Linda Griffin’s Post

RAG can improve LLM responses, but it's not a magic wand. DataStax's Dom Couldwell discusses RAG challenges, opportunities, and real-world use cases with TechZine. #RetrievalAugmentedGeneration https://ow.ly/MnI450R3oqn

RAG can improve LLM responses, but it's not a magic wand. DataStax's Dom Couldwell discusses RAG challenges, opportunities, and real-world use cases with TechZine. #RetrievalAugmentedGeneration https://ow.ly/Oyzx50QAmpn

RAG can improve LLM responses, but it's not a magic wand. DataStax's Dom Couldwell discusses RAG challenges, opportunities, and real-world use cases with TechZine. #RetrievalAugmentedGeneration https://ow.ly/TW7R50QFtRg

https://lnkd.in/eVD2U99e As an accelerant of modern software, open-source maintains a cherished place in tech. Who can argue with free stuff, decentralized control, and free speech? But open and closed-source AI cannot both dominate in the limit: if centralizing forces hold, scale advantages will compound and leave open-source alternatives behind. Perhaps GOOG’s AGI issues (few and far between for now), could be a harbinger of open vs close source models…

Hey! We just released a new video explaining what GavelGen does in under a minute. Check it out! #generativeai #llmops #gavelgen

Have you ever tried Googling something mid-conversation? Congrats, you've performed #RAG! I had this realization recently: you know how sometimes, when you're in a conversation, you discreetly Google something to sound more informed? That's essentially what Retrieval-Augmented Generation (RAG) does. RAG is like having an AI that excels at finding the most relevant information mid-conversation to give you the best possible answer. It's not just making things up; it’s searching for the latest and most accurate data, then integrating it seamlessly into the response. In essence, RAG ensures that answers are not only generated but also enriched with the most current information. It’s like having a super-smart assistant who always has the right answers. So, to end these endless debates - RAG vs. Fine-Tuning - they are different tools for different purposes. Both are needed. And for those who thought there's some mystery behind these searched results - they're simply added to the LLM prompt. As simple as that.

Today, we are seeing a complete meltdown in publicly traded tech companies. There's blood in the water, and value to be had for those paying attention. With that context, in addition to last night's results from MSFT and GOOGL, I wanted to re-share the article I wrote this week that can help you navigate this turbulence. https://lnkd.in/eMmmQNFn

Quick Read on "How I selected the GenAI Platform for my use cases" - https://lnkd.in/ewQyeymt

Article on how the world wound up with the most reckless unleashing of high-risk, advanced technology since the atomic bomb. In the case of the atomic bomb, they had a good reason -- end the war and save more lives in the end. In the case of consumer LLM chatbots it was panic in an old big tech company desperate to protect itself and become ever larger. They had milked monopoly products for a decade or two too long, had become a security nightmare, became an increasing systemic risk for the economy, and earned a reputation for being the highest bidder among really big, really dumb money in M&A. I'm speculating a bit here, but it's not entirely speculation. While I completely appreciate the need to focus on profitable products people want to use, and few are more frustrated by hundreds of billions of dollars worth of research molding on shelfs, by going around MSFT research direct to small AGI cults, MSFT also went around the ethical processes for commercialization of technology. This didn't just break a few things in the china shop -- it literally risked all of society. With all due respect to the CTO, CEO, Vice Chair, and Chair of MSFT, that is way beyond their authority or competencies. They were not anointed God. Market power and wealth does not a responsible person make, and wealth and market power certainly does not ensure integrity or competency in advanced technology. What this experience demonstrated in quite clear form was that big tech must be broken up. Not just restrained by antitrust like the previous MSFT case, but divided up into smaller pieces with less systemic risk and market power. They've become much too large and high risk. That's what their behavior is screaming -- that's the message that should be received loud and clear by Congress and SCOTUS. This is how market failure speaks when governments have ignored their behavior for too long, how the twisted soul reveals itself, and how the power cultures beg to be put out of their misery -- preferably before a meltdown of our economy occurs. Microsoft isn't smart about AI. It's reckless, unethical, and incredibly wasteful. They've proven over and over again they can't be trusted with the power they have amassed. They should all be fired for wildly reckless behavior.

Must (long) read for anyone who wants to understand how the "most complex software object ever created" happened and where AI is headed in the near future: https://lnkd.in/emKZQ_Ry