1. ) AI will increase demand for software = more jobs 2.) You still need one person who operates the AI chatbots 3.) You still need coding skills to operate the AI chatbots (right questions, validation of answer) 4.) Software Engineering is maybe 20-40% about coding. The rest is Analaysing bugs( AI can't do that in a full project alone), Analysing and defining requirements with the customer (AI cant do that), communication with customer and other stakeholders of the process (AI cant do that), documentation of steps, Deployment of software (CI / CD) which an AI chatbot cannot do (alone) Based on that, i think salaries could even increase on the short term for single engineers, since they could be more powerful and faster than before and deal with increasing demand! on the long run, prices per day for software development maybe fall, since it is "easier" to get things done so the single engineer keeps the same salary but has to deliver more output to compensate falling prices. let s see how it turns out. but 1 thing is clear, software engineer won't go anywhere the next x-years! What i think is that hard skills will be more "replaceable" by tools (like it always was happening in the past). but only the non complex/non creative skills! and soft skills wont be replaceable anyhow. so to be safe, learn soft skills and lean towards project management and people managment! communication skills will be more and more the key to success!
I have been in and out of so many rabbit holes with TH-cam on AI tools and I like your style and presentation. I will be in a Volvo rabbit hole now going through your other vids! Please keep up with your content. I look forward to your AI CODING 101 course. Thank you
Your AI-related content is amazing and truly valuable! I really appreciate that you're not just creating low-quality AI content for the sake of making money. Here's a bit about my experience: I'm a frontend developer with solid backend knowledge since I started as a full-stack developer. I've been using Cursor for two months now, and I've noticed something important. The challenge in development has never been just writing code-it's about thinking, analyzing, and solving problems. However, a concern I have is about newcomers to the field. With every piece of information handed to them so easily, they might not develop any real problem-solving skills. What will happen to them in the long run? It's an open question. Thank you for creating such thoughtful and meaningful content! Keep up the great work.
This made my day, thank you! Yes, I came to that same realization as AI was making me faster! Hopefully AI Coding 101 will help newcomers make the most of the tools while also learning problem-solving skills!
Technical PM but also inevitably debugging. So true, building 95% of the app is doable, but adding one feature, sometimes a vital one, takes hours with AI.
I think startups will have less need for developers and will be quick to adopt LLMs for coding. However, some industries, like Health and Finance, are still very weary of using LLMs for coding due to privacy and security concerns. If I were looking for a job in software development right now, I would look in hospitals, health insurance, banks, finance, etc.
Yes, I think this is very on point! I think we will see many more *small* teams with low funding, addressing many small problems/opportunities. It's an exciting time to be an entrepreneur IMO. And for more stable jobs, I think you nailed it - being a software engineer and blowing everyone's mind at a 'traditional' business like healthcare, finance, construction, manufacturing, etc etc is a great opportunity right now.
system design/architecture are extremely crucial when you work with AI. I wouldn't say you need to know how to code to be able to create something meaningful with AI. But you should def know how to talk like a technical PM. Looking forward to your new way of learning to code course.
Yeah I agree - understanding system design and architecture becomes important much sooner than it used to. And in many cases, talking like a technical PM may be enough to build some products. I think it's a bit of a scale and ultimately depends on how technical your product is. Thanks for watching! I'll be releasing the AI Coding 101 videos starting in the new year!
Hey, thanks for pointing that out! Could you make the course one long video? Weekly releases'll get us lost in all the daily AI news; it's hard to keep up. Hope you understand. Good luck
Thanks for watching! It will be a pretty detailed and long course so I plan to release it over time as it will take me a long time to film and edit the whole thing. Perhaps when I am done I can create a compiled version!
You need to know how to code to create an effective plan for AI assistance in designing your apps. Additionally, due to the limited output window of LLMs, their application is mainly suited for basic apps, like the examples often shown by TH-camrs. In real-life scenarios, I have encountered many challenges that LLMs cannot resolve.
Well said Micbab. I will plan to include real work on a medium/large codebase a part of AI Coding 101 to show how to overcome some of those challenges when the code grows too large for LLMs to easily handle.
Just subscribe! The course videos will start going up in the new year. I'll also create a discord community where you can ask questions and such! Stay tuned :)
@@VoloBuilds Thank you. I am an absolute struggling newbie. I signed up for a phyton class on Udemy and a few more. I have some IT certs. I know how to prompt fairly well due to my linguistic background but coding is still abstract to me and I wanna ask a coder, "Does it matter if a line of code is indented or not?" Nobody can ever answer me that. I'd like to take a step by step class like the one you described.
AI is just "Autocorrect on steroids" and it cannot think outside the box. It can just replicate what all other humans already did and remix it but it cannot create something new.
It is a very interesting topic. The question I would ask is: are *people* able to create "something new"? I would argue we are similarly remixing data that we have been exposed to. I explored this further in this video: th-cam.com/video/l-9EUBbktqw/w-d-xo.html - would love to hear your perspective on it!
I think they are getting caught up on what coding "is". In my mind, "building software with prompts" is still coding. You are still specifying how you want a piece of software to work by defining rules for it to follow. Knowing the right words/phrases and how software works helps you do this much more quickly.
@ Replit AI Agent will suggest it and implement it. It can dig a hole that you may have to crawl back out of depending on your use case. It’s a personal example I have learned firsthand about what you’re saying. It’s good to educate yourself because the AI may end up doing something that doesn’t align with what you intend. The DB layer isn’t something a non coder may be looking at on day one so it can easily be missed/overlooked during implementation
@@jonasRaymondl you are also right - eventually the price will go down for sure. But how long is eventually? Either way, the best thing to do is to learn how to use these tools and direct them to do what we want. Currently, learning coding or at least system architecture is an easy and effective way to guide the models.
Layoffs are a small part of the bigger puzzle. Many of the tech layoffs were due to tax code changes. Also, I expect a huge shift from large companies to individual makers or small teams since things are much easier to build quickly and people are building software outside of traditional corporate roles. Look at the growth of the indie hacker community as an example. If you believe any of the official stats, software engineering jobs are projected to grow significantly (www.bls.gov/ooh/computer-and-information-technology/software-developers.htm) but I personally don't put a lot of weight into this. I don't think we can accurately predict anything past a year or two. Knowing how to code or at least how software works will enable you to take advantage of the current AI wave more than anything else.
Haha for some reason this seems to be the preferred course of action for a lot of people - not exactly sure why. So much better to take advantage of the opportunity, build your skills, and finding meaning!
If you have topic suggestions, let me know! If you mean sponsorship kind of stuff, I am not taking on sponsors at this time but appreciate any "SuperThanks" contributions!
I'll just learn COBOL, it survived rise of internet, rise of enterprise java, and it will survive this AI revolution you guys are experiencing right now.
I'm a retired software engineer. The new frontier models are already superhuman at some coding tasks. I can't imagine it will take longer than 2-5 more years before human labor is worthless.
Yeah I use Sonnet 3.5 for 90% of my coding and it gives me a massive speed boost. As a result I'm able to get more done and solve problems more quickly. But there is an endless pit of problems to solve - and we are in a position to learn how to direct the AI to solve the problems that matter most to us. So I think it's very valuable to learn how to best direct the AI. Jumping on this stuff now makes it possible to take advantage of the shifting landscape - not to mention avoiding the nihilistic depression of assuming everything is pointless.
I retired a senior software engineer. If you are having problems getting AI to produce high quality code, the issue is in your prompting, NOT the AI models. As shown in the tests, the newest models can already code at a competition level. That has been my experience both with Claude, and O1.
This is true for a lot of the cases. The interesting thing is that knowing how to code dramatically improves your prompting! Most people who don't understand software won't even know what to ask. This is what makes learning to code valuable in the future.
@@Steve-xh3by I see that English is not your native language, and maybe that is why you cannot see the difference: "You retired a senior SWE" means you made someone else retired ... Like you forced them into retirement... "You are a retired senior SWE" is a statement about you .... It means you retired ... Where are you from ?
so is it really worth it to pursue coding right now? i am seriously don't know, this is my passion and also i want to do it for money so i will learn it very seriously, could u please let me now please
Here is my take , AI is already replacing devs , start ups need less workforce to excel. the demand for devs will fall significantly. and eventually AI replaces every aspect of development . If they need a single dude to prompt AI instead of 5 software devs , that is called "REPLACEMENT"
While this makes sense on the surface, it does not consider the full market dynamics caused by the reduced cost. It also misses the significant precedent we have of other software automation and improvements adding MORE demand for software. I made a whole video talking about the economics of this (Jevons Paradox) and why I think AI will counter-intuitively create *more* software jobs: th-cam.com/video/56H0n98V6hg/w-d-xo.html
There is no Ai that can just code on the basis of prompting, as Volo said he spent two hours to troubleshoot, and it didn't fix anything. I spent 12 hours to troubleshoot, and the code gets worse. We are far from Ai taking over any jobs during this time. It can happen maybe tomorrow, or it can take ten years or hundred we just don't know what'll happen. But big corporates like Nvidia, Microsoft, Google or Apple, will not allow using these Ai models to code everything until they are sure its fully secure.
thats zhe one opinion on that. i think, demand will increase! because now more software products can be build at a lower cost ! back then when google was introduced, the demand increased also ! it is the same eith chatgpt now..its google on steroids
Never listen to a CEO whose company has thousands of developers saying coding is dead.
The only genuine honest take on AI and coding on the internet! Thank you 🙏
:) Thank you very much!
1. ) AI will increase demand for software = more jobs
2.) You still need one person who operates the AI chatbots
3.) You still need coding skills to operate the AI chatbots (right questions, validation of answer)
4.) Software Engineering is maybe 20-40% about coding. The rest is Analaysing bugs( AI can't do that in a full project alone), Analysing and defining requirements with the customer (AI cant do that), communication with customer and other stakeholders of the process (AI cant do that), documentation of steps, Deployment of software (CI / CD) which an AI chatbot cannot do (alone)
Based on that, i think salaries could even increase on the short term for single engineers, since they could be more powerful and faster than before and deal with increasing demand!
on the long run, prices per day for software development maybe fall, since it is "easier" to get things done so the single engineer keeps the same salary but has to deliver more output to compensate falling prices.
let s see how it turns out.
but 1 thing is clear, software engineer won't go anywhere the next x-years!
What i think is that hard skills will be more "replaceable" by tools (like it always was happening in the past). but only the non complex/non creative skills!
and soft skills wont be replaceable anyhow. so to be safe, learn soft skills and lean towards project management and people managment! communication skills will be more and more the key to success!
You nailed it!! Completely agree
@@VoloBuildsthanks🎉
Until you have 10x context scaling and pooling of agents.
Keep up the great work Volo!!!
Thank you!!
I have been in and out of so many rabbit holes with TH-cam on AI tools and I like your style and presentation. I will be in a Volvo rabbit hole now going through your other vids! Please keep up with your content. I look forward to your AI CODING 101 course. Thank you
Haha that's awesome - enjoy the videos! Let me know if there's anything specific you'd like to see in AI Coding 101!
Your AI-related content is amazing and truly valuable! I really appreciate that you're not just creating low-quality AI content for the sake of making money. Here's a bit about my experience: I'm a frontend developer with solid backend knowledge since I started as a full-stack developer. I've been using Cursor for two months now, and I've noticed something important. The challenge in development has never been just writing code-it's about thinking, analyzing, and solving problems.
However, a concern I have is about newcomers to the field. With every piece of information handed to them so easily, they might not develop any real problem-solving skills. What will happen to them in the long run? It's an open question.
Thank you for creating such thoughtful and meaningful content! Keep up the great work.
This made my day, thank you! Yes, I came to that same realization as AI was making me faster! Hopefully AI Coding 101 will help newcomers make the most of the tools while also learning problem-solving skills!
Technical PM but also inevitably debugging. So true, building 95% of the app is doable, but adding one feature, sometimes a vital one, takes hours with AI.
Yes, hearing this all the time!!
This is such a balanced, and honest take. you gained a subscriber.
Really appreciate it! Hope you have a wonderful day!
Hey Volo! Fantastic video, thanks for sharing the information. Jay
Thanks, Jay! Hope you had a very Merry Christmas and wishing you a happy new year!
@ Thank you, same to you
Excellent take
I think startups will have less need for developers and will be quick to adopt LLMs for coding. However, some industries, like Health and Finance, are still very weary of using LLMs for coding due to privacy and security concerns. If I were looking for a job in software development right now, I would look in hospitals, health insurance, banks, finance, etc.
Yes, I think this is very on point! I think we will see many more *small* teams with low funding, addressing many small problems/opportunities. It's an exciting time to be an entrepreneur IMO. And for more stable jobs, I think you nailed it - being a software engineer and blowing everyone's mind at a 'traditional' business like healthcare, finance, construction, manufacturing, etc etc is a great opportunity right now.
Thank you 🙏
Thanks, Nassim :) Hope you are doing well!
system design/architecture are extremely crucial when you work with AI. I wouldn't say you need to know how to code to be able to create something meaningful with AI. But you should def know how to talk like a technical PM. Looking forward to your new way of learning to code course.
Don't you think talking like a Sr Dev to an LLM will give better results than a PM or a PO..?
Yeah I agree - understanding system design and architecture becomes important much sooner than it used to. And in many cases, talking like a technical PM may be enough to build some products. I think it's a bit of a scale and ultimately depends on how technical your product is. Thanks for watching! I'll be releasing the AI Coding 101 videos starting in the new year!
Waiting ❤❤❤
Hey, thanks for pointing that out! Could you make the course one long video? Weekly releases'll get us lost in all the daily AI news; it's hard to keep up. Hope you understand. Good luck
Thanks for watching! It will be a pretty detailed and long course so I plan to release it over time as it will take me a long time to film and edit the whole thing. Perhaps when I am done I can create a compiled version!
@VoloBuilds I can't wait 🫡
You need to know how to code to create an effective plan for AI assistance in designing your apps. Additionally, due to the limited output window of LLMs, their application is mainly suited for basic apps, like the examples often shown by TH-camrs. In real-life scenarios, I have encountered many challenges that LLMs cannot resolve.
Well said Micbab. I will plan to include real work on a medium/large codebase a part of AI Coding 101 to show how to overcome some of those challenges when the code grows too large for LLMs to easily handle.
That is great. Thank you. Where do i sign up. That is exactly what I need.
Just subscribe! The course videos will start going up in the new year. I'll also create a discord community where you can ask questions and such! Stay tuned :)
@@VoloBuilds Thank you. I am an absolute struggling newbie. I signed up for a phyton class on Udemy and a few more. I have some IT certs. I know how to prompt fairly well due to my linguistic background but coding is still abstract to me and I wanna ask a coder, "Does it matter if a line of code is indented or not?" Nobody can ever answer me that. I'd like to take a step by step class like the one you described.
AI is just "Autocorrect on steroids" and it cannot think outside the box. It can just replicate what all other humans already did and remix it but it cannot create something new.
It is a very interesting topic. The question I would ask is: are *people* able to create "something new"? I would argue we are similarly remixing data that we have been exposed to. I explored this further in this video: th-cam.com/video/l-9EUBbktqw/w-d-xo.html - would love to hear your perspective on it!
I think those people who say AI is going to replace coders are the ones who don’t know how to code and don’t even want to spend time leading it.
I think they are getting caught up on what coding "is". In my mind, "building software with prompts" is still coding. You are still specifying how you want a piece of software to work by defining rules for it to follow. Knowing the right words/phrases and how software works helps you do this much more quickly.
Drizzle or no for database control?
Haven't tried it!
@ Replit AI Agent will suggest it and implement it. It can dig a hole that you may have to crawl back out of depending on your use case. It’s a personal example I have learned firsthand about what you’re saying. It’s good to educate yourself because the AI may end up doing something that doesn’t align with what you intend. The DB layer isn’t something a non coder may be looking at on day one so it can easily be missed/overlooked during implementation
And O3 price will be insane right?
Yes great point - I forgot to mention this - and it will also be slower than current models since it has to go through the reasoning steps.
for now , give it some time , it will be cheap
@@jonasRaymondl you are also right - eventually the price will go down for sure. But how long is eventually? Either way, the best thing to do is to learn how to use these tools and direct them to do what we want. Currently, learning coding or at least system architecture is an easy and effective way to guide the models.
Are you reading the news about layoffs?
Layoffs are a small part of the bigger puzzle. Many of the tech layoffs were due to tax code changes. Also, I expect a huge shift from large companies to individual makers or small teams since things are much easier to build quickly and people are building software outside of traditional corporate roles. Look at the growth of the indie hacker community as an example.
If you believe any of the official stats, software engineering jobs are projected to grow significantly (www.bls.gov/ooh/computer-and-information-technology/software-developers.htm) but I personally don't put a lot of weight into this. I don't think we can accurately predict anything past a year or two.
Knowing how to code or at least how software works will enable you to take advantage of the current AI wave more than anything else.
just dont code dont learn nothing just stay at home or sleep on the street..AI does everything for you.
Haha for some reason this seems to be the preferred course of action for a lot of people - not exactly sure why. So much better to take advantage of the opportunity, build your skills, and finding meaning!
Can we contribute to the course
If you have topic suggestions, let me know! If you mean sponsorship kind of stuff, I am not taking on sponsors at this time but appreciate any "SuperThanks" contributions!
I'll just learn COBOL, it survived rise of internet, rise of enterprise java, and it will survive this AI revolution you guys are experiencing right now.
Hahaha long live COBOL!
You need an engineer to perform good tasks
I'm a retired software engineer. The new frontier models are already superhuman at some coding tasks. I can't imagine it will take longer than 2-5 more years before human labor is worthless.
Yeah I use Sonnet 3.5 for 90% of my coding and it gives me a massive speed boost. As a result I'm able to get more done and solve problems more quickly. But there is an endless pit of problems to solve - and we are in a position to learn how to direct the AI to solve the problems that matter most to us. So I think it's very valuable to learn how to best direct the AI.
Jumping on this stuff now makes it possible to take advantage of the shifting landscape - not to mention avoiding the nihilistic depression of assuming everything is pointless.
I retired a senior software engineer. If you are having problems getting AI to produce high quality code, the issue is in your prompting, NOT the AI models. As shown in the tests, the newest models can already code at a competition level. That has been my experience both with Claude, and O1.
This is true for a lot of the cases. The interesting thing is that knowing how to code dramatically improves your prompting! Most people who don't understand software won't even know what to ask. This is what makes learning to code valuable in the future.
You "retired a senior SWE" or "you are a retired senior SWE" ?
@@Rami_Elkady There is no difference between those two sentences. They are functionally equivalent in the English language.
@@Steve-xh3by
I see that English is not your native language, and maybe that is why you cannot see the difference:
"You retired a senior SWE" means you made someone else retired ... Like you forced them into retirement...
"You are a retired senior SWE" is a statement about you .... It means you retired ...
Where are you from ?
so is it really worth it to pursue coding right now? i am seriously don't know, this is my passion and also i want to do it for money so i will learn it very seriously, could u please let me now please
Here is my take , AI is already replacing devs , start ups need less workforce to excel. the demand for devs will fall significantly. and eventually AI replaces every aspect of development . If they need a single dude to prompt AI instead of 5 software devs , that is called "REPLACEMENT"
While this makes sense on the surface, it does not consider the full market dynamics caused by the reduced cost. It also misses the significant precedent we have of other software automation and improvements adding MORE demand for software. I made a whole video talking about the economics of this (Jevons Paradox) and why I think AI will counter-intuitively create *more* software jobs: th-cam.com/video/56H0n98V6hg/w-d-xo.html
There is no Ai that can just code on the basis of prompting, as Volo said he spent two hours to troubleshoot, and it didn't fix anything. I spent 12 hours to troubleshoot, and the code gets worse. We are far from Ai taking over any jobs during this time. It can happen maybe tomorrow, or it can take ten years or hundred we just don't know what'll happen. But big corporates like Nvidia, Microsoft, Google or Apple, will not allow using these Ai models to code everything until they are sure its fully secure.
@@Jay-pw5hy Well , the milestones are inevitable , AI is advancing too fast that every opinion may or is rather vague!
thats zhe one opinion on that. i think, demand will increase! because now more software products can be build at a lower cost ! back then when google was introduced, the demand increased also ! it is the same eith chatgpt now..its google on steroids