Thanks a lot Bran, learning a lot from you day in day out, just started my role as a dev from school in the past 2 months, watching your videos to level up 😇
Thanks Bran for taking your time to create great content, I have only 2 of your video and will say God bless you, as I am on my journey to become a software engineer
Using these models to generate code is crazy, but it’s not crazy to ask it for information about specifically what you’re working on: that’s how I’ve used it at least. You can treat it as a rubber duck that can actually respond back and maybe every now and then it gives you a good idea. It doesn’t usually listen when I blanketly ask it to not give me code unfortunately
It's not perfect yet, I agree! But it is useful, e.g. it does give me new insights and approaches to problems sometimes. But I'm always rewriting the code it gives me. In the example of the linting script, it saved me loads of time figuring out the babel docs.
Let me give you example: ask random person on the street "why use decorators in classes" and get amazed how they will be absolutely clueless about what do you want from them. Adding more information and providing a context will give you always better response from either ai or person. My take on this is "garbage in, garbage out".
Why have anything if it isn't immediately perfect from its inception, I guess eh? C'mon dawg, are you for real... The transformers paper was literally published in 2017, not even 10 years passed, and look what's come of it despite being early days.
@@branvandermeer"Prompt Engineering" is 5% of what a programmer has to know to do their job daily. The idea that this will be anything more than a flash in the pan career (as automating AI tasks is already here) is laughable.
@@FlintBits I'm saying that prompt engineering is not a title, it's a skill. Just like software engineers have to learn certain languages and have to learn test-driven development, I think prompt engineering is also a skill that makes sense. As AI's get better; they're going to save us more and more time, and they're going to increase quality of our work, so that 5% might be 10% in a few years, and might increase from there. That's why I think it's important.
Very interesting video. I will try and dedicate time next year to learning more about this from the resources you've provided
Thanks a lot Bran, learning a lot from you day in day out, just started my role as a dev from school in the past 2 months, watching your videos to level up 😇
Thanks Bran for taking your time to create great content, I have only 2 of your video and will say God bless you, as I am on my journey to become a software engineer
Thanks for your kind words! What would you say is your biggest hurdle to take in becoming a software engineer?
thank you.
Using these models to generate code is crazy, but it’s not crazy to ask it for information about specifically what you’re working on: that’s how I’ve used it at least. You can treat it as a rubber duck that can actually respond back and maybe every now and then it gives you a good idea. It doesn’t usually listen when I blanketly ask it to not give me code unfortunately
I actually use ChatGPT to generate code, albeit in a very specific way. My next video will be about that ;)
Why do we have AI when we have to trick it into giving useful answers?
It's not perfect yet, I agree! But it is useful, e.g. it does give me new insights and approaches to problems sometimes. But I'm always rewriting the code it gives me. In the example of the linting script, it saved me loads of time figuring out the babel docs.
Let me give you example: ask random person on the street "why use decorators in classes" and get amazed how they will be absolutely clueless about what do you want from them. Adding more information and providing a context will give you always better response from either ai or person. My take on this is "garbage in, garbage out".
Why have anything if it isn't immediately perfect from its inception, I guess eh?
C'mon dawg, are you for real...
The transformers paper was literally published in 2017, not even 10 years passed, and look what's come of it despite being early days.
😍
Prompt Engineering 🤣🤣🤣
It's not my term, it's the industry's term. 🤷🏼♂️
@@branvandermeer"Prompt Engineering" is 5% of what a programmer has to know to do their job daily. The idea that this will be anything more than a flash in the pan career (as automating AI tasks is already here) is laughable.
Also also quit shitting on real Engineers.
@@FlintBits I'm saying that prompt engineering is not a title, it's a skill. Just like software engineers have to learn certain languages and have to learn test-driven development, I think prompt engineering is also a skill that makes sense. As AI's get better; they're going to save us more and more time, and they're going to increase quality of our work, so that 5% might be 10% in a few years, and might increase from there. That's why I think it's important.