Posts

Showing posts from November, 2023

Azure Flowcharts - Default Outbound Access

Image
If you deploy a virtual machine in Azure and it doesn't have explicit outbound connectivity, it's assigned a default outbound access IP. This IP address enables outbound connectivity from the resources to the Internet.  Original image source   NAT gateway is the recommended approach to have explicit outbound connectivity. A firewall can also be used to provide this access. Related article - Default outbound access in Azure

My PartyRock App #4 - SWOT Analyzer

Image
I provided this prompt to the PartyRock App Builder - " Perform a SWOT analysis of two entities " Within seconds it came up with a form with all the text input widgets and a text generation widget hooked to the Claude Instant model with the following prompt -  Perform a SWOT analysis comparing [Entity 1] and [Entity 2] . Include their key strengths, weaknesses, opportunities and threats based on the provided details. Check out the  SWOT Analyzer and an example  comparing two entities. Also see -  Public PartyRock Apps The Inside Story of PartyRock My other apps: Global Cuisine Explorer Ask Me Anything Mock Interview GeoFlix

AI Calms the Beast Within

Image
  More of my AI co-creations

Intro to Large Language Models - Andrej Karpathy

Image
Andrej Karpathy who has worked on AI projects at Google, Open AI and Tesla has the gift of explaining complex topics in a simple & accessible way. It is great that he shares his knowledge though his YouTube channel . Large Language Models are the core technical component behind systems like ChatGPT, Claude, and Bard. In his recent 1-hour long talk " Intro to Large Language Models ", he covers a wide range of LLM topics including:  LLM Inference  LLM Training  LLM dreams  How do they work?  Finetuning into an Assistant  Appendix: Comparisons, Labeling docs, RLHF, Synthetic data, Leaderboard  LLM Scaling Laws  Tool Use (Browser, Calculator, Interpreter, DALL-E)  Multimodality (Vision, Audio)  Thinking, System 1/2  Self-improvement, LLM AlphaGo  LLM Customization, GPTs store  LLM OS  LLM Security Intro  Jailbreaks  Prompt Injection  Data poisoning  LLM Security conclusions He has also shared the slide deck (links in the YouTube video description). Also see -  A 15-day email

Dr. Vint Cerf on AI

Image
It was fascinating to watch and hear Dr.Vint Cerf, one of the fathers of the Internet and currently the Chief Internet Evangelist of Google, speak about AI  in a recent NDTV interview . Till technology finds a way to trace their origins, one way to tackle the dangers of deepfakes could be to teach people to question motives, according to the Internet pioneer, Vint Cerf. He suggested one thing to teach everybody: critical thinking - "Ask yourself, where did this come from? What's its purpose? Am I trying to be persuaded of something that I shouldn't be persuaded of?" . People need to be taught that not everyone has "your best interest at heart". Also see - Vint Cerf Infographic from Google [PDF] Shortened transcript -  Q - Everyone is talking about the future of AI but let me ask you the big question which so many people have on this from a technology standpoint are we getting to a point where AI systems could become self-aware? A: Well, I hope the answer i

This Week I Learned - Week #47 2023

Image
This Week I Learned -  * The Applied Skills assessment lab gives you the opportunity to demonstrate your skills by completing tasks in the associated technology, such as Azure, Microsoft 365, or Power Platform.  You will perform tasks in a virtual environment.  Most Microsoft Applied Skills assessment labs for a given scenario typically contain 12-16 tasks; however, the number can vary depending on the skill domain being assessed. You will have 2 hours to complete the lab with the ability to request an additional hour if needed as you progress through the lab. For most Applied Skills credentials, you’ll be informed of your results within minutes, but it can take up to 24 hours for results to appear in your Learn profile’s Credentials tab.  *  Amazon Elastic Compute Cloud (Amazon EC2) UltraClusters can help you scale to thousands of GPUs or purpose-built ML accelerators, such as AWS Trainium , to get on-demand access to a supercomputer. They democratize access to supercomputing-class

AI is Like a Diabolical Diplomat

Image
More of my AI co-creations

A PartyRock app to Prepare for Interviews

Image
I remixed an existing Resume Optimizer PartyRock app to include a chatbot that conducts a mock interview based on the provided job description.  Check out Mock Interview I used this prompt for the Mock Interviewer chatbot - I want you to act as an interviewer. I will be the candidate and you will ask me the interview questions related to the [Job Description] . Include questions related to [Missing Skills] . I want you to only reply as the interviewer. The inter-widget interactions that PartyRock orchestrates without users having to write any code makes it an amazing tool to build simple apps. Also see -  Public PartyRock Apps The Inside Story of PartyRock My other apps: Ask Me Anything GeoFlix SWOT Analyzer

The Inside Story of PartyRock

Image
Andy Jassy's explains how PartyRock came to be - Earlier this year, some AWS developers created an internal tool for Amazonians to experiment with Foundation Models from Amazon Bedrock, and it quickly became a word-of-mouth hit across AWS. Because the interface only required a user to add in a description of the app they wanted to build in a prompt box, the tool could be used by teammates across a wide spectrum of roles and technical expertise. Some built apps for productivity, while others made ones for amusement, or that tapped into personal passions. It was cool to see builders get inspired by using generative AI in a more hands on way, and by sharing their apps with each other, they were actually sharing key generative AI techniques in action. Based on the popularity and feedback, we launched this simple to use but powerful app to the world for free, naming it PartyRock (in reference to it being a fun and collaborative way to experience Amazon Bedrock). It is built on the beli

Ask Me Anything PartyRock App

Image
I noticed that most of the public apps that are out there currently use Anthropic's Claude or Claude Instant Models. To compare, I wanted to test how AI21's Jurassic and Cohere's Command Models would respond to the same query when asked.  I created my second PartyRock App to scratch that itch - Ask Me Anything . Here's a snapshot that you can check and remix for your own use case if you find it beneficial. Post any query & see which Model provides the best answers for a particular prompt & context. According to the FAQ, new and updated models may be added regularly. Currently 6 LLM models are offered: Anthropic's Claude Anthropic's Claude Instant AI21's Jurassic-2 Mid AI21's Jurassic-2 Ultra Cohere's Command Stability AI's Stable Diffusion XL   There are excellent resources on Prompt Engineering from AWS , Anthropic , AI21 and Cohere . The  Ask Me Anything  app can be used to experiment with Prompt design for each of the available Mode

PMP Process Groups, Knowledge Areas, Processes

Image
Project Management Professional (PMP) is an internationally recognized professional designation offered by the Project Management Institute (PMI).  The closed-book exam consists of 180 multiple-choice questions written against the PMBOK (Project Management Body of Knowledge)  specification and the PMP Code of Ethics .  Since January 2021, the exam is tested on three different domains : People (42%) Process (50%) Business Environment (8%) The exam is based on tasks from five performance domains or process groups , encompassing a total of 49 processes. The PMP Examination Specification  weighted these groups as follows: Initiating the project (13%) Planning the project (24%) Executing the project (31%) Monitoring and controlling the project (25%) Closing the project (7%) Mapped to these five process groups are ten project management Knowledge Areas : Project Integration Management Project Scope Management Project Schedule Management Project Cost Management Project Quality Management Pr

Public PartyRock Apps

Image
I started compiling this list of public PartyRock Apps soon after PartyRock was released to learn how GenAI enthusiasts are crafting their prompts & to explore creative solutions. I have updated the list according to the categories mentioned in the PartyRock Generative AI Hackathon submission guidelines - 1. Interactive Learning Experience - virtual science lab, language lesson, etc; SWOT Analyzer TransLitify Prompt Doctor Code Reviewer   Bird Buddy Icebreaker Topic Finder Late Night Current Event Comedy Daily Language Learner Claude Prompt Engineering Interactive Tutorial Car Advisor Yoga Flow Coach Quiz Creator DynamoDB Single Table Designer DynamoDB Modeling Application AWS Cloud Formation Template Generator AWS Best Practices Recommendation Assistant SAP-on-AWS Blog & Documentation Linker Azure Helper App Data Strategy Builder_Intern Assist Multi-Level Why Questions Generator Data Governance_Productivity Assistant  Oracle to PostgreSQL Data Migration Tool AWS Product Detail

GeoFlix - My First App on PartyRock

Image
PartyRock from AWS is a fun way to build simple Gen AI apps. PartyRock uses foundation models from Amazon Bedrock to turn your ideas into working PartyRock apps through its code-free app builder (similar to OpenAI's GPT Builder ). Amazon Bedrock is a fully managed service that makes foundation models (FMs) from Amazon and leading AI companies available on AWS available through an API.  You have to sign in with a Google, Apple or Amazon account - In the next step you can choose a username - After I tried out some of the featured apps on the home page, I was able to build my own in minutes and publish it publicly - GeoFlix, check it out !  All I did was to provide the prompt that "the app should identify and suggest movies that prominently feature specified locations or place names within their storylines." I edited the pre-created app name to GeoFlix and clicked on the "Make public and Share" button.  And that's all it took to get a URL for my app GeoFlix

Idea Checklist

Image
The Tata Imagination Challenge is an annual idea competition open to college and postgraduate students in India and Tata employees around the world. Registrations for 2023 are closed but their Idea Guidebook  [ PDF ] has good tips & suggestions on how to approach problems and solutions: Plausible impacts to consider while identifying an idea - Improve or change an existing practice. This could be anything from the way people shop to the way they consume news. Create new business avenues or help existing businesses perform beyond their expectations. Help people or communities who are less fortunate. It could be an idea for generating employment, providing education, better healthcare, etc. Help nature or animals because unless we act now, the future won't be so bright. Idea Checklist - Is your idea innovative and original? Is your idea clearly defined and easy to understand? Does your idea solve a real problem? Have you done research and analysis before arriving at the solution

This Week I Learned - Week #46 2023

Image
This Week I Learned -  * PartyRock is a fun way to learn the fundamentals of foundation models and then collaborate with other developers. Anyone can access PartyRock through the website, and you don’t need an Amazon Web Services (AWS) account. PartyRock uses foundation models from Amazon Bedrock to turn your ideas into working PartyRock apps. You don't need to be a developer or ML engineer to start creating your own apps. Just write a text-based prompt to describe what you want your app to do in the PartyRock app builder. * Amazon Bedrock is a fully managed service that makes foundation models (FMs) from Amazon and leading AI companies available on AWS available through an API. It is the easiest way to build and scale generative AI applications with foundation models. PartyRock provides builders with access to FMs from Amazon Bedrock in a code-free app-building playground to learn the fundamentals of prompt engineering and generative AI. *  Virtual Assistants (VA) are built by