The Ultimate Guide To ChatGPT & Midjourney

Create a ChatGPT account on openai.com, verify with a phone number, then access chat.openai.com to start chats, with optional ChatGPT Plus and GPT 3.5 vs GPT 4 choices.

Explore the chat interface basics, including choosing GPT-4 or GPT-3.5, composing prompts, regenerating responses, and managing chats with names, feedback, clipboard, and history controls.

Open the three-dot menu to access ChatGPT settings, choose a color theme (light, system, or dark), and manage data controls like exporting conversations and the chat history and training toggle.

Learn how to edit prompts after generation, update prompts in place with new context and requirements, and rerun outputs to refine marketing copy for courses.

Discover how chat sessions are siloed and lack memory, as full transcripts are fed to the model with every prompt, creating context without true memory, and learn about token limits.

Learn how to achieve continuous output from ChatGPT after a stop by using continue or regenerate. Understand how prompts, output limits, and resume behavior affect translations and longer writings.

Learn how the interface is adjusted by hiding the left column to enlarge text, keeping the focus on prompts and the generated outputs in ChatGPT.

Identify the training cutoff for ChatGPT 3.5 and 4 as September 2021, and explain why it lacks real-time data by default, with web browsing available via ChatGPT Plus.

Examine ChatGPT hallucinations, also called delusions or confabulation, where models answer confidently yet err. Learn to cite sources and verify facts to mitigate this risk.

Explore what chatgpt refuses to answer, including illegal activity, child sex abuse, hateful or violent content, and risky medical or legal queries, with notes on guardrails and jailbreak attempts.

Explore how tokens encode text and define the context window in GPT models, and learn how to avoid hitting per-message and total token limits in GPT-3.5 and GPT-4.

Master prompt engineering by learning to craft high-quality prompts that optimize ChatGPT and other large language models, using the four key ingredients of a winning prompt formula.

Set the scene with an initial context or initializer to guide a general purpose language model. Define instructions, specify the task, provide input data, and apply constraints and formatting rules.

Master the act as and imagine as syntax to set initial context for ai models with simple prompts like you are a writing tutor, and learn when it works.

Explain the actual task within a prompt as the core instruction, using context, examples like writing a tweet or generating code, per the prompt formula.

Place context, task, and input data in the recommended order to craft effective prompts. Learn to structure input data, wrap text, and separate formats such as text, JSON, or code.

Explore how to design prompts with output formatting constraints and specific output formats, including lists and tables, to control results in ChatGPT and Midjourney prompts.

Design and test prompt-driven article and Wikipedia summarizers by using an initial context, constraints, and bullet-point outputs to generate concise, actionable summaries.

Learn to summarize YouTube videos and podcasts by extracting transcripts and prompting ChatGPT as a transcript summarizer. Practice crafting prompts, handling transcripts in chunks, and tracking speakers and word counts.

Explore zero-shot, one-shot, and few-shot prompting by starting with no examples, then adding input-output examples to guide tasks like sentiment analysis and output format.

Learn to design and apply output templates for prompts, using end-shot prompting to specify a desired format and ensure repeatable results with format templates, such as JSON or structured data.

Learn to improve responses by instructing the model to ask follow-up questions before answering, gathering information to fill gaps for complex prompts like scripts or marketing copy.

Discover how ChatGPT accepts tables as input and produces tables as output, including markdown table formats and export options for easy HTML or markdown reuse.

Learn to create and interpret Markdown with ChatGPT, mastering headings, formatting, lists, code blocks, links, images, tables, and live rendering in an editor.

Learn how to format data with ChatGPT across csv, json, and markdown tables, and output data in csv or json for easy integration with Excel, Pandas, or Python.

Learn how to use incremental prompting by breaking complex tasks into sequential, single‑purpose prompts, outlining and drafting steps to improve results for writing, coding, and multi‑file tasks.

Practice self-critical prompting to have the model critique its own output, improve blog posts with detailed resources, iteration, and post-interview reflection.

Explores chain of thought prompting. Uses think step by step prompts to elicit intermediate reasoning and explain your reasoning step by step, improving model accuracy on reasoning tasks.

Learn to cue ChatGPT to start its output at a specific point, control results with bullets or numbers, and place starter text for precise outputs.

Create mindmaps for the course using markmap.js.org with markdown syntax and headings, generating nested levels; explore few-shot and one-shot prompting for prompt engineering.

Learn to prompt for multiple perspectives and audiences, generating tailored outputs—tweets, blogs, or videos—from distinct personas.

discover how the ultimate guide to ChatGPT & Midjourney demonstrates using ChatGPT to generate prompts for itself, including an interactive prompt assistant and its practical caveats.

Evaluate online prompting resources and separate useful prompts from scams and junk. FlowGPT lets users share and vote prompts, including Facebook ad headline prompts, discussions on prompt quality and ethics.

Examine why the ChatGPT web browsing feature is temporarily disabled due to abuse that could bypass paywalls, and how re-enabling it could let ChatGPT read and summarize articles.

Enable ChatGPT web browsing with browse with bing and plugins, then choose it in a new chat to fetch current data, such as 'Last Night' by Morgan Wallen.

Explore web browsing techniques beyond a single google search, including stock market queries and top gaining stocks. Learn to verify data, read content, and summarize multiple news stories.

Demonstrate a more complex web browsing prompt by extracting the first five startup names from the Hacker News jobs board, then fetch their funding and output a markdown table.

Explore common bugs in web browsing prompts, including navigation errors and timeouts, that derail article summaries, with guidance on diagnosing why prompts fail.

Discover how web browsing with chatgpt encounters a time limit of about one minute per request, hindering tasks like summarizing multiple articles.

Discover how to enable and install ChatGPT plugins, explore the plugin store, and use plugins to generate diagrams and perform real-world tasks within chat.

Explore the AskYourPDF plugin for ChatGPT, showing how to download, index with a vector database, and query PDFs to summarize research papers and manuals.

Explore how the ChatWithVideo plugin analyzes YouTube videos from a URL, generating summaries and key points from transcripts, with comparisons to other plugins.

Explore the OpenTable plugin in a chat-based format to search and reserve restaurants by city, date, party size, occasion, and other parameters, with example in New York City.

Enable two factor authentication for select store plugins, scan the QR code with an authenticator app, and sign in to Notable using GitHub to access an interactive notebook.

Explore two ways to use ChatGPT for code: leverage the brand-new code interpreter to run Python in the browser and use plain ChatGPT to write, debug, and improve code.

Enable the code interpreter in chat GPT and run python code inside the chat session. See outputs, plots, and data analysis generated on the fly with numpy and matplotlib.

Explore data analysis and visualization with the code interpreter by uploading CSVs, JSONs, or images from Kaggle, using the IMDB top 1000 dataset.

Use the code interpreter to generate and customize movie data visualizations, including histograms, scatter plots, line plots, and heat maps of correlations with matplotlib and seaborn.

Explore a UFO sightings dataset with latitude and longitude to plot a US map, using scrubbed data, pandas, and matplotlib, while noting code interpreter limitations.

Explore using the code interpreter to process images in the browser, converting pngs to jpgs, applying black-and-white filters, and automating batch workflows with zip uploads.

Learn to use the code interpreter to upload PDFs or text, extract and summarize content, metadata, and pages, and search for keywords like bicycle or cycling.

Explore video manipulation with the code interpreter and GPT-4, including zipping a video folder under 500 megabytes, unzipping, extracting two-second clips with MoviePy, and concatenating them into a video.

Learn how developers can use ChatGPT as a co-pilot for writing small, self-contained code, debugging, translating, and tutoring, while understanding its limits for full applications.

Master iterative prompts for developers, building apps piece by piece and fixing code through repeated testing. Learn to manage context windows and minimize explanations to generate reliable, efficient code.

Learn how to use ChatGPT to write basic Python functions, test code before integration, and apply it to real-world tasks like renaming lecture files for Udemy bulk uploads.

Learn to use ChatGPT to craft shell one-liners and small scripts that sum durations of mp4 videos across nested directories using ffprobe.

See how ChatGPT helps craft a bash script to track daily disk space usage, convert it to an executable, and place it in user local bin for easy access.

Observe how ChatGPT sometimes returns broken code, and learn how identifying bugs helps it fix itself, using Python batch upserts as an example.

Demonstrates converting code between languages and frameworks with ChatGPT, from curl to python requests and fastapi to flask, while learning scala and using chatgpt as a coding tutor.

Explore how to fix bugs in ChatGPT generated code by iteratively refining a bouncing DVD logo with CSS transform, correcting boundary logic, and emphasizing multi-file structure and performance.

Explore building a simple URL shortener with Python, Flask, and SQLite, guided by ChatGPT, including basic authentication, testing, debugging, and iterative problem solving.

Midjourney converts text prompts into diverse images, from photorealistic to stylized, via prompts and the Discord bot, with no free tier and plans starting at $10 a month.

Explore DALL-E 2 as a Midjourney alternative, noting its free tier and API access. DALL-E 2 yields impressive prompts-to-images but lags Midjourney in detail and control.

Compare Stable Diffusion with Midjourney as an open-source, customizable image generator; learn about DreamStudio, free credits, and running models locally.

Sign up for Midjourney through Discord using the /subscribe command, choose from basic, standard, or pro plans (basic $10/month, yearly $8/month), and learn fast vs relaxed generations.

Learn to generate images with Midjourney on Discord via the /imagine prompt, start in newbies channels, and refine four square images using prompt best practices and adjustable parameters.

Message the Midjourney bot in Discord to keep your generations organized. Use the imagine command in direct messages to generate and refine images with variations.

This lecture explains that most generated images are publicly visible in the showcase, even from private messages, and shows how stealth mode—available to pro plan subscribers—keeps images private.

Explore MidJourney's community guidelines to understand PG-13 restrictions and blocked prompts, including gore, nudity, violence, and celebrity content, and consider stable diffusion for more flexible local use.