Gopal Kakivaya

Oracle today announced the general availability of HeatWave #GenAI, including in-database LLMs, an automated in-database vector store, and the ability to have human-like contextual conversations in natural language. https://lnkd.in/gK-Q44gc

3

We do have a 45 TOPs NPU in Copilot+ PCs. But it’s not all about speed. You can get speed with GPUs too - even faster! But it’s about efficiency. It’s about battery life on your laptop that you are only allowed to fly (in the US) with a 100WHr battery capacity. Its about minimizing thermal throttling. Its about sustainability. When you get 45 TOPs with single digit watts, you can compute almost all the time. Something like Recall would totally trash your device experience when not offloaded to efficient hardware even if it is able to ‘run’. Phi Silica would take over your system resources without efficient hardware and if you want generative AI running in your games, you may soon end up thermally throttled.

195

4 Comments

This Oracle #HeatWave #GenAI demo will show you how you can have contextual conversations in natural language informed by your unstructured content. https://lnkd.in/gJbAXEND

3

Summary of keynote from Google Cloud Next 24. Gives a glimpse of enhancements centered around AI as a theme in Google Cloud, Vertex AI and Gemini..... https://lnkd.in/grbfrf9H #google #googlecloud #gemini #ai #vertexai

3

🚀 Google Cloud Advances Generative AI Capabilities with Gemini 1.5 on Vertex AI 🚀 See this blog by Amin Vahdat: https://lnkd.in/gqSjpB-z Google Cloud has announced significant enhancements to its Vertex AI platform with the release of Gemini 1.5. 🧠 Gemini Pro 1.5: Optimized for complex tasks demanding high-quality output. Its expanded 2 million token context window 📖 enables comprehensive analysis of extensive documents, codebases, and multi-modal data 📊, driving breakthroughs in natural language understanding, code generation, and creative applications. ⚡Gemini Flash 1.5: Engineered for high-throughput, cost-efficient processing of large-scale workloads. While retaining a substantial 1 million token context window, its lower latency and cost profile make it ideal for real-time applications like chatbots 💬, captioning, and content extraction. 📰 Beyond model enhancements, Google Cloud has introduced features specifically tailored for efficient context management: 🧠 Context Caching: This feature enables the caching and reuse of previously processed context data, significantly reducing both latency and compute costs for applications with long input sequences. 🔁 Batch API: The ability to process multiple requests simultaneously optimizes throughput and resource utilization, particularly beneficial for high-volume scenarios. ⏫ Controlled Generation: This feature provides fine-grained control over the length and style of generated outputs, allowing for greater customization and alignment with specific requirements. 🎛️ These advancements represent a major step forward in generative AI, enabling developers and enterprises to build more sophisticated, context-aware applications that can process and understand massive amounts of information. ⚙️ Google Cloud offers a diversified model portfolio and specialized tools with optionality, This empowers users to unlock new levels of efficiency and innovation across a wide range of industries and use cases. 🗝️ #GoogleCloud #VertexAI #Gemini15 #GenerativeAI #MachineLearning #AI #Innovation Congrats Amin Vahdat Burak Gokturk Warren Barkley Nenshad Bardoliwalla Jason Gelman Lillian McNealus Erwan Menard Ting Liu Riham Mansour

63

Vector search for Amazon MemoryDB is now generally available! See here: https://lnkd.in/gkMYq6H3 You can use vector search for MemoryDB to store, index, and search vectors. You can develop real-time machine learning (ML) and generative AI applications with in-memory performance and multi-AZ durability. You can use the MemoryDB API for use cases such as Retrieval Augmented Generation (RAG), anomaly (fraud) detection, document retrieval, and real-time recommendation engines. You can generate vector embeddings using AI/ML services like Amazon Bedrock and Amazon SageMaker, and store them within MemoryDB. With this launch, Amazon MemoryDB delivers the fastest vector search performance at the highest recall rates among popular vector databases on #AWS. You no longer have to make tradeoffs around throughput, recall, and latency. Check it out! Amazon MemoryDB is a durable, in-memory database for ultra-fast performance. It has microseconds read latency and low single-digit millisecond write latency, security features (encryption, authorization, authentication, and more), a flexible open source API, and a 99.99% availability SLA. Customers use MemoryDB for their high throughput and low latency demanding workloads. You can learn more about MemoryDB here: https://lnkd.in/gA3wk3mv #awscloud #genai #aiml #memorydb #vectors #vectorsimilaritysearch #vss #rag #performance

168

1 Comment

AWS continues to add capabilities to Amazon Bedrock that make it the easiest way to build high quality, low latency, and cost-effective GenAI applications. Today the team announced:   - Custom Model Import, which lets customers easily bring their own models (from SageMaker and S3) to Bedrock to leverage its leading GenAI features, such as RAG, agents, Knowledge Bases (guardrails), inference deployment, and advanced customization to keep teaching and refining models.   - Model Evaluation, which lets customers quickly select predefined evaluation criteria (like accuracy and robustness) and run reports that help them understand which is the best model for them.   - Meta’s new Llama 3 models, now available on Bedrock and can support a wide range of use cases including text summarization and classification, sentiment analysis, language translation, and code generation.   - New Cohere Models, Command R and R+, are state-of-the-art FMs customers can use to build enterprise-grade generative AI applications with advanced RAG capabilities, in 10 languages, to support their global business operations. These will arrive on Bedrock in the coming weeks.   - New Amazon Titan models with Amazon Titan Image Generator (GA today) helping customers produce studio-quality images, at low cost, using natural language prompts (and applies an invisible watermark to help reduce the spread of disinformation)… along with Amazon Titan Text Embeddings V2 model, which reduces storage and compute costs, while out-performing other leading models on text embeddings.   Tens of thousands of customers are already building real-world production-ready GenAI apps and we're excited for even more customers to get started. https://lnkd.in/g4rkak47

2,849

53 Comments

One does not simply waive a wand and make DevSecOps applicable to MLOps. Nice, concise summary of some of the basic versioning challenges associated with AI/ML model building, and why it cannot be treated as simply sequential or iterative. https://lnkd.in/e7MeXBN8

1

Be sure to check out Vamshidhar Kommineni and Vishnu Charan TJ as they demonstrate insane levels of throughput from Blob storage alongside our much improved Blobfuse client! (~8Tbps of ingress, ~14Tbps of egress) @ #MSBuild https://lnkd.in/g6S6TxPu #AzureStorage #AzureBlobStorage #BlobStorage

63

4 Comments

Oracle today announced the general availability of HeatWave #GenAI, including in-database LLMs, an automated in-database vector store, and the ability to have human-like contextual conversations in natural language. https://lnkd.in/gqZ2ebDS

1

PostgreSQL 17 Beta 1 is now available in Amazon RDS Database Preview Environment! Try it and give early feedback! https://lnkd.in/gEeeAZ_H #Amazon #RDS #PostgreSQL

95

2 Comments

Today, AWS announces the general availability of Amazon OpenSearch Service zero-ETL integration with Amazon S3, a new efficient way for customers to query operational logs in Amazon S3 data lakes eliminating the need to switch between tools to analyze data. Customers can quickly get started by installing out-of-the-box dashboards for AWS log types such as VPC Flow, WAF, and Elastic Load Balancer. Customers who use OpenSearch Service also use Amazon S3 as a cost-effective way to store infrequently-accessed operational log data. To perform analysis on Amazon S3 data and correlate data across multiple sources, customers previously had to copy that data into OpenSearch Service to take advantage of its rich analytics and visualization features that help them understand data, identify anomalies, and detect potential threats. However, continuously replicating data between services can be time consuming, expensive, and hard to maintain. With OpenSearch Service zero-ETL integration with Amazon S3, customers can access operational log data stored in Amazon S3 using OpenSearch Service, making it easier to perform complex queries and visualizations on their data without any data movement. https://lnkd.in/dN_YaCqd

52

2 Comments

AlloyDB provides 2x better price-performance than self-managed PostgreSQL. See how in this article from Sridhar Ranganathan and Shesh Ranganath.

14

Oracle Database 23ai, previously known as Oracle Database 23c, is now Generally Available. Get an inside look at some of the major features and learn where you can download the software. https://lnkd.in/gzdApb57

2

For anyone interested in Confidential Computing on Arm platforms, you must watch the update that Dan Handley from Arm provided at Linaro Connect. He covered the current state, what's coming next in the Confidential Compute Architecture (CCA) specification, and what to expect. Watch the recording at https://lnkd.in/dRiCz9TH.

25

Check out this great article by Kara Saucerman! It talks about CETO, a high-performance gateway serving as the front-end for Microsoft AI services. It leverages YARP (Yet Another Reverse Proxy) and the latest version of .NET.

9