Generative AI in a Weekend: Learn ChatGPT and OpenAI

Explore how strong AI and narrow AI differ, and see how current systems like self-driving cars and fraud detection aim to simulate human-like intelligence.

Explore how narrow AI like OpenAI's Chat GPT generates content and assists tasks, from planning trips to coding, while avoiding sensitive data and verifying outputs as starting points.

Explore using ChatGPT to code, learn, and design, from writing a leap year function and unit tests to designing REST APIs, database schemas, and exploring design patterns.

Explore technology with chatgpt and identify top cloud engineering technologies, from cloud platforms and containers to infrastructure as code and security, plus docker exercises and prompt-driven insights.

Generate ideas with ChatGPT by crafting prompts to understand DevOps, its culture, principles, and tool chain from development to monitoring. Explore prompts that reveal DevOps tools using analogies.

Explore how generative ai is not deterministic; the same prompt often yields different responses, and adjusting a few parameters can influence output variability.

Maximize learning efficiency with playback speed and speed watching to save time, using 1.25x to 2x while maintaining comprehension. Begin at 1x, then gradually increase and invest in yourself.

Explore how ChatGPT works, clarify terms such as artificial intelligence, machine learning, generative AI, large language models, and foundation models, and discuss its limitations and use cases.

Compare machine learning, which learns from millions of examples to build predictive models, with traditional programming that relies on hand-written rules, and distinguish narrow AI from general AI.

Learn how machine learning uses features and labels to make predictions, distinguish regression from classification with real-world examples like house price, used car price, spam detection, and loan decisions.

Learn the traditional machine learning workflow from obtaining and cleaning data to feature engineering, model creation, evaluation, and deployment, and key terms like training, evaluation, and inference.

Navigate the AI turmoil with a pragmatic mindset and learn to use AI solutions without fear, embracing APIs you can consume and integrate to work faster.

Generative AI learns from examples to create new content, distinguishing it from AI and machine learning. It covers text, code, and image generation as core use cases.

Generative AI relies on huge volumes of data, analyzing images, text, and code from sources like Wikipedia, books, and open source code to train models such as GPT-3, with infrastructure.

Generative AI uses self-supervised learning, training on vast text data to predict the next word without explicit labels. The model learns word relationships and context through repeated prediction and backpropagation.

Explore how generative ai learns through self-supervised training by predicting the next word within context, using probabilities and controllable parameters like temperature, top_k, and top_p to shape output.

Learn how generative artificial intelligence text models use tokens instead of words to handle context, meanings, and relationships, and why token limits require splitting long content.

Explore how generative AI models predict the next word from information learned from internet data. See how supervised fine tuning with labeled prompts and responses trains models to answer questions.

See how supervised fine tuning and reinforcement learning from human feedback build a reward model, rank responses, and tune a generative AI to align with human values.

Compare predictive machine learning with generative AI, highlighting inputs, outputs, use cases, data needs, and training time, from features and predictions to prompts and generated content.

Discover foundation models and large language models, like GPT, that are pre-trained multitask systems for chat, review analysis, and article summarization, with multimodal capabilities across text, video, audio, and image.

Discover the generative AI landscape from foundation models like GPT and DALL-E to OpenLaMA, and see how OpenAI, Azure OpenAI, and Google Cloud PaLM APIs power apps like ChatGPT.

Explore OpenAI Playground, a web-based tool to experiment with the OpenAI API, using chat and completion modes, multiple models, and adjustable parameters.

Discover how to use the OpenAI Playground for text summarization, login and model basics, and practical prompts. See examples like summarizing course descriptions, reviews, and generating hashtags.

Explore text classification with generative AI, labeling sentiment (positive, negative, or neutral) and categories for news and reviews, and return a json object with sentiment and explanation.

Explore extraction features by learning to pull text specs into a JSON format and answer questions from documentation using prompts. Practice crafting prompts and generating outputs for technical specifications.

Explore ideation with generative AI: generate use cases, ideas, names, and interview questions, including cloud and DevOps trends like automation, Kubernetes, serverless, hybrid cloud, and AI/ML, plus cloud cost practices.

Learn to stay ahead of technology by building a daily learning habit through the in28Minutes learning pledge: commit to 28 minutes of study for 28 days.

Explore prompt design fundamentals to craft effective prompts for foundation models like ChatGPT. Apply best practices—clear instructions, examples, and experimentation with prompt design frameworks such as RTF, CTF, and RASCEF.

Design prompts with output formats like json and word limits, test think-step-by-step vs. no-step, and instruct the model to act as an expert while saying 'I don't know' when unsure.

Master zero shot, one shot, and few shot prompts by incorporating examples to guide responses and even format outputs as json objects.

Explore prompt frameworks like RTF, CTF, and RASCEF to craft precise prompts by defining role, persona, context, action, steps, format, and examples.

Master how to tune OpenAI Playground outputs by adjusting max tokens, temperature, top P, and penalties; learn through practical examples that parameter choice is an art, not science.

Learn to integrate OpenAI API into apps, generating language, code, images, audio, and fine-tuning models, with GPT-4, GPT-3.5, DALL-E, and Whisper.

Understand OpenAI API pricing, billed per 1,000 tokens, varying by model. Note independence from ChatGPT Plus, with GPT-4 at $0.06 and GPT-3.5 Turbo at $0.002, plus a free $5 credit.

Generate a top three cloud and DevOps trends using the OpenAI Playground. Then run the same prompt from Python in Google Colab, configuring the OPENAI_API_KEY and using the ChatCompletion API.

Explore the OpenAI chat completion response, including id, created timestamp, model, and content, and understand finished reason, token usage, max tokens, and per 1k tokens.

Learn to run the OpenAI API with prompts that include examples using system, user, and assistant messages. Explore how example inputs and outputs drive the chat completion flow.

Learn to build a chat bot with the OpenAI API’s chat completion, using system messages and examples to control behavior and reduce hallucinations for customer service and education.

Explore OpenAI API image and audio capabilities by generating images with DALL-E, selecting sizes, and using whisper for transcribe, text-to-speech, and translate; follow notebook examples.

Learn to tune large language models by creating a jsonl training dataset, uploading it to Google Colab, and running a fine-tuning job on gpt-3.5-turbo to produce a customized model.

Explore embeddings as vector representations in a high-dimensional space that capture semantic relationships and contextual information, then measure text similarity for recommendations, clustering, and outlier detection.

Explore embeddings in a hands-on Colab example, calculating text embeddings with OpenAI, storing them in a pandas DataFrame, and using dot product to find the most similar sentence.

Explore Lang Chain to interact with multiple language models via abstractions for text, chat, and embeddings, and learn to switch between OpenAI, Palm, and other providers.

Explore LangChain fundamentals by configuring OpenAI, installing libraries, and using prompt templates to create flexible prompts; switch models easily, run chat models, and generate embeddings.

Split long articles into chunks with LangChain’s recursive text splitter, compute embeddings, and perform similarity search with FAISS to retrieve relevant sections, then answer questions with a QA chain.

Explore LangChain summarization, comparing the stuff approach with the map-reduce method that summarizes documents chunk by chunk and then combines them for long documents.

Discover how to build and run a LangChain sequential chain that executes a tool prompt and then a summary prompt, using OpenAI, templates, and LLM chains to summarize Terraform.

Explore the various ChatGPT plans, including free, plus, and enterprise options, and understand access to GPT-3.5 vs GPT-4, features like plugins and data analysis, and the limits.

Enable ChatGPT plugins to extend capabilities with trusted third-party apps. Remember that plugins are a paid feature requiring ChatGPT Plus and careful attention to data sharing and settings.

Experiment with GPT-4 enabled plugins to access Wikipedia, Kayak, Link Reader, and other tools, then generate up-to-date facts, travel options, and diagrams for insights.

Enable advanced data analysis in GPT-4, upload countries.csv, and use Python with pandas to describe data, create visualizations, and perform image processing.