Make the best of GPT-5 potential as a coder by preparing for it ahead of time

Transcription for the video titled "Make the best of GPT-5 potential as a coder by preparing for it ahead of time".


Note: This transcription is split and grouped by topics and subtopics. You can navigate through the Table of Contents on the left. It's interactive. All paragraphs are timed to the original video. Click on the time (e.g., 01:53) to jump to the specific portion of the video.


Intro (00:00)

I recently came across this tweet which stopped me on my tracks. This gentleman who was present at Y Combinator's tool kickoff event heard Sam Altman suggest to people to build with the mindset of GPT-5 and that AGI will be achieved relatively soon. Most GPT-4 limitations will get fixed in GPT-5. So in this video, I'll try to explore what this will mean for the world of coding. And I'll try to inspire you and excite you on why this year is the best year to start learning how to code in Python and using GPT API and also do our best to improve on it.

Ai Applications And Learning Platforms

Using AI for Programming (00:30)

I had some preliminary thoughts on this, which I tweeted. I said, I'm predicting that the arrival of GPT-5 will be the first time the coding will have fundamentally transformed. GPT-4 was the first time we can meaningfully use AI assistance while coding, but I feel like GPT-5 will, by eliminating most of the limitations GPT-4 allow us to become true architects and design and build so much more complex applications without much dealing with the actual writing of the code itself. Based on this, I suggest that the single developer becomes a team and meta learning of concepts of what is available possible becomes more important than detailed knowledge about them. What I mean by this is that after the arrival of GPT-5, the idea that a single developer will be able to handle the task that an entire team can handle previously has become a reality. So therefore, a single developer will become like a project manager or a CTO who has access to dozens of expert coders who know the details. In this aspect, knowing what each expert does within a larger project is more important than knowing how to do that in detail yourself. Knowing the general capabilities of each framework library package necessary for your project will allow you or one or I to use GPT-5 as that expert for that part. By saying this, I'm not referring to some agent-like system in which multiple AI-powered agents do this work, but I mean this is with a single dev in the loop, knowing exactly the general aspects of what a programming language is capable of, and each one of its framework libraries or packages that are useful for the project at hand. So a general I'm trying to consider that a single developer needs to know the general landscape of the land, the general possibilities, and which programming languages, libraries or packages allow for that to happen, rather than to directly mess it themselves strictly. That's why this is tied to the second point metal learning of concepts. What I mean by that is that the bird's eye view of programming paradigm you're building with will be more important than spending weeks mastering it. For example, knowing what levels of abstractions you can have for your models and views in Django, Django being a backend framework for Python, in general will be more important than knowing how to exactly code them yourself. This brings me to my point that knowing or understanding a language in general is going to be very important, especially if you're just starting to learn Python. But this also should apply if you just want to get better at it that understanding what python offers in general what its capabilities are is going to become more important than to actually be able to write the syntax perfectly yourself. That's why I'm proposing that if you're beginning to learn Python just for the first time, spending about eight hours chatting with ChatGPT, completely focused on understanding the fundamentals of the language is going focused on understanding the fundamentals of the language is going to actually be sufficient to get you start coding right away. So, talk to ChatGPT for the Plus version, right?

Learn Python in 8 hours (03:20)

Not 3.5. Talk to chat GPT about learning Python for eight hours straight, fully focused, don't cheat, just absorb as much information, keep asking questions. Now if you're not able to do it 8 hours straight, I just set this as a general goal. You can do it 2 hours a day, whatever, but I'm just setting some kind of standard which is achievable quickly. I do believe that this should give you a general understanding of Python. Now I do also say that have a paid course, YouTube series or pre-Python learning website as a reference. Any decent one will do. For example, there's a really good course from W3Schools that you can use. But I have also built one for a full Python course using AutoStreamer. I'll talk more about AutoStreamer at the end of the video, which gives you the general capabilities of python in a single website with audio history of python interactive python basic idea is to start chatting with gpt4 asking teach me python from scratch but having a structured reference course such as the Learn Python Fast course I built so that you know exactly what you need to learn. So you can ask more in-depth questions about functions or modules and packages or object-oriented programming. I'm proposing that a full eight-hour chat with ChatGPT, along with a structured course, should give you the understanding you need to start coding with Python using AI assistance. And the rest of the learning will take place while you're building the apps. And the mastery will come by continuous chatting with GPT and AI assistance while building apps yourself.

Live Stream (04:51)

To demonstrate how and why this is possible, I'll be hosting a live stream tomorrow at 9am Los Angeles time. During this exact process that I'm suggesting, I won't go for eight hours, but I will take a good few hours to demonstrate and display how I would go about it if I were to start learning Python right today. I will also outline how to actually, if you already know Python, how to get better at it using the same method. I'll drop a link to this in the description. The message that I'm trying to convey is that you don't have to spend weeks or months trying to master Python from scratch, all you need is just the general understanding of it. And if you're willing to use AI assistance, you can start building useful apps using Python right away. And the second part is simple too. First, you have to accept that you need to use AI assistance and in preparation for GPT-5, this is the best thing you can do because talking to AI assistants, whether it's Copilot, Cursor ID or ChatGPT Plus will give you the necessary practice and the ability to be able to communicate your thoughts properly to build the apps that you're looking to build. The next question to ask is which AI assistance to use. I have tried many and there's a bewildering array of options which you can choose from. But I've settled personally on a fundamental three, the GitHub Copilotilot which allows you to do code completions within the visual studio code or cursor ide definitely cursor ide which is pretty much very similar to Visual Studio Code, but it is the AI first code editor. And of course ChatGPT Plus for GPT-4. I think these three is all you need to start building useful apps and to bring your vision to reality. After spending the eight hours learning the fundamentals of Python, all you have to do in the second day and every day after that is to think of an idea and try to code it with ChatGPT, Copilot, and Cursor. And I will demonstrate how I use each one of them in the live stream tomorrow. Now, the important part is to ask about every little error message and every new concept that pops up. Keep asking and probing incessantly. Do this for eight hours a day, every day. Now, if you don't have time to do this and break it apart, it doesn't matter. Do it one hour here and one hour there, but just make sure to keep doing it continuously with determination. This is pretty much it. As long as you have the determination, you should become a novice app builder in 10 days, an intermediate in a month, and an unstoppable force of nature in about six months and this is especially true when GPT-5 is released. If you decide to join me in the live stream, or you can watch it afterwards as well, I will try to demonstrate what is possible currently and what might what might be possible with GPT-5. Try to inspire you and excite you about what you can actually start doing right now, today. Now that brings me to the final point. Communicating your ideas clearly to GPT will be the number one skill. Since more and more we will only need to have general knowledge rather than specific, and to be able to communicate the needs of your project to GPT clearly will be more important. It will be like leading a team of half a dozen expert developers in different fields. The better we can communicate the plan and trajectory of the project to GPT will be the superpower in post GPT-5 world of coding. Imagine yourself like an architect which can plan, design, and lead an entire mega architectural project, construction project, without laying the brick one by one itself.

AI Architect (07:57)

All this being said, just as the architect in our example had to learn different materials, construction techniques, or styles of building, it is essential that you understand what tools are available to you, and to collaborate with AI assistance with what you know and continually expand and enhance what you know by continually studying and learning by building and creating useful and interesting apps i'll put a link to this course in the description as well which brings me to AutoStreamer which I'd like to take a moment to talk about.

AutoStreamer (08:30)

AutoStreamer for me started as a fun project. I was trying to see how much of live streaming I can automate while also being useful teaching educational content. AutoStreamer allows you to create content in real time, which allows you to record, learn, first of all, learn yourself or record your videos and share with people later, or create live streams, and it automatically builds a course website while it's doing all this. I put a link to a live stream, I have that, feel free to watch that. But in this live stream, we were building this course and this course was built in this webpage one by one while also playing the audio history of Python one by one. And at the same time, creating this website, which is fully deployable. And as a matter of fact, I have deployed it at railway.apps at the end of the live stream. I actually show you how you can deploy it as well. Currently it's available only at my Patreon shop for a limited time for $200. If you need more information, please read the Patreon post and watch the live stream that you can have access directly from the website. It's also and I'll put a link in the description as well. Feel free to also contact me at Discord if you have any questions about it.

Patreon (09:44)

I also would like to talk about my Patreon membership. Currently, there's over 200 interesting code project files and some exclusive walkthroughs for patrons, and I will be creating more exclusive content for preparing for the release of GPT-5. I have spent over 2000 hours in 2023 building these projects. So by becoming a patron, you have access to all the code files, which you can download and get inspired by. Just simply just go to the post and find the code download things at the bottom and download them and just run them get inspired you can also provide these as examples like few shot examples to GPT-4 and in the future GPT-5 to allow it to create more interesting apps.

Project Echohive

EchoHive (10:33)

I also created an entire website,, where you can actually watch every one of my projects for free. It also has the code download links for patrons for convenience. Just find the project you're looking for and go to its post and find the code download links at the bottom. It also has a community search. For example, if you're looking for embeddings, if you're looking to learn more about embeddings, you can search for that. If you're learning to want to learn more about link chain for example, it is a real time search. But you can just simply browse. It also has a shuffle function you can use. Just click on that and it shuffles. So if you're a patron and just want to experiment with some exciting projects, this is a good way to find what you're looking for. And if you do decide to become a patron, I just would like you to know that. I really appreciate it. And thank you for your support. I hope you found this video useful. I hope to see you in the live stream. And I hope to explore this topic much more in upcoming videos. Thank you for watching and I'll see you in the next one.

Could not load content

Great! You’ve successfully signed up.

Welcome back! You've successfully signed in.

You've successfully subscribed to Wisdom In a Nutshell.

Success! Check your email for magic link to sign-in.

Success! Your billing info has been updated.

Your billing was not updated.