Parker Rex
All videos
Parker Rex

ChatGPT Noobie to Pro in 7 Minutes (2025 Updated)

March 8, 2025

Watch on YouTube@parkerrex

A practical, updated quick-start to go from ChatGPT newbie to pro in 7 minutes, with tips you can apply to business workflows, prompts, and Mac integration.

Quick start: what to use and how#

  • You’re on the paid version. GPTs (custom bots) aren’t essential—they’re often overrated.
  • Projects are the core: a collection of your uploaded files to train ChatGPT plus a space for custom instructions.
  • Example: build a project with your work samples and a set of instructions that guide how to turn your input into a polished output (e.g., a blog post from your story).
  • Use Projects to centralize your context and accelerate repeatable outputs (like drafts for common business tasks).

Prompts: how to structure for best results#

  • Prompts are what you tell the model. More context usually means better results. Four levels of prompt detail:
  • Level 1: Basic prompt (e.g., write an email to my boss about a raise). The model knows little about you.
  • Level 2: Add context (you’re a mid-level marketer at Company X, with specific daily tasks).
  • Level 3: Add work examples stored in the project (your past outputs, results, metrics).
  • Level 4: Add persuasive-writing examples and successful emails (from real-world wins) to guide tone and structure.
  • Practical takeaway: provide as much relevant context and examples as you can to nudge the model toward your desired output.

Model types and when to use them#

  • Two model paradigms:
  • Chain-of-thought / deeper reasoning models: slower, better for complex, multi-step problems.
  • Chat models: fast, back-and-forth iterations for drafting and quick turns.
  • Parker’s guideline:
  • For complex tasks, start with the deeper-thinking option (01).
  • For quick pings and drafts, use the faster option (03 mini or GPT-4-like).
  • Core idea: tailor the model to the task complexity and desired turnaround time; provide ample context regardless of model.

Tools you should know about#

  • Search: like a Google inside the chat, but not ideal for deep research.
  • Not available on chain-of-thought models; use a chat model if you need quick lookup, but be skeptical of accuracy.
  • Deep research: go-out-and-back-with-sources mode.
  • Best for papers, reports, citations, and links. Returns sources and evidence you can cite.
  • Image: generate visuals when needed (cartoony style, but handles quick visuals).
  • Canvas: a connected document editor that opens alongside ChatGPT for editing workflows.

Mac app and the Companion feature#

  • Companion is a key productivity boost:
  • Bound to a hotkey, keeps context on hand without leaving your current task.
  • Drafts and prompts you can refine on the fly.
  • How to use it effectively:
  • Draft a raise email using a targeted prompt, then use snippets to produce variations.
  • Snippets are reusable prompts/templates. Example: a concise prompt that outputs multiple variations.
  • Snippet strategy:
  • Use a concise prompt (e.g., “GPT key concise”) to trim output.
  • Generate multiple variations:
  • Variation 1: direct and professional
  • Variation 2: warm and appreciative
  • Variation 3: crisp and authoritative
  • Variation 4: confident and results-focused
  • Practical caveat: the Mac app is great for focus sessions, but don’t rely on it exclusively for heavy, day-to-day work—use the main app for longer work sessions.
  • On-screen context: Companion can show auto-suggestions and even take screenshots of your current screen to inform prompts.
  • Quick tip: you can start typing in Chrome or other apps, and Companion will offer app-specific prompts or context-aware suggestions.

Practical takeaways and best practices#

  • Build a strong project: store your best work samples and your custom instructions in one place.
  • Always add context: the more you tell the model about who you are and what you’ve done, the better the output.
  • Use deep research for credibility: when you need sources, evidence, or citations, use the deep research tool.
  • Leverage Canvas for documents: keep drafts and edits organized side-by-side with ChatGPT outputs.
  • Use Snippets to scale: create a library of prompt templates and tone styles; remix outputs quickly.
  • Use the Companion for focus: don’t rely on it for every task, but use it to accelerate repetitive workflows and rapid drafting.
  • Don’t overdo it: if a draft is too long or verbose, trim with concise prompts or request variations focused on brevity.

Next steps#

  • If you want more depth, Parker suggests a Master Class on the channel to level up further.
Transcript

here's everything you need to know about using chat GPT in 3 minutes to be competent on Parker X I led a product team at a startup that did $50 million before we exited after that I chased the Silicon Valley unicorn status for three years until pivoting to going all in on a the whole point of this channel is to just cut the fluff and get into what actually works for businesses so on the left we have Chachi PT on the right we have our little curriculum so first of all when you jump in we're on a paid version of this you'll see a couple things there's gpts I wouldn't worry about using these they're kind of a flop but if you do click on it you'll see a bunch of different ones that people have made and some of them are helpful but for the most part not great instead you would use a project a project is a collection of different files that you've uploaded to train chat GPT on as well as a place to make custom instructions so if I were writing an email to my boss trying to get a raise I might put a project together with all the files of all the work that I've done all the the um projects that warant why I should get that raise as well as some custom instructions I can go into one in here just to show you where I have a storytelling to blog custom project and then here I have a set of instructions that it needs to follow so that I can just talk to this project and it'll take a story that I tell and then turn it into a Blog so that's that on the prompts this is what you actually tell the computer it's what you write and the more information the better so when you write a message that's a prompt and I want to kind of explain like the different levels of it so the different levels might be hey write me an email to my boss about why I need a raise well guess what this thing doesn't know a whole lot about you so the more information you can give the better level two might look like I'm a mid-level marketer working at X company doing y daily tasks you fill that stuff in obviously and I'm writing on my boss to get a raise this is why I think I need the raise and then you tell a little bit more level three would be all of that plus the examples of the work you might do that in the project and then level four would be perhaps adding examples of persuasive writing and examples of successful emails that you might have pulled off Reddit or something so there's a trend where basically the more information you provide the better and that's how a prompt will work now there's two different types of models a model is basically kind of the the engine or the brain behind the uh the AI and so when you open up a new chat you'll see all these things right now the ones with the GPT in front of them you can basically just split them into chat models and Chain of Thought models Chain of Thought models take longer they're thinking it's a chain of thoughts that it's having and it'll take a bit longer but it's for advanced reasoning I always go with 01 so if you're doing something that is complex that you wanted to think a lot about then you use o1 chat models are the Ping ponging back and forth real quick you probably use o03 mini or GPT 40 and for the purposes of of this video that pretty much covers it um but yeah you always want to basically provide more than uh than you'd think when you're when you're talking to these so next up we have a couple tools so we have search which you'll notice that it's not available on the Chain of Thought models but if I did a search with a chat model you can think of it as like a version of Google it's not very good I don't recommend using this just use Google when it comes to deep research this is actually really good at going out and if you're writing a research paper and you wanted a ton of information on here about um links support of evidence research it would go out and do that so that's what you would use deep research for and it would come back with all the links and a ton of information other tools that are in here image you can generate images with this they look kind of cartoony um but it does the job and then a canvas would be where it tags this tool called canvas where it will open up a document editor right next to it so uh lastly I want to just quickly touch on the Mac app so the really The Shining factor of the Mac app is the fact that it has what's called a companion so so if you notice this thing at the bottom just opened and that's the companion so it's tied to a hotkey and then I can ask questions so it's really so I never lose context of what I'm working on I can just bring it up so if I did have let's say um draft a email to my boss I'm a mid-level marketing person and I want to ask for raise I've done a lot of good work that has generated a lot of profit make a draft for that so I'd send that off and now I'm going to get an example that I can then run a uh prompt on so I'd get this it's just okay and then I can grab the example and as you can see because it wasn't very specific it's like super wordy so you'd never want this so then I would use a snippet you're always going to want to use these Snippets and I'll put them in the description but I do GPT key concise and it will have this text replacement and then I can paste the response that I got in and now it'll come back with a bunch of different variations so now I get variation one is direct and professional kind of chops it down a little bit variation two and be warm and appreciative and then I think variation three is crisp and authorative authoritative but I'm not sure we'll see in a second confident and results focused okay cool so that's kind of the beauty of the companion now if I popped this button it would open up the Mac App but again I don't recommend it for using it day-to-day like if you're actually doing a lot of work in a dedicated Focus session then use the actual app and um besides the companion part there is this where it can work with the different apps so you can connect it and it'll actually have context of what's on screen some other stuff that most people don't know actually so if you just start typing in um let's say if I open a new chat I'm gonna open the companion no I want a new one and I start typing Chrome it'll start giving me auto suggestions on what I want to write um sometimes it'll show the app too so if it's an app that doesn't technically work with it'll show that but then you can click it it'll take a screenshot of what's on the screen I know we ran a little bit over but I hope that's a good I hope that's a good starting point for your journey with chat gbt if you want to learn more about it there's a master class on my channel and you can learn a whole bunch on there until next time