Why Not Include AI

[00:00:00]
>> ThePrimeagen: And so why no AI is part of this course. You probably noticed that I haven't talked anything about AI thus far, which, like developer productivity, AI. Well, I have two reasons why I didn't talk about it, the first time I did this course, I actually did show off co-pilot.

[00:00:12]
This is, like the week co-pilot came out. I also talked about it in this course because they gave me access to it right away. And it was a lot of fun, it was such a cool experience using it for the first time. I still remember to this day going like, if player one loses, console player one lost.

[00:00:28]
I'm like, and I'm like, and then it's like, else this, I'm like, wow, that's a magical. And then I hit Enter. And it's also, else tie, and I was like I didn't even think about that, I remember just thinking, wow, that's like, the most magical thing I've ever used in my entire lifetime.

[00:00:44]
But remember, at that point in life, there was none of that beforehand, and now there's just something that's like, spewing out, sometimes accurate things and like, that was magic to me. And so I have this whole theory about AI in general right now, what's happening? If you listen to anyone talk right now on the internet, just anyone, ask them about how do you feel about ChatGPT4 versus cloud.

[00:01:06]
And what will they say? Man, cloud is so good, ChatGPT4, it's just ass, it's just the worst thing I've ever, it's just so bad. It's just like, dude, two years ago, we just ChatGPT3, and it was the craziest thing you've ever seen. What's happening? And that's what I like to refer to as the, what's it called?

[00:01:23]
The inevitable decay of amazement, which is, we had nothing. So GPT3 was incredible and you found it to be the most amazing thing ever. If you were to use it today, you'd say it's completely terrible, but why was it so great back then? It's cuz there was nothing else like it beforehand, and so just felt so amaizing.

[00:01:39]
And every release feels slightly less amazing, and you feel slightly less like captivated by it. And now we're just arguing over these really fine tuned differences between cloud versus GPT4 versus one versus whatever. And inevitably, when GPT5 comes out, I'm sure it will be so much, much better, but it still won't have the same level of amazement.

[00:01:58]
Right here, and we'll still be arguing which one is better all along, they only getting slightly actual better on their error rates. There'll still be a bunch of errors. So, I didn't really want to talk about it for that one reason and number two, the second one is that a lot of junior devs when they're starting out.

[00:02:11]
They replace their ability to learn with AI. If you are using AI in such a way that you're not asking If you had to say, if you wanted to make a game and you're like, hey, make me Tetris. You didn't learn anything, right? If you're like, hey, how do I paint a square?

[00:02:26]
You take painting a square and you make your own block. I don't, I don't think AI is bad at all for saying, hey, show me how to make a square so I can make whatever they call those poly he drawn, whatever I for, there's like a certain name for Tetris things.

[00:02:37]
I forget what they're called. So you can make your own, right? So many people jump forward now. And I talk to a lot of college students and they all are saying the same thing, which is I'm the only person that doesn't use it and I'm always the only person doing any of their projects.

[00:02:50]
Everyone is copying me on the labs now. Everyone's doing this, cuz if they can't immediately get the answer from the AI, they don't know what they're doing anymore. So it's like, don't become learned helplessness. This is why I'm not showing the tools. I just don't think they're great unless if you really know what you're doing, and at that point, almost always, you don't need them.

[00:03:07]
Other than, like, I don't know how to do CSS center, the text flex box. What are the four properties I need? Okay, there they are, I could have just used Google or AI, just I don't have to click through three links to get to the thing, or have to save 9000 ads.

[00:03:18]
So I'm just choosing AI to avoid ads, right? It's a little bit different, and so I didn't wanna talk about that. And lastly, the products are just moving so fast. When I first did this one, it was just Copilot. Now, how many different AI things can you do?

[00:03:32]
You can even run your own with Facebook Llama. You can have this, you can have all these mini ones. You got Cursor, you got Strawberry. You got all these different ones all popping and so it's like anything that I show you today will become deprecated and irrelevant within like the next three months.

[00:03:46]
All the tools that I showed you now have been around for the last 30 years, 40 years. Grep, I think might be even 50 years old. And so it's like, these are really old, well-tested things that are going to continue to be around probably for the next 50 years and just keep on being useful to you.

[00:04:01]
So I wanted to keep it focused on that, cuz I didn't wanna get distracted with things that feel shiny and cool. And maybe they do help you and make you more productive. If all you're doing is building Laravel endpoints. Yeah, they're pretty cool, because every Laravel endpoint effectively looks the same.

[00:04:16]
So you can just go pop, pop, pop, pop, pop, pop, dun, dun, dun, dun, dun, this feels Feels great, this is fast. I can see why you like AI. But if you're building anything bespoke at all, even remotely bespoke, the speed and enjoyment you get from it just goes drastically down.

[00:04:29]
And so I didn't want to go over that stuff, cuz I just find it largely uninteresting in its current state. Maybe in like five years, it'll be super cool. And my whole tune about everything will completely change, willing to eat crow. Totally willing, but right now, there you go, there's my only AI rant.