead in this exciting era of AI-driven applications! (https://epsilla.com) #VectorDatabase #RAG #Chatbots #GenerativeAI #NoCodePlatforms

31

8 Comments