Pallavi Vanacharla’s Post

In my time spent speaking with our customers, I’ve noticed the conversations around AI begin to shift. While the race to use AI to improve customer experience is definitely ON, AI applications are costly and plagued by performance issues. I have seen an incredible amount of demand from customers for a way to speed up time to ROI and optimize performance. That’s why our new integration with NVIDIA NIM will be so powerful for our customers. Now, they can leverage the power of New Relic’s observability platform to AI monitor applications built with NVIDIA NIM. This helps them gain a comprehensive view of their AI stack, along with key metrics on throughput, latency, and costs while ensuring data privacy.  This is only the beginning. Learn more via our blog post: https://lnkd.in/giPUDMNB

Have you heard? New Relic and NVIDIA have made it easy for organisations to reduce the complexity and costs of developing, deploying, and monitoring generative AI apps built with NVIDIA NIM 🎉

AI observability is here!

🤖 🧠 There is strong momentum behind NVIDIA NIM. It lets developers build GenAI applications faster, with greater stability, and at lower cost by using industry-standard APIs and the option to self-host AI models in their own managed environments. With this integration, customers can use New Relic AI monitoring for applications built with NVIDIA NIM, all with a simplified setup and enhanced data security.

New Relic + NVIDIA = 💚 Secret’s finally out! 📣 New Relic and NVIDIA released the first observability integration making it easy for companies to monitor the health and performance of their AI applications built with NVIDIA NIM. Jemiah Sius shares how this integration helps developers monitor their AI stack and deploy cost effective and high performance models 👇 Learn more: https://lnkd.in/gWJFDJke

Some great points from Rodrigo below..

Pumped about this! NVIDIA Inference Microservices (NIM) are incredible. They provide prebuilt and optimized deep learning and LLM containers that make it easier to get value from AI. But...what's the fastest way to confidently deploy these models, manage them, monitor them, and make sure they're safe, all on one screen? Well, DataRobot of course! Stop by our booth 1603 at GTC to learn more! #gtc2024 #nvidia #aiops #genai

I’ve been getting into tech content creation and have been slowly sharing my thoughts and takes on LinkedIn! Nvidia has been on a monstrous tear lately and it’s because they’ve positioned themselves to be the foundational layer of the AI industry. If you put the AI industry into a “triangle” format, they are the bottom most layer. This means the entire AI industry is built off the hardware produced by companies like Nvidia.

How far can you go with a single Nvidia DGX A100 Station to train modern LLMs? As it turns out, quite far!!! Today we're releasing **phi-1.5**, a 1.3B parameter LLM exhibiting emergent behaviors surprisingly close to much larger LLMs, despite being trained for less than 2 weeks on only 8 A100s. To get a first feel for what the model is capable of, see the picture below with an example completion and a comparison to Falcon 7B and Llama2-7B. How did we achieve such a model? Well, we used the same recipe as for phi-1 (1.3B parameters model that achieved >50% on HumanEval just 3 months ago), which we call Textbooks Are All You Need. I made an explainer video for the method, recalling the results on phi-1 as well as discussing the new phi-1.5: https://lnkd.in/gTievxiE Moreover, we have decided to open-source phi-1.5 as we believe that it can be a great tool in the community's quest to better understand LLMs, particularly towards mitigating their shortcomings (hallucinations, toxic/biased content generation ..). See the report for a lot more discussion: https://lnkd.in/gkDU7m3U. Our models phi-1.5 and phi-1 are available right now on Hugging Face and Azure ML: https://lnkd.in/g9fTfDBP & phi-1 https://lnkd.in/gStZ-tcD . We can't wait to see what the community will discover with them. Finally, the relatively low cost of developing the phi models has all sorts of deep implications, foremost for the environmental impact of AI, but also it means that this type of exploration can be vastly democratized. I personally can't wait to see what all of you are going to do with this technique, and what is the frontier of what can be achieved at the one billion parameter scale!

There are two things to take note of in this particular statement: More parameters does not neccessarily give you a better result. It’s knowing your model and what it takes to train and improve on the results. As for AI gaining level of self awareness, I personally think we are about a decade away from that true “self awareness.” By that, I mean level of self awareness from higher primates vs all other animals type of self awareness. I am not worried yet in that regards. What I am worried is that companies pushing for “self aware” AI as a marketing terms.