MR Rangaswami’s Post

Rodrigo Liang, CEO & co-founder of SambaNova Systems, discusses AI and how recent innovations in AI hardware will affect technology and enterprises on an episode of the Gradient Dissent. Liang explains that production interference will be the “dominating factor” in the cost of AI. Listen to the full episode to hear more of Rodrigo’s insights on the future of AI below. #AI #GradientDissent #Technology

Cascading multiple LLM experts can be a way to specialize each model in a specific task (Mixture of experts) and can be a one better way to address the current issues with LLMs like hallucinations, catestrophic forgetting, reduced inference times. However, it will not solve problems to build a general Intelligence. Tasks are always connected to each other in terms of semantic meaning and there should be a one big model to address it. The current Autoregressive nature of LLMs may be missing objective function of general intelligence which cannot be acheived through the current LLMs transformer structure according to Yann LeCun.

We're almost there! if you're in or around San Francisco tomorrow, 26 October, join the first FemInno outgoing conference outside motherland Armenia. We called it FemInno Satellite and plan to move onwards. Check the very impressive lineup below and join us for meaningful insights on "AI and Inclusive Innovation. #FemInnoSF23 #FemInnoSatelliteSF23 #FemInno23

Our guest this month on Tech Talk with UB VPCIO Brice Bible is truly a pioneer in artificial intelligence research. If you're interested in #ai, then this is a great listen for you. Thanks to Venu Govindaraju for coming on the podcast and sharing his incredible insight and experience. Learn more about the unique history that University at Buffalo has with AI research and how UB is currently working on life changing AI research for good. #podcast #ubuffalo #ubit

The conversation continues to be all things #AI right now and this is a great contribution to that space! Jesse, Erin, and I speak with the creators of Ludia to learn how AI can help make a framework built upon tenets of learner accessibility (#UDL) something that is accessible to all educators.

⚡Wavelngths of Cloo: ⚡S1E3: Soul of a Human, Energy of a Machine - Part 1 In this episode, the truth about PEMOS is unveiled to Cloo during the discovery of the imminent challenge on their planet - Magrathea. Follow Cloo’s adventures while aligning with the wavelengths. Click here to find out what happened next in S1E3 Part 2 https://lnkd.in/g9ij-PSF How do you think Cloo reacted to the big revelation? Comment down below! #WavelngthsOfCloo #cloobotx #genai #comicart

This is how the entire idea of Beatoven came into existence... In this episode, Mansoor spills the beans on how he found a like-minded co-founder who shared an equal passion for music and tech, kicking off their journey in the Gen AI space. However, Beatoven.ai's true journey began when they stumbled upon the perfect problem to solve after extensive exploration and rigorous market validation. 🤔🎵 Watch the entire episode to discover Beatoven's complete story 👉🏼 https://lnkd.in/gV-kEs_3

Is there a mathematical conception of AGI? In my whiteboard conversation with Google DeepMind colleague Marcus Hutter, we discuss his theory of Universal Artificial Intelligence which proposes a candidate combining universal prediction with optimal action. Universal prediction attempts to solve the problem of how to predict the next observation given any unknown sequential distribution. The problem has an elegant solution via Solomonoff induction, which combines Occam's razor (via Kolmogorov complexity) with Bayesian inference. Universal agents are then those agents that can learn in general environments. Starting with the Solomonoff prior and acting optimally accordingly gives us the famous AIXI agent - a candidate for AGI. In general, Shane Legg & Marcus Hutter formalize the notion of universal intelligence. Youtube: https://lnkd.in/efpTHsEy Apple Podcasts: https://lnkd.in/gmHeAkFS Spotify: https://lnkd.in/gGABB73Q

One of the most common attitudes with respect to AI today is the so-called “doomerism,” the idea that AI technologies are inevitably fated to present an existential risk to humanity. This talk from Bryan Cantrill takes that idea on head first, systematically examining the theoretical risks versus the reality on the ground, taking a skeptical but thoughtful view to how we balance the potential of the technology with the varying risks AI may - or may not - represent. Watch: https://lnkd.in/gmDkVVsZ

Techstars Columbus Spring 2024 Demo Day Red Carpet Interview with Ryan Lunka of Doohickey Watch Now: https://lnkd.in/eaMS-FGC Join us on the red carpet at the Techstars Columbus Spring 2024 Demo Day! In this exclusive interview, Elio Harmon, the dynamic host of 614Startups, is joined by Ryan Lunka, Co-Founder of Doohickey. Discover how Doohickey is revolutionizing system integration with its conversational low-code platform. System integration is often a daunting task—it's complex, time-consuming, and can quickly become a costly endeavor. Doohickey simplifies this process, allowing companies to effortlessly integrate their systems with the help of an AI assistant. Learn how this innovative platform makes integration fast, seamless, and painless, saving you valuable time and resources. Don't miss this insightful conversation about the future of system integration and how Doohickey is leading the charge. Like, comment, and subscribe for more exclusive content from Techstars Columbus Spring 2024 Demo Day! #Techstars #Columbus #DemoDay #614Startups #Doohickey #SystemIntegration #LowCode #AI #StartupInnovation