I couldn't agree more with you on AI topic. 100% the same thoughts about it. And yeah, for people who are just starting to code, first get very deep into the language with all try and errors and a bunch of documentation read. We for sure need to understand what the code does.
Other reasons of burnouts: Agile, incompetent leadership, too much non-techies which don't understand the industry, and unaccountability of leadership. I really miss when companies could fire people for doing a bad job rather than keep people on for office politics bias.
Truth! I love merit based teams…as long as the merit is measured equally! Sometimes people are strong because they’re amazing at enabling others to be more proficient and productive rather than doing insane work themselves.❤️👊
I agree that AI will steepen the competency distribution between intermediate and expert developers. What are your thoughts on the feedback loop of LLMs being trained on generated code? Some say it'll fold in on itself considering most of whom use AI assistance are beginners who are seeing their code compile but aren't identifying elusive logical errors, polluting the data the LLM trains on. In any case, fundamentals are essential, indeed. It's because of expert's deep understanding of constraints, problem domains, and applicable solutions that using an LLM becomes more of a code review exercise rather than a problem solving one. At that level, I imagine it's far more difficult to accept code that "looks right," whereas that "close enough" allure is the primary pitfall to those (including myself) who don't know any better. Perhaps the best course of action is instead of asking for the answers from the smart classmate, who stakes no interest in anyone else's growth, is to go straight to the source and consult mentors whose goal is deliver understanding. AI is the smart kid in your classes. And in most cases, the teacher is virtually guaranteed to outperform the smart kids in their subject. It's true that it does happen that a smart kid does know better than their teacher, but it's doubtful that even the best LLMs perform at this level in any field and are certainly not interested in anyone's growth (regardless of any speculation of consciousness lol) tl;dr Learning this stuff is hard and takes time, but that investment pays dividends. AI is a tool can make it easy but at the risk of the wrong ways at the most critical times. Get better to use AI more effectively.
You said this so beautifully, friend! It’s exactly this - learn the fundamentals then speed run it with AI. It’s exactly the reason in high school math teachers will show you the long way, before showing you shortcut!
Heh Andrew I'm a backend python developer and I want to make my own SaaS base on RAG for that thing I need to learn and do a lot of things but the issue is that you know no one knows everything like the every library of AI then I search it like the thing I want to make is already exist so I use that Instead of writing my own is this is correct !
Don’t be afraid of flat out copying existing project ideas! Literally - this is how competition is had. People copy ideas all the time & put their own twist on it & are very successful!
The bottom-line has never changed though, has it? Velocity of adoption of new tools and methodologies to drive outcome has always been the root driving factor of Software Engineering. Not programming. You learn this early in Computer Science. You can even make money today if you can do x86 assembly and C optimizing old N64 games with patreon support, that just happens to be the technical skillset needed for that task (alongside marketing the project, and soft skills in building a community). Same applies to any other domain, just be pragmatic within your space. A reductive perspective and complacency on learning and adapting has been, always, big enemies of technologists. The current incarnation of it just changes names.
Super interesting take!! What about open source software that turns into successful biz? Don’t you think that start with the passion of programming rather than trying new methodology?
@ Thanks for the encouragement! Well unless you already have one made, I would love a video where you share your take on some kind of roadmap for a novice software developer to become a great one.
@@stoneyNET I’m definitely interested in doing a video on the behavior/social/personal roadmap! I try hard not to dive into the specifics of teaching how to code because it literally changes every day 😂👊 Lots of folks interested in how to actually navigate the career path in sense of climbing the ladder, 9-5, etc!
Great video, Andrew. But yeah Junior's are really fucked. Last year I did not focus on using ai, I used to scour through shadcn, next docs and official github again and again I broke it I created a new branch and understood how data tables work etc etc. Well but yea as you said right, the efficiency at which learning complex stuff has improved cause I did my fundamentals well . Now I can just write code and what visuallixing in my head ver very fast. No disrespect to devs entering this field but you need to have domain knowledge cause coding will be becoming a soft skill by the end of this decade , so having knowledge like in business, communication, medical understanding domains pain point will be crucial were you will have the competency to solve the issue faster in that domain. 😢 yeah breaks my heart,
Well said my friend! I do think we’ll start building cooler more complex things and AI won’t fully take over. Still lots of fun stuff to get our hands dirty with!👊👏
So true man, so true.
Thankfully I have been full onboard with AI since it started. I learned to run it locally too. Has been crazy.
That’s fun!! How’s it been running locally???
I couldn't agree more with you on AI topic. 100% the same thoughts about it. And yeah, for people who are just starting to code, first get very deep into the language with all try and errors and a bunch of documentation read. We for sure need to understand what the code does.
So glad this resonated with you Bohdan!❤️👊
Is there ever a year where devs aren't struggling? People say that every year.
I think every year CAN be a struggle if you fall into the trap ..otherwise it’s a great career
@@hyperadapted do you think developers won’t fall into these traps next year?👀
We've optimized to the point where we optimized ourselves out of a job
@@Thato_M-00 I don't think this is true (yet)! Just a different set of problems ❤
Other reasons of burnouts: Agile, incompetent leadership, too much non-techies which don't understand the industry, and unaccountability of leadership. I really miss when companies could fire people for doing a bad job rather than keep people on for office politics bias.
Truth! I love merit based teams…as long as the merit is measured equally! Sometimes people are strong because they’re amazing at enabling others to be more proficient and productive rather than doing insane work themselves.❤️👊
Let me summarize most important points of this this video:
1. Use AI tools in a responsible way
2. Keep learning and keep perfecting your craft
🫡🫡🫡 any big points you think I’ve missed??
I agree that AI will steepen the competency distribution between intermediate and expert developers. What are your thoughts on the feedback loop of LLMs being trained on generated code? Some say it'll fold in on itself considering most of whom use AI assistance are beginners who are seeing their code compile but aren't identifying elusive logical errors, polluting the data the LLM trains on.
In any case, fundamentals are essential, indeed. It's because of expert's deep understanding of constraints, problem domains, and applicable solutions that using an LLM becomes more of a code review exercise rather than a problem solving one. At that level, I imagine it's far more difficult to accept code that "looks right," whereas that "close enough" allure is the primary pitfall to those (including myself) who don't know any better.
Perhaps the best course of action is instead of asking for the answers from the smart classmate, who stakes no interest in anyone else's growth, is to go straight to the source and consult mentors whose goal is deliver understanding. AI is the smart kid in your classes. And in most cases, the teacher is virtually guaranteed to outperform the smart kids in their subject. It's true that it does happen that a smart kid does know better than their teacher, but it's doubtful that even the best LLMs perform at this level in any field and are certainly not interested in anyone's growth (regardless of any speculation of consciousness lol)
tl;dr Learning this stuff is hard and takes time, but that investment pays dividends. AI is a tool can make it easy but at the risk of the wrong ways at the most critical times. Get better to use AI more effectively.
You said this so beautifully, friend!
It’s exactly this - learn the fundamentals then speed run it with AI.
It’s exactly the reason in high school math teachers will show you the long way, before showing you shortcut!
Heh Andrew I'm a backend python developer and I want to make my own SaaS base on RAG for that thing I need to learn and do a lot of things but the issue is that you know no one knows everything like the every library of AI then I search it like the thing I want to make is already exist so I use that Instead of writing my own is this is correct !
Don’t be afraid of flat out copying existing project ideas! Literally - this is how competition is had. People copy ideas all the time & put their own twist on it & are very successful!
The bottom-line has never changed though, has it? Velocity of adoption of new tools and methodologies to drive outcome has always been the root driving factor of Software Engineering. Not programming. You learn this early in Computer Science. You can even make money today if you can do x86 assembly and C optimizing old N64 games with patreon support, that just happens to be the technical skillset needed for that task (alongside marketing the project, and soft skills in building a community). Same applies to any other domain, just be pragmatic within your space.
A reductive perspective and complacency on learning and adapting has been, always, big enemies of technologists. The current incarnation of it just changes names.
Super interesting take!! What about open source software that turns into successful biz? Don’t you think that start with the passion of programming rather than trying new methodology?
As a soft dev student it’s a little bit scary hearing this but I still hope for the best.
Don’t be afraid! Lots of amazing times to come ❤️❤️👊
Let me know if there are any videos you’d love to see!
@ Thanks for the encouragement! Well unless you already have one made, I would love a video where you share your take on some kind of roadmap for a novice software developer to become a great one.
@@stoneyNET I’m definitely interested in doing a video on the behavior/social/personal roadmap! I try hard not to dive into the specifics of teaching how to code because it literally changes every day 😂👊
Lots of folks interested in how to actually navigate the career path in sense of climbing the ladder, 9-5, etc!
@@_andrewpeacock That makes sense, behavior/social/personal would be super cool!
@@stoneyNET I’m on it ❤️👊
Great video, Andrew.
But yeah Junior's are really fucked.
Last year I did not focus on using ai, I used to scour through shadcn, next docs and official github again and again I broke it I created a new branch and understood how data tables work etc etc.
Well but yea as you said right, the efficiency at which learning complex stuff has improved cause I did my fundamentals well . Now I can just write code and what visuallixing in my head ver very fast.
No disrespect to devs entering this field but you need to have domain knowledge cause coding will be becoming a soft skill by the end of this decade , so having knowledge like in business, communication, medical understanding domains pain point will be crucial were you will have the competency to solve the issue faster in that domain. 😢 yeah breaks my heart,
Well said my friend!
I do think we’ll start building cooler more complex things and AI won’t fully take over. Still lots of fun stuff to get our hands dirty with!👊👏
Thank you.
Thank YOU! Appreciate you tuning in❤️👊