Sarath Sasidharan

What does a Nobel Laureate highlight as AI value blocker? AI Models? . . . Not quite! . . . GPUs and Infra? . . . Not really!! . . . It’s data. Of High quality! So, when model and GPU info in the market sweeps you away, remember to get your data in shape and in place!!! https://lnkd.in/dxaxe9sC #AI #ItStartsAndEndsWithData

3

#MapReduce - invented by Google in 2004, led to the rise of Hadoop and the big data ecosystem, spawning hundreds of new companies selling "Data Analytics" today. #Kubernetes - invented by Google in 2014, led to transforming how cloud applications are deployed and scaled, spawning countless companies that today sell "Managed Kubernetes". #Transformers - invented by Google in 2017, led to, basically everything you hear about AI today. True innovation is a marathon, not a sprint. We're not innovating for today's headlines, but for tomorrow's transformations. Or as Sundar puts it in conversation with Emily Chang, we "...are listening to our own music."

74

1 commentaar

Announcing "Med-Gemini". Building on the core strengths of Gemini 1.0 and Gemini 1.5, we introduce Med-Gemini, a family of highly capable multimodal models that are specialized in medicine and can be efficiently tailored to novel modalities using custom encoders. We evaluated Med-Gemini on 14 medical benchmarks spanning text, multimodal and long-context applications, establishing new SoTA performance on 10 of them, and surpassing the GPT-4 model family on every benchmark where a direct comparison is viable, often by a wide margin. Not long from now, Med-Gemini will be available through the Google Cloud MedLM APIs. Read the full research here: https://lnkd.in/eGRcYbpp

25

Unlock the power of your data with Large Language Models (LLMs). Discover how to leverage cutting-edge techniques like Retrieval Augmented Generation (RAG) and Reinforcement Learning from Human Feedback (RLHF) to drive innovation, improve customer experiences, and streamline operations. Google Cloud can guide you through the journey to a smarter, more efficient future.

3

Wow!! Dynamic scaling and extensive mongoDB AND Vector integration?? Truly demonstrates leadership in the NoSQL space!! This service just gets better and better!

3

Thirsty? Building a high-performance #RAG system is like mixing the perfect gin and Tonic. Tonic.ai, a Databricks ISV partner, offers Tonic #Textual, a tool that converts your unstructured data (PDFs, Word Docs, etc.) into AI-ready formats. It integrates with #UnityCatalog, ensuring both accuracy and privacy. Key business benefits: -- Faster decision-making: Tonic Textual preps your unstructured data, delivering faster and more accurate insights. -- Better productivity: Automating data prep saves your teams hours of manual work. -- Improved accuracy: By adding metadata (like names or dates), AI reduces errors and delivers more precise results. -- Seamless integration: Tonic Textual works with Databricks Unity Catalog and Mosaic AI for easy scalability, so your #AI solutions can grow with your business. Wanna do it? Read here for instructions: https://lnkd.in/gAGGNRNT Tomer Benami, Prasad Kona, Eric Peter

73

14 commentaren

Love the following quote in this recent InfoQ article about our new built-in Gen AI actions: "Azure Logic Apps document parsing and chunking capabilities unlock many automation possibilities. These features, from legal workflows to customer support, allow businesses to harness AI for more innovative document processing. " <-- couldn't agree more

20

Insightful article Key insights - 99.9th percentile query on Redshift and Snowflake scans about 300 GB data which can also be processed in a single node with DuckDB. Specialized engines reading and writing directly on open table formats outside the confines of MPP systems can simplify many use cases. https://lnkd.in/gc8-bQXe

1

In AI news, Amazon is looking to upgrade "Classic Alexa" with more AI, but might charge for it - in a sign of how AI is being commercialised (and always an interesting indicator of how businesses will follow). Conversational AI is not a product that a majority of users are willing to pay for. Paying $5-to-$10 a month just to have a "now you can chat with your shopping cart!" capability isn't going to attract buyers. But cutting through the noise of the current online shopping experience would be super valuable for me. My preferences (which you know already, thanks to my shopping cart) and my personality (which we can discuss after I've read your EULA) could ensure that I keep coming back to you for those things that I am now looking elsewhere for. It's the same in B2B products. Everywhere you look today you see ✨AI magic sparkles✨ which I'm supposed to interact with and eventually pay for. While I do see some organisations being won over by pushy AI sales, I think the majority will expect it as part of the core capability of your product or service. In short: organisations won't buy "the AI", but they're buying an outcome that your product is going to deliver better, faster, or cheaper than the competition. https://lnkd.in/gF_TW3qV via Reuters

33

4 commentaren

I’m officially re-certified for the Google Professional Cloud Architect! Compared to the previous certification a couple of years ago, the questions seemed a lot more complex. This time as part of the exam study, I used Generative AI for : 1. summarizing the study material 2. creating flash cards 3. generating sample questions 4. reasoning for answers 5. overviews of case study and 6. summarizing technical best practices. This helped cut down the study time from 4 to 2 weeks. In addition, one can generate unlimited number of sample questions and answers for creating practice tests. If I could give a tip, please try out NotebookLM for any study preparation. #GoogleCloudLearning #GoogleCloudCertified #NotebookLM

130

12 commentaren

It’s great to see Azure Cosmos DB at the top of the list for preferred database vendors for implementing RAG patterns. Beyond RAG and operational data, Cosmos DB is also utilized for chat history, semantic caching, and user feedback in Gen AI applications. Chat history and user feedback are crucial for understanding what your users are asking and evaluating the quality of the responses provided. I would recommend connecting your chat history with Microsoft Fabric via mirroring to unlock real-time, advanced analytics and BI scenarios for your Gen AI application. #genAI #database #VectorDatabase #azure

56

2 commentaren

Pure gold from Mike Swantek... https://lnkd.in/dfYKT9Yx. Have a look by yourself! In a nutshell: "Data Strategy is the foundation for implementing a modern data platform, and leverages Microsoft's own internal data estate journey"

23

In this Inside #Java Newscast you'll get a look at the goals and history of Project Leyden, and how to evaluate Project Leyden in your projects. https://lnkd.in/edDYwDMF

1

Learn about Retrieval-Augmented Generation (RAG) with vCore-based #AzureCosmosDB for #MongoDB! Discover how to build advanced AI chat applications: https://lnkd.in/eT7BavgX 🚀🤖

1

Neo4j introduced Apache Lucene-based vector indexes on nodes in December last year and on relationships in March this year. This is a short and sweet course on how to create a knowledge graph with vector indexes on text chunks, use few shot learning capability of LLMs to generate Cypher queries from natural language questions, use that query to traverse the graph, perform keyword search, vector search, geospatial search to create the context for LLM to generate the answer. https://lnkd.in/g7fC8wxn

12