Although I agree with your basic premise (programming is about problem solving, first and foremost), I have close to 25 years of experience as a programmer! I have experience with around 40 different programming languages...but last year (2023) I applied for over 400 jobs and the most common issue was that I didn't have 10 years of experience in any one thing... your point might be right! But, the people hiring don't seem to understand that message yet.
Recruiters and the system of hiring has been completely warped by the amount of Indians offshore applicants. Nowadays, they don't considered "the storyline of your work experience" as of anything valuable, or something they should curiate to see your potential as a great candidate. Yes a guy like you with 25 years of experience I will pick in a heart beat. But these days, these brainlet hiring system have to go through thousands of applicants because of the offshore flood gate from India...so they simplified it down to matching keywords, and the Indians are PRO at fabricating their resume to match keywords on the job posting...literally they all do this and it flood the system with their resume instead of Americans one with great experience. So hence why more offshore are getting hired and our applications are getting dumber and more cheaply made.
@@KampouseThe correct answer. These corporations are not dumb, they create excuses for reasons to not hire "on paper" but common sense is seeing you're extremely adaptable to any language as needed and that experience comes with a big price tag. We have to stop telling these salesmen and chair holders to dream big when they themselves are doing zero of the work.
Absolutely agree. You should "specialize" to finding solution to any technical problems for businesses using your brain, computer and LLM. And that all!
Yea, I'm able to build things I only dreamed of before just because GPT is filling all the information gaps. I can feed it documentation on any framework or library and learn to use it quickly. Its helped alot with the dev ops stuff like you said. This year, I built a Sass product and a server authoritive web game which is way beyond anything I've done on my own before.
Donald Knuth was the guy whose premature optimization quote is always missing the entire context of "the crucial 3% whose big gains that can easily be obtained should always be optimized".
I disagree. The times I had to fix data pipelines created by full stack developers tell me otherwise. A full stack with chatgpt is still not a data engineer 😂😂😂😂😂
Chat GPT just needs to get a little better :) I hear you though, that's why I said in the beginning, it may not be perfect, or even good, but sometimes, working, is enough to play your hand in a market, product or feature.
Chat GPT is definitely supplemental at the least. I like its ability to extrapolate information, so that I have a high-level understanding of any particular concept. At that point, I ask it to provide details to get a more thorough understanding. It's basically a feedback loop where I test my own understanding, ask Chat GPT, and improve my understanding with its responses in order to ask it better questions to further improve my understanding.
YES, I agree as an experienced developer with the help of these A.I. tools you can implement the backend, the frontend, even devOps. BUT in contrast to this argument is if you are new to software development how do you know what you need to know in order to develop a restful api, or a database if you had never developed any of these things? Maybe I am using the wrong A.I. tool, I tried out co-pilot on a brand-new Windows machine which already had co-pilot built into the os and it was useful for building out a template for a react project which I already knew how to create. But ended up being more of a problem when trying to create the backend api endpoint. Because I never created one before. At the end I found myself solving the problem after reading the docs from the official Golang site. Just my two-sense. I use A.I. but only when I have some understanding for what I am trying to build, who knows maybe skill issues.
@@denisblack9897 Well compared to coding, learning stocks feels doable even if it is still hard. Even made a few bucks already. and the way I search for stocks had some crossover with Micheal Burry, so im kinda getting somewhere.
I know exactly what you mean, in 9 days , I built MVP sass for a company, Nuxt front end, Python Fast api backend, with some crazy ass concurrency magic. Spun up a VPS, setup everything in containers, with single click deployment. And more. It’s bananas what you can do. BTW, you sound like you may have had an extra drink or two. 😉
Should I even spend more time learning edge cases of frontend framework and be expert in it? Or I am just burning my time and should be more into getting knowledge of SaSS infra in general? I am currently working as frontend dev and I am reflecting to your comment specificaly
@@dinesee1984 After you get a general understanding of the front-end, study the back-end and become a full-stack. From there you can decide if you want to pursue DevOps, or Database engineering or any other specific field.
@@dinesee1984 Of course you need to! That's only something u can do when doing PoC/MVP stuff. When building the actual thing, this bullsh1t won't work, AT ALL
I don't totally agree with your idea of no necessity to have specialization anymore because one should understand what is provided by the bot (maybe not always, but most of the times). In your case, you already had years of relevant experience before using a bot, but this is not true for everyone. In contrary, I think those who want to take a career path in IT should aim for one specialization and that's specializing in programming. This is because it includes multiple sub-skills such as problem-solving, methodical thinking, cloud computing, Git and CI/CD, database, Linux and API. Once you are a good programmer, then you can be most productive using a bot etc. A skilled artist can produce a masterpiece with just a pencil and paper, but someone with little skill having all mediums at his disposal might produce something that someone might pay for or hang on the wall, but if you have a choice between the two work of art which one you'd likely to have?
I guess that's my point though, I've been doing this long enough to know what I need from it. There is still a lot I don't know, but this allows someone with a lot of experience in multiple areas to move very quickly. A beginner with AI can't replace me, I'll still be much more efficient due to that past knowledge. A beginner can catch up in any one area, much faster with AI.
My take on this is AI will eventually wipe out ‘Generalist’ SDEs as these problems don’t require thorough technical depth. However , there will still be a market for ‘Specialist’ SDEs to go one level deeper. E.g performance critical applications on legacy codebase + business knowledge
I kind of disagree. You need to learn the basics of OOP and perfect your programming/coding skills as much as you can, and then you will be able to use AI tools much more effectively in a wide variety of programming languages, technologies and technical problems you might face in your projects. The better you are at the topic at hand, the better you will be able to use the AI tools to improve your productivity. The most important thing is to develop a programming mindset, and you ain't gonna do that if you don't perfect your skills. If you learn one technology well, the transfer is great, and with AI, you become a lot faster and more productive. I am currently testing some web apps i created in JS and Node.js, and i started working with those technologies less than a week ago. So, i already excelled a lot, of course I'm using GPT a lot, but i also understand a lot about application development since I'm a professional Java Developer, primarily a backend, but learning frontend is not an issue if you are experienced in other areas of programming. Its actually very intuitive, and I'm learning super fast. So yes, use AI, but also don't forget to perfect yourself and specialize, that will never get old. Then your brain will rewire, and you will be able to implement programming solutions much more effectively. In other words, you will be better of coordinating whatever you have to do using your AI assistants.
the endpoint of all this 'muhhh improved efficiency' is that software (and likely all digital 'content') will trend in price towards 0 since so much of it will be easily replicable and generic and 'solved' thanks to Generative AI and so many people becoming muhhh more efficient
If allowed, AI can learn from AI, but it's important to remember that the foundation of AI's knowledge is based on human knowledge. However, the potential danger lies in the future when the majority of training sets will be produced by AI. This could impact human ability if there are any inaccuracies or errors in the AI.
It's not just IT being hammered in the job markets. Other fields are down too, the economy is not nearly as good as what is being advertised. It's not good for most people out there right now. Ultimately, AI may cause a reduction in jobs, but it should only be temporary. It's going to up the bar and rush in a whole lot of new investment and business ideas. It's just not a good time right now and that doesn't have much to do with AI... yet. I think we'll see this used as another tool to create an explosion of innovation in the computing world. Similar to the advent of the internet.
@@realchrishawkes I hope you're right but, my feeling is that, because only so much can be consumed, there is also a limit to how much can be created. As the efficiencies continue we'll be creating 2X, 3X, or 10X more by the same amount of people. Who will consume all of that product? If there aren't consumers for it, we'll enter a deflationary state for products, salaries, etc. Basically everything.
who cares if you're more 'muhhh efficient' unless you are CAPTURING that efficiency IE you deliver a SASS you OWN or you own the agency you are coding for etc.. etc...
Companies paying tens of millions every month for a sass product, may be able to do it for way less soon. Hopefully, that means more competition as well. Most likely that's what it's going to turn out to be. We'll still have a need for creative people who can get business goals accomplished quickly.
@@realchrishawkes one area I am bullish on when it comes to improved 'muhh efficiency' is Game DEV since I think tiny studios have the potential to erode AA and AAA companies out of business (and they definitely deserve to end)
One challenging thing in this department is these game engines are so complex. One minor update here or there, which they do all the time and things don't work. I'll be very happy when I can get more help in this area. I've spent countless hours dealing with some unity or unreal update, where some default setting wasn't clicked or only affected something because of x, y and z. It's going to be difficult for these bots to keep up with the engine changing every other day.
If you're more of an entrepreneur type programmer you'll actually need to deploy the things to production which you won't be able to do if you're specializing.
@@FlygOnLiTe As mentioned in the video you have LLMs for that which are pretty good at generic stuff. If you are enterprenerial you can ask what you are missing about deployment, config, domains, etc
We're headed for a world where the best programmers will be entrepreneurial. It's already been that way for a long time, but highly paid specialist were needed to reach goals for many products and businesses. They're still needed for many established products and businesses, but the playing field is evening for the entrepreneurial programmer with these AI tools.
@@realchrishawkes I believe I understand your point and I partially agree. That’s a future I’d love to be a part of but I don’t think we are headed there. It feels like LLMs can completely take your job if you can just write some code and deploy it as it’s done many many times and they can be trained on it. What a programmer can do that an LLM can’t? My only answer is get specialized on highly technical stuff that only experts can do and there’s very little training data on them.
You laugh and then you copy. I like how chris is finally admitting that chat GPT is useful and can replace specialised programmers. Last year, he was making multiple videos as to why GPT is so bad and cannot replace developers.
It's still not even close to doing what software engineers need to do day to day. People last year were saying we'd have no jobs already. I admit, this is definitely a great tool to have. It writes some horrible code as well.
@@realchrishawkes "It's still not even close to doing what software engineers need to do day to day. " You think this is still true of most juniors? I was one of the ppl who disagreed with you last year. I said then and still say now that seniors will be even more needed (for a while), paid more but u will be expected to be able to do WAY more with AI tools. But everyone below senior who isnt REALLY REALLY REALLY GOOD is fked.... I still think I disagree with your belief that AI will cause an overall explosion of tech employee demand due to a greater ability to produce, like with past tech milestones... I think AI is a bottleneck and will lead to crazy monopolization, and way less jobs overall in the coming years.
Although I agree with your basic premise (programming is about problem solving, first and foremost), I have close to 25 years of experience as a programmer! I have experience with around 40 different programming languages...but last year (2023) I applied for over 400 jobs and the most common issue was that I didn't have 10 years of experience in any one thing... your point might be right! But, the people hiring don't seem to understand that message yet.
Recruiters and the system of hiring has been completely warped by the amount of Indians offshore applicants.
Nowadays, they don't considered "the storyline of your work experience" as of anything valuable, or something they should curiate to see your potential as a great candidate.
Yes a guy like you with 25 years of experience I will pick in a heart beat. But these days, these brainlet hiring system have to go through thousands of applicants because of the offshore flood gate from India...so they simplified it down to matching keywords, and the Indians are PRO at fabricating their resume to match keywords on the job posting...literally they all do this and it flood the system with their resume instead of Americans one with great experience.
So hence why more offshore are getting hired and our applications are getting dumber and more cheaply made.
no one want to invest in hire lol
People hiring never understand anything.
@@KampouseThe correct answer. These corporations are not dumb, they create excuses for reasons to not hire "on paper" but common sense is seeing you're extremely adaptable to any language as needed and that experience comes with a big price tag. We have to stop telling these salesmen and chair holders to dream big when they themselves are doing zero of the work.
Absolutely agree. You should "specialize" to finding solution to any technical problems for businesses using your brain, computer and LLM. And that all!
Yea, I'm able to build things I only dreamed of before just because GPT is filling all the information gaps. I can feed it documentation on any framework or library and learn to use it quickly. Its helped alot with the dev ops stuff like you said. This year, I built a Sass product and a server authoritive web game which is way beyond anything I've done on my own before.
What tools/agents have you found to be most useful?
Also, anything that's useful with local llms too?
Donald Knuth was the guy whose premature optimization quote is always missing the entire context of "the crucial 3% whose big gains that can easily be obtained should always be optimized".
Good point.
I think "security research" will be a good area to specialize in with all this ai generated code.
Agreed.
I disagree. The times I had to fix data pipelines created by full stack developers tell me otherwise. A full stack with chatgpt is still not a data engineer 😂😂😂😂😂
Chat GPT just needs to get a little better :) I hear you though, that's why I said in the beginning, it may not be perfect, or even good, but sometimes, working, is enough to play your hand in a market, product or feature.
Chat GPT is definitely supplemental at the least. I like its ability to extrapolate information, so that I have a high-level understanding of any particular concept. At that point, I ask it to provide details to get a more thorough understanding.
It's basically a feedback loop where I test my own understanding, ask Chat GPT, and improve my understanding with its responses in order to ask it better questions to further improve my understanding.
Totally agree!
YES, I agree as an experienced developer with the help of these A.I. tools you can implement the backend, the frontend, even devOps. BUT in contrast to this argument is if you are new to software development how do you know what you need to know in order to develop a restful api, or a database if you had never developed any of these things? Maybe I am using the wrong A.I. tool, I tried out co-pilot on a brand-new Windows machine which already had co-pilot built into the os and it was useful for building out a template for a react project which I already knew how to create. But ended up being more of a problem when trying to create the backend api endpoint. Because I never created one before. At the end I found myself solving the problem after reading the docs from the official Golang site. Just my two-sense.
I use A.I. but only when I have some understanding for what I am trying to build, who knows maybe skill issues.
Spot on.
@@jokelot5221 thank you
I stopped trying to learn how to code and now trying to learn a bit about stocks instead. But I still love hearing about how coding is changing.
Stocks lol😂
You can’t trade against bots with best in class wires and internet connection to make orders closest possible to instantly
@@denisblack9897 Well compared to coding, learning stocks feels doable even if it is still hard. Even made a few bucks already. and the way I search for stocks had some crossover with Micheal Burry, so im kinda getting somewhere.
I know exactly what you mean, in 9 days , I built MVP sass for a company, Nuxt front end, Python Fast api backend, with some crazy ass concurrency magic. Spun up a VPS, setup everything in containers, with single click deployment. And more. It’s bananas what you can do. BTW, you sound like you may have had an extra drink or two. 😉
I think he always speak like in this one.
Should I even spend more time learning edge cases of frontend framework and be expert in it? Or I am just burning my time and should be more into getting knowledge of SaSS infra in general? I am currently working as frontend dev and I am reflecting to your comment specificaly
@@dinesee1984 After you get a general understanding of the front-end, study the back-end and become a full-stack. From there you can decide if you want to pursue DevOps, or Database engineering or any other specific field.
@@dinesee1984 Of course you need to! That's only something u can do when doing PoC/MVP stuff. When building the actual thing, this bullsh1t won't work, AT ALL
Oh common it's not that big of a productivity boost. It can help you write code but doesn't have the full context of your enterprise application.
Yeah I agree. Once AWS allows AI to change the configuration by customer command, deploying large scale architecture under 30 min might be possible
I don't totally agree with your idea of no necessity to have specialization anymore because one should understand what is provided by the bot (maybe not always, but most of the times). In your case, you already had years of relevant experience before using a bot, but this is not true for everyone. In contrary, I think those who want to take a career path in IT should aim for one specialization and that's specializing in programming. This is because it includes multiple sub-skills such as problem-solving, methodical thinking, cloud computing, Git and CI/CD, database, Linux and API. Once you are a good programmer, then you can be most productive using a bot etc.
A skilled artist can produce a masterpiece with just a pencil and paper, but someone with little skill having all mediums at his disposal might produce something that someone might pay for or hang on the wall, but if you have a choice between the two work of art which one you'd likely to have?
I guess that's my point though, I've been doing this long enough to know what I need from it. There is still a lot I don't know, but this allows someone with a lot of experience in multiple areas to move very quickly. A beginner with AI can't replace me, I'll still be much more efficient due to that past knowledge.
A beginner can catch up in any one area, much faster with AI.
Hey Chris! Nice to hear you
I'm using Tabnine. It can keep up with me, and my style of programming. Using it with VS Code and Python.
How is it ?
And Chris, create some Unreal tuts for pros :)
What if AI creates an optimized programming language for itself and then we would only need to write the prompts and check the results ?
It's probably already doing that. English may be the next big programming language.
My take on this is AI will eventually wipe out ‘Generalist’ SDEs as these problems don’t require thorough technical depth. However , there will still be a market for ‘Specialist’ SDEs to go one level deeper. E.g performance critical applications on legacy codebase + business knowledge
I kind of disagree. You need to learn the basics of OOP and perfect your programming/coding skills as much as you can, and then you will be able to use AI tools much more effectively in a wide variety of programming languages, technologies and technical problems you might face in your projects. The better you are at the topic at hand, the better you will be able to use the AI tools to improve your productivity. The most important thing is to develop a programming mindset, and you ain't gonna do that if you don't perfect your skills. If you learn one technology well, the transfer is great, and with AI, you become a lot faster and more productive. I am currently testing some web apps i created in JS and Node.js, and i started working with those technologies less than a week ago. So, i already excelled a lot, of course I'm using GPT a lot, but i also understand a lot about application development since I'm a professional Java Developer, primarily a backend, but learning frontend is not an issue if you are experienced in other areas of programming. Its actually very intuitive, and I'm learning super fast. So yes, use AI, but also don't forget to perfect yourself and specialize, that will never get old. Then your brain will rewire, and you will be able to implement programming solutions much more effectively. In other words, you will be better of coordinating whatever you have to do using your AI assistants.
Love your coding philosophy !
the endpoint of all this 'muhhh improved efficiency' is that software (and likely all digital 'content') will trend in price towards 0 since so much of it will be easily replicable and generic and 'solved' thanks to Generative AI and so many people becoming muhhh more efficient
Digital content has been trending towards zero for a long long time now.
Awesome video. THank you for sharing... I wish I could find some new work.. Let me know if you need any assistance.
If allowed, AI can learn from AI, but it's important to remember that the foundation of AI's knowledge is based on human knowledge. However, the potential danger lies in the future when the majority of training sets will be produced by AI. This could impact human ability if there are any inaccuracies or errors in the AI.
We're going to have multiple industries created just for fact checking AI output.
There are some darker coding that AI is not trained. Finding exploits for example.
If everyone can do everything and senior devs are super powered, will this lead to a reduction in jobs?
It feels like it already has.
It's not just IT being hammered in the job markets. Other fields are down too, the economy is not nearly as good as what is being advertised. It's not good for most people out there right now.
Ultimately, AI may cause a reduction in jobs, but it should only be temporary. It's going to up the bar and rush in a whole lot of new investment and business ideas. It's just not a good time right now and that doesn't have much to do with AI... yet. I think we'll see this used as another tool to create an explosion of innovation in the computing world. Similar to the advent of the internet.
@@realchrishawkes I hope you're right but, my feeling is that, because only so much can be consumed, there is also a limit to how much can be created. As the efficiencies continue we'll be creating 2X, 3X, or 10X more by the same amount of people. Who will consume all of that product? If there aren't consumers for it, we'll enter a deflationary state for products, salaries, etc. Basically everything.
15 years and not an expert in any area. That is real. But, stupid job descriptions will ask for experts while only requiring 5 years of experience.
I knew you would change your mind
Well, I still think there is a lot of hype. It's a great buddy system to code with though. Much better than Google.
Great video!
Thanks!
mic is low
My bad, I usually record at a lower level and amp it up, but forgot to.
is that a oracle logo next to unreal? D:
It’s the oculus logo
Yeah, VR game development with Unreal, testing on Oculus.
who cares if you're more 'muhhh efficient' unless you are CAPTURING that efficiency IE you deliver a SASS you OWN or you own the agency you are coding for etc.. etc...
Companies paying tens of millions every month for a sass product, may be able to do it for way less soon. Hopefully, that means more competition as well. Most likely that's what it's going to turn out to be. We'll still have a need for creative people who can get business goals accomplished quickly.
@@realchrishawkes one area I am bullish on when it comes to improved 'muhh efficiency' is Game DEV since I think tiny studios have the potential to erode AA and AAA companies out of business (and they definitely deserve to end)
One challenging thing in this department is these game engines are so complex. One minor update here or there, which they do all the time and things don't work. I'll be very happy when I can get more help in this area. I've spent countless hours dealing with some unity or unreal update, where some default setting wasn't clicked or only affected something because of x, y and z. It's going to be difficult for these bots to keep up with the engine changing every other day.
Not the best take, now is the time to specialize. It’s what an LLM can’t do, at least under the current paradigm.
If you're more of an entrepreneur type programmer you'll actually need to deploy the things to production which you won't be able to do if you're specializing.
@@FlygOnLiTe As mentioned in the video you have LLMs for that which are pretty good at generic stuff. If you are enterprenerial you can ask what you are missing about deployment, config, domains, etc
We're headed for a world where the best programmers will be entrepreneurial. It's already been that way for a long time, but highly paid specialist were needed to reach goals for many products and businesses. They're still needed for many established products and businesses, but the playing field is evening for the entrepreneurial programmer with these AI tools.
@@realchrishawkes I believe I understand your point and I partially agree. That’s a future I’d love to be a part of but I don’t think we are headed there. It feels like LLMs can completely take your job if you can just write some code and deploy it as it’s done many many times and they can be trained on it. What a programmer can do that an LLM can’t? My only answer is get specialized on highly technical stuff that only experts can do and there’s very little training data on them.
you never did
I just want to really good at c language all other language will be taken care by chat gpt for me
You’re in for a surprise lol all those languages will be considered legacy in like 2-3 years trust
Oh point is to embrace ai tools 😳
You laugh and then you copy. I like how chris is finally admitting that chat GPT is useful and can replace specialised programmers. Last year, he was making multiple videos as to why GPT is so bad and cannot replace developers.
It's still not even close to doing what software engineers need to do day to day. People last year were saying we'd have no jobs already. I admit, this is definitely a great tool to have. It writes some horrible code as well.
@@realchrishawkes "It's still not even close to doing what software engineers need to do day to day. "
You think this is still true of most juniors? I was one of the ppl who disagreed with you last year. I said then and still say now that seniors will be even more needed (for a while), paid more but u will be expected to be able to do WAY more with AI tools. But everyone below senior who isnt REALLY REALLY REALLY GOOD is fked.... I still think I disagree with your belief that AI will cause an overall explosion of tech employee demand due to a greater ability to produce, like with past tech milestones... I think AI is a bottleneck and will lead to crazy monopolization, and way less jobs overall in the coming years.
wow so early
Hella