I been in the field for almost 2 years as a software engineer, and everything you said was spot on. I love using chatGPT. I've used to get started and when I have ideas for complex ideas, I feed the ideas to ChatGPT and get instance feedback if the solution I think is possible. This helps me to be more productive. Software engineers need to embrace change(actually everyone regardless if its your job or real life).
If software developers become obsolete then any other job will have the same faith. Simples as that. I don’t know why the fear only happens about coding… what about math and physics automation?
@@davedsilva look like you dont even have a job or knowledge about coding. i can use mine hacking website and selling those data on black market. i can even make spyware and sell it to someone
Once all of the wealth is concentrated in the hands of a few thousand tech bros, they will of course need services. Being so filthy rich, means that you basically don't have to do any task that you find to be boring. New positions will spring up that are unheard of today. For example, "Butt wiper". I know if I was worth a trillion dollars, I would want to pay someone to wipe my butt instead of doing it myself.
We coders aren’t obsolete, but our field has become incredibly saturated. Now, anyone with even a basic understanding can create sophisticated applications without any coding experience. The game has changed; creativity, not just technical skill, is now the key. With powerful tools available, anyone with a spark of imagination can bring their ideas to life without spending a fortune on developers. I started coding back in the mid-'80s, but it’s clear this territory is no longer exclusive to us
Yep, execution is becoming a lot easier, so creativity is going to become a lot more valuable. Using code to make tools that are actually useful. But it will be easy for anyone to copy… meaning lots of competition. The cost of software is going to drop severely.
Considering the speed of the development, software engineering as in programming will change a lot. Just imagine how chatGPT will look like in 5-10 years. Or even 20 years, 50 years. Many developers will loose their jobs, because ChatGPT can generate code way faster. What you do in one hour, chatGPT does in 10 seconds.
Anything that is well documented out and is a straight-forward question ChatGPT and O1 can perform exceptionally. Anything which requires nuance and wholistic understanding both conceptually and factually, like communication of business problems and solving them, they perform really poorly on
you can't extrapolate. What if this is 90% of as good as it will ever get? We don't know how good it is capable of getting. We don't know that it will be economical compared to humans. What we do know now is that they can NOT write code that solves complex problems faster than a software developer can. Suppose this algorithm is perfect and correct. What about the training data? It is trained off of open source code online, most of which is bad code. How can it correct itself if it doesn't know what is correct? The more code that AI writes, the more AIs that will be trained off of that code. And all of the errors will compound and explode. There might not even be enough data to train these things properly.
@@awesomedavid2012You're correct, Sir. I haven't heard anyone rave about how great and novel the software is that is being generated. It can do pretty good at regurgitating what it was trained on but not at prividing new innovation.
People don't understand just how complex an app ecosystem are right now. There are so many things you have to update/maintain/compliance be it an api or libary or even just Google or apple terms and conditions or security updates etc.. AI is not going to do that for you.
It already is. Developers are paid primary to code. And if AI can do that for you. What stops are person not well versed in tech doing your job for you. They can take a crash course on the diffrent tech stacks and chose from there. Programming has abstraction layers and with more them more people can code. This means supply goes and demand for experts goes down. meaning we get paid less. you dont want you job to on the same level as a fast food worker. Which it is going to be.
@@bravo90_ I've been programming in my niche and I just cannot see how your statement jives with my reality. There is simply too much complexity and unique problems that generative AI cannot solve, let alone even attempt to solve. Experience is the best teacher and computer system are nowhere near being able to simulate a brain
But the problem is not for senior programmers. Here juniors are the one who suffer. No compony will give you such big responsibility and technical jobs right after university. It used to be, you get hired to code under supervision , and learn from your seniors, in 2 or 3 years you become better programmer. But now if they start automating junior jobs, sw engeenering become a mid senior job. It become like cybersecurity.
I have a friend who works at aconstruction buildings. He is a manual worker. He is probably the most happy and mentally stabile people I know. In his job he does details, not lifting heavy stones or something. He works with cables and paints walls, etc. He doesn't have to worry about all that. Programming is more respected and people consider it much better than what my friend does. But still I think soon I will give up on that IT industry. I spend 3 years in IT and it is too much for me. Always uncertainty, infinite learning and fight. I just wanna be happy honestly, even if I am not the richest person in the room.
"and i will see you in the next onnnnneeee" ah Tim, you truly have no idea how many times I hear you say that on a weekly and even daily basis haha and it always puts a smile on my face
Programmers are going the same way as human cashiers at the checkout in the supermarket. The Lidl in my neighbourhood used to have 5 human checkout points. Now they have 8 self scan checkout points and 1 human cashier and guess what? There is a waiting line for the selfscan points and even when there is nobody at the human cashier point, customers will still prefer the selfscan machines.
Well, here in the UK the opposite is now happening. People don't really like those machines for large numbers of items. As such, several large supermarket chains are now pulling the machines out and employing more people for checkout roles. Indeed, Booths in the north of England has got rid of all their machines. Asda, as I recall, intend to halve the number of machines. So your argument doesn't hold water here in the UK, at least.
@@paraglide01 surveying their customers, listening to complaints and observing the huge lines of people waiting at the manned checkouts while the self service ones are pretty free. For a large shop with a cartfull of items, people get to chat to the checkout person and they only have to wait for the goods to be rolled to them for easy packing, not have to fumble around scanning and doing the packing which takes a lot longer. Lastly, people just feel they are unpaid employees when they have to do all their own scanning and packing.
Whether you agree or not, AGI (Artificial General Intelligence) is the beginning of the end for complex tasks being solely human. The next model, or perhaps the one after that, will conquer these challenges. But make no mistake-it's only a matter of time before it arrives.
Not sure. I've read some academics believe we may need a new paradigm beyond LLM to go significantly further with reasoning and what we call "intelligence". They argue that the rate of improvement of AIs is slowing down, and that you cannot get incrementally to AGI on the path we are currently using. That we cannot scale indefinitely using the current foundations. Who knows?
AGI is not going to happen for many years. LLMs are not a viable route. o1-preview, as I've shown on my channel, cannot do basic maths. Without that you just have a fancy writing machine, not an intelligence that will perform any function.
It is going to end software developers once they add agent-based architecture into the mix. Frog in boiling water type thing. They already have the tech they're just rolling it out slowly to avoid regulations.
yeah like they barely put any effort into marketing / hyping up o1' s release unlike 4o's. They just randomly decided to make it accessible and let us knew via twitter.
@@hendrx On this one in 5 to 10 years. I just retired as a dev and I can see what’s coming from testing these models. Right now they are great tools to accelerate your work, but realistically do you actually think you will be writing software in 5 to 10 years at the rate this technology is improving? You will just tell them what you want written and it will generate the code. This new model was able to actually do graduate level physics problems out of Jackson. This is the standard Electromagnetic’s book used in Masters and PhD physics courses. I used the book for my masters in Physics and I was blown away seeing this thing to some of those problems which require partial differential equations to solve.
@@jimpresser3438 Your argument was used 5-10 years ago. AI can still only solve rather simple problems, or answers with straight answers, such as those from a university level text book. It cannot create ideas or solutions, it can only apply the ideas from it'a dataset. If you just reitred, you'd definitely know that in a real world software problem, you cannot plug and place something from stackoverflow. You can only do that if the problem is simple. I use AI everyday at work, yet the reality I encounter is not the one you are describing. It can only speed up the developing process of the simple parts of a system, and it is still performing rather poorly. At this point AI is in the same boat as fusion energy it's "only 10 years away!" I have no need to worry seeing how much struggles in real life tech stacks firat hand.
I’m still in high school, but I think the o1 model will help programmers by saving time instead of taking their jobs. Plus, AI will open up new jobs for people like designers and engineers.
I retired from my previous industry (public education: English teacher, school counselor, and admin) about two years ago. And that's when this wave of LLMs started taking off (not exponentially, but definitely increasing in pace). I've had several friends and recruiters tell me that my main advantage for job seeking will be my broad soft skills gained from my previous industry: dependability and reliability, communication skills (oral and written), creativity and imagination, teamwork (both vertical and horizontal), critical thinking across multiple spectra, adaptability, empathy, and eagerness to learn (and engage others in my learning).
Tim, that advice is outdated, especially for juniors. The people getting hired today are already in the industry. If you're just starting out, it's really tough. Employers won't hire anyone without experience, but you can't get experience without a job. They're outsourcing to cheaper countries, and a lot of entry-level tasks are being automated. The only places still hiring juniors are government programs or large corporations, usually through CS graduate programs. To be honest, you're more likely to get hired as a plumber, electrician, or handyman than as a developer right now. The tech industry is still great in many ways, but the job market for developers is brutal unless you have a master’s degree or strong connections. If those opportunities ever existed, I would’ve found something by mid-last year, not spent 15 months job-hunting. If anyone disagrees, that's fine, but it's not just my opinion-this is the reality of the current job market.
It seems to me you are ascribing the result of various trends to AI, and it s wrong from my pov for various reasons: 1 AI lets "bad developers" access the brain output of "good developers" That means that the California tech bros are losing competitivity very, very rapidly, vs the rest of the world. If you go to East Asia, you can get coders for a tenth of the US price, and thanks to LLMs the difference in output has vastly been reduced. 2 People who went to prestigious unis and were able to secure employment through the name, resting on their laurels because they had exclusive access to knowledge, are now facing competition from self taught devs from remote places. 3 Because LLMs make the knowledge available in various languages, non eng speakers can now access and learn more easily. Local mgmt is thus more inclined to hire them since they are able to communicate properly. Getting barely eng speaking Kor mgmt to instruct a CA based eng only dev teams was a nightmare. Now they have the same capabilities in Korea, they will hire Koreans, as they are cheaper as well. 4 Increasingly "software devs" are seen as just IT, and companies require less dev skill, and more Network / Systems / Infra skills, since the automation trend is picking up, and that needs a broad understanding of things. All in all, don t be an "average to good" dev in the US, you are too expensive and if you can t work for the giants the other companies will outsource the job. And do not be a "I can only code" IT person, you can have less advanced knowledge in coding and develop an understanding of the other fields, that will help with getting a job tremendously since you are then versatile in entry positions.
@@cayjutler1262 He's already wasted 15 months - how long do you suggest he should slog away on a project? Will his project make him stand out from the tens of thousands doing the same thing? Will his project counter the shortcomings his CV obviously has to employers? It might be better to spend the time retraining in a profession that's actually hiring juniors or developing a side hustle like affiliate marketing. Developers are 10 a penny and we're never going back to the heyday of easy tech jobs for everyone.
@howtopassthat This is just partly true, but not 100% true. The market is competitive? yes this is correct. But it's not like fresher or junior got 0% chance like the way you are implying. They still have a chance, even the emerging of AI, AI now is just an assistant, nothing more. For 10 years in the future, I cannot say, but right now, no sight of replacement happening because tool is still a tool, and this tool still need human intervention in order to function properly. So instead of looking at negative sides, we should moving forward and keep updating ourself to not become obsoleted, then the chance still many. But if you are lazy, "you already dead".
If you truly love software engineering (notice I didn’t say just coding), and never quit, and keep learning, and keep growing, you WILL stand out and you will succeed. Also notice I didn’t say it won’t be hard. In life things worth keeping and cherishing are usually hard anyway. For context, it took me 7 years to break into the field. I “quit” and got back at it many times during that process. Can’t tell how many times I felt dumb as a rock. Now I’m coming up on becoming a senior full stack engineer soon. I hope this is encouraging!
Thanks Tim, this is refreshing to know. The points you highlighted are very good. I Have always seen AI in programming as and advance code completion, nothing more.
I'm still a 2nd yr CS student and 2b honest this whole AI advancements used to bother me. This was until I started using chatgpt. Instead of just asking it to generate code for a task so that I can quickly finish and submit my weekly practicals, I started having "conversations" about whatever task I want to do and this helped me a lot in spotting its mistakes in the code it generates before I'd even intergrate that code in my project. In a way it gave me time to focus more on what could go wrong on the project/practical I'd be writing. After I finish writing my code, I'd normally paste some of my files on it so that it can check if I didn't miss anything that can maybe mess up my code, especially with handling exceptions coz I'm not very good at that. Great video, noted👌
When i was a young engineer we were told to try and "Engineer out the technicians". AI "WILL" be able to Engineer out the Engineers! It will not stop with engineers, AI will be everywhere and SOON! This is not a bad thing in the long run. Now this does not engineer out innovation, in fact I think it will accelerate it! A few months ago there were blogs saying AI has peaked-lol. I remember hearing all these people saying look we are flattening out, but now here we are with another WOW with AI. We have not even scratched the surface of what the current customer level AI can do! I can not even imagine on what is going on in the heart of OPEN AI right now with GPT 5!!! We are in the dawn of a new technology revolution and AI will lead the way! AI will lead in things like cancer research, anything medical, energy maybe it will solve fusion, technology, longevity, space exploration, food and farming, and so many more i can not even think of right now!!!
AI code is going to give every person a developer side kick, that also knows how to implement numerical methods to solve problems. Engineers will have AI write tools for them to solve engineering problems, but won’t replace engineers. Scientists can have it help model their data, visualize, etc. it’s going to save so much time for STEM workers. There will still be software engineers and developers to direct the AI, but far fewer humans. One developer can now do the work today that it took 2 or 3 people before AI. Obviously that means the job market is now saturated.
Perfect analysis. This is an evolution and a tool, and this changes our software development methodologies and know-how, but this is just the beginning. When AI will be fully automated with more sophisticated app templates, coder work will be limited to know how to ask a customised app responding to customer's enquiries. And after, there will be another evolution. Just think that the biggest human evolution is the communication between people. LLM/AI is the communication between humans and machines. And at last, there will be the communication between machines. That's the point where humans will be over.
Developers are safe now, until next year when autonomous AI agent will be coming. I dont think it is a 'natural evolution' of software development, it is more a revolution and a lot of programmers risk to loose their job in the next two years.
@@tradingviewindicatorguides895 I say that a lot of developers risk to loose their jobs, not every developers. Maybe not the techleads. And I didn't say that is was optimal at the moment, but in a near future, maybe with AI agents, it will certainly replace many developers.
I absolutely agree with Tim, AI is like our Friend, and it is not a replacing the jobs, in fact it's creating more jobs for software developer and prompt engineer
Hi Tim, great great video. Thanks for restoring my confidence. I am learning java mainly using Ai actually. And it is true, things are changing in many other fields too, for ex I used to be a manual turner metal engineer in the past and not it is all computerised in general, incredible what this machines can do now. Not that I fully agree but you kinda need to keep up with the times somehow if you want to work. Thanks again
I fully agree with you. Coding is/ can be very tiring. So if you can save brainpower by using this tools for the low-mid level stuff, you can focus on an the more difficult things with your limited brainpower.
Full disclosure I am a Senior Software developer. Just watch Star Trek the Next Generation. They tell the computer what they need and it just does it. That's where we are headed. These LLM models are at best 1st Gen releases and they are already this good. In 2 or 3 more generations I expect they will be so good that software development will really be a thing of the past. They are already creating models that can generate rudimentary 3d games. Ten years from now at this rate I am pretty sure we will have perfect responses to our requests no matter what they are.
@@TheBrothersTalk On a larger timescale though, robotics isn't where it needs to be now to replace fine motor tasks even if the "brain" (the LLM models) are smart enough to know what to do, the "body" (the robots) aren't precise enough to execute the actions yet.
Wow i think you are absolutely right coz it means even all other jobs shall be replaced too if the software engineers are replaced there is no way other jobs can survive
It has majority for communication happens on the web. Post is to slow in this world. If you could get physical products online then. post would be dead
Like this guy is going to give any sort of honest advice, or feedback. I’ve learned most of these software RUclipsrs are just like like any content creator… say anything as long as it aligns with their bottom line(money/views).
Yep. Tetris, Snake, and I'm sure they'll crack PacMan at some point. I can tell you from experience last year, getting GPT to write a more complex game from scratch ended up with several days of herding it to do what I could have done in half a day, complex code that was not understandable, and repetitions of code because it kept forgetting it had already written utility functions. It worked - but it was enormously inefficient and the code base was really unmanageable.
@@CricketFantasies5 would it not have been more fun and more insightful to have understood how to code it yourself, even if you used the AI to help you? That seems to me to be more advantageous than getting it to produce a black box which you don't understand and cannot debug.
@@cbnewham5633 well, I am in content creation field. So, I don't want to spend my time learning to code, rather I want to use it in whatever meaningful way I can. At normal circumstances, I could have bought a subscription of an already present software built by a developer, but AI helped me develop my own software (only back-end) saving my money and time (to learn coding and the software works perfectly fine. That's why, I say that AI will give the software developers a tough challenge as more and more people will do it themselves or the developers will face severe competition among themselves leading to a portion of them becoming jobless.
Ai is going to take our jobs. You asked what I'm going to do instead of run full speed at a brick wall? I started my own business using ai. I'm looking to expand my empire now. The problem with dev is the amount of work that goes into it just to be made obsolete in only a matter of time.
@@JVWorksIt sure did not happen but the models become much better. There will be software engineers, but fewer. This is the same moment as CAD introduction to mechanical engineers, now one ME can do the work of five.
True, true... I work in automotive industry as a industrial engineer but my background is automation programmer. All modifications on machines have to be outsourced because we do not have team for that. As I understand the topic very well, I'm able to implement those modifications nearly for free with help of AI as my "junior programmers". So there is power in using those tools, I save a lot of money because usually even small software additions costs thousands of euros plus external companies can't to it ASAP because they have their own projects. We can apply something like 80/20 here. 20% of a problem is actual coding, but I will have to still do the 80% of that job. On the other hand, those 20% of coding usually takes 80-90% of that time. With AI, this time can be decreased by 50-80%, what is great. at the end, everyone is happy, corp. that I work for valuates me even more and I'm able to do what I love at position where I shouldn't code at all.
"The data flywheel refers to the self-reinforcing cycle where the more data a system collects, the more value it can provide, leading to a compounding effect on insights and innovation." Unfortunately for people just getting into this field, the flywheel has just started. There's no going back. Enjoy the ride as AI takes over
The way I think this will play out: a lot more people will start producing code with AI. For example, an analyst will use it to write some python to write a colab script to connect data. Millions will start writing code like this. This code will break and they won't be able to fix it. That's where you come in. You will be so busy fixing code, so learn how to specialise in AI code repair and you will always have a job.
Hello bro can you make a roadmap video from beginner to advanced (absolute essential topics and projects) for your videos to learn python . There are so many good videos, so I couldn't understand where to start. Thank you for your time for this comment and content.
Interestingly, the time spent writing code jas shifted to making sure that AI output actually matches my needs properly. Following it blindly is a sure way to break things
Its just a matter of time when we have new softwares that are built on AI. Then its a new paradigm, but sure people will be in the loop for sure but we dont know what will be entirely automated that used to be code by hand. Time will tell🙌
Been a developer for over 25 years, have been trying to get these new AI tools to accomplish tasks so I can get more done. It's always close but has yet to be 100% yet. The issue I keep seeing is that these companies keep trying to sensor things by adding invisible prompts and its screwing up the code return. The best i've got was running models locally but i've only been able to run smaller models due to the cost of hardware.
It´s great for new features, I imagine that those models are great for startups, however, most software development is 95% searching the right place and making sure that what you do doesn´t break things and 5% writing code.
100% mirrors my viewpoint as a fellow millennial. The yak-yak software engineers will become obsolete, right !! The future of application/system development will be impossible for the layman, only an engineer can think these things through. Lay people do not know what they want/need unless its given/offered to them.
I gotta be honest, I don't think AI will ever fully take over software development. However, I could see it doing to the developer field what stable robust servers did to network engineers. Like when I first started working as a network engineer, the formula was simple. One administrator per server or 100 users. However, by the time I left you might have a team of 3-4 managing 100 servers and thousands of employees. I can't even imagine what it's like today since it's been about 17 years. So, I'd guess you'll still have developers, but you'll need much fewer per project and AI will do most of the simple lifting.
Or, like with pretty much every other field where automation has happened, the amount or complexity of things being produced will increase, which means the same amount of workers will be required. What I'm more concerned about is that companies will find a way to use these changes as an excuse to exploit workers even more.
its not the end for software developers, although i expect the number of jobs in the industry will be reduced and the workflow will be a bit different. and yes, like u are saying, coding will gradually become higher level and higher level although there will still be demand for those that actually understand the low level stuff.
I like to compare AI with everyday life tools. In the past we could clean with brooms. They still exist but now we also have vacuum cleaners, leaf blowers or outdoors vehicles that make the task easier or even vacuum robots that will do it themselves. Same for cutting grass, from scissors to manual mowers to all kinds of machines. Also remember when frameworks appeared? I think it's all part of progress and it's more tools that get added. We still need to be part of the process to make them work as we want.
Supply and demand. Productivity will go up radically, cost will go down and hence more demand for software. This may translate into more or fewer programmers (probably fewer) but "everyone" will not become a programmer. That said, I think Tim is correct with what he is saying, but I think at some point most programmers who work on repetitive type apps will have a very difficult time of it.
The thing is biologics is not keeping up with AI and machine. One should use ai to evolve biologics. I do not mean zombie, but robust super biologic to keep up.
Everything is a panic. You’ll find articles panicking about not enough people having babies, and within the same article, warning that the cities are over populated and we’re using up all the natural resources. People need to choose, pick one or the other. And realistically, how much code does the world really need? At what point should we start focusing on things like clean water, mental health, not working 80% of our waking hours, etc?
Change the way you think, it does not effect developers in negative way, but a oppotunity write more code and effective. So as a developer my self it make life easier not misserable. replacing a human with AI , we are not their yet.
I went for this field cause I love coding and it seems like that is going away in the next years. I don’t know how I could possibly feel encouraged right now..
Welp, I just started learning Python about a week ago and now I see this video?! Realistically what is the point of continuing with my journey when I currently know incredibly little and AI seems like it will be a step ahead whilst I learn. Back to the trades I guess....
Why are people still learning programming? I’m confident there’s little room left for freshers in this field. At this rate, within 1 to 2 years, only a handful of freshers will be needed to support senior developers. Experienced programmers will remain in demand as some human oversight will still be required. Yet, thousands are competing for the same roles, and senior developers will get them in the near future. I don’t understand why people are still rushing into this field. Freshers should look into new opportunities and have the courage to pursue something different. After all, humanoid bots are likely to take over most jobs soon anyway! I'm keen to understand how my view might be wrong
The problem is that there is no safe job. You want to be medic or psychologist? If AI can replace developers, it will be also able to give medical examination and do mental support to patients. Lawyers? The same. Managers? It will be able to very boost productivity of managers, less managers needed. Manual workers? Sure, but there will be much more of them in the market (former white collars) and humanoids robots will slowly start to take over.
I need it to immigrate and I have already spent so much learning it. Giving it up would mean I've spent a soul crushing amount of energy on it that won't be replenished in a very long time if ever. I also know testing, so I keep it in might as a fall-back position. Why I ever started coding? This is the only thing I find interesting in IT.
It would be interesting to see let’s say a 1 year difference between someone who just started to learn coding through AI vs. someone who has been a developer for 5-10 years but just started to use AI. Choose a starting date and compare your project/startup after 1 year. How far did they get? Was it actually useful? Interested to see where this all goes!
I can't help with a direct comparison, but I've been doing coding for 35 years. AI is a useful tool, but for anything that's complex you really have to know what you are doing - AI hallucinates and the code doesn't always address edge cases. Great for boiler plate stuff or getting started with APIs you haven't used before, but I wouldn't just ship code straight from an AI without proper reviews.
This is just a cope. Developers are definitely going the way of the dodo. We're at the early stages of this AI thing and it's already this good. In a couple of year a handful of non developers will be able to huddle around a table or over zoom discuss a problem and design a structured solution. The AI will be able to build that solution and make it available in one afternoon. Developers will be replaced by structured planners
Dev channels like this can't scare away their audience or else no one will watch their videos anymore. They have to keep giving junior devs hope that they're education and channel views still have meaning.
It’s far more likely that software engineers would rise to the role of design and AI prompting and tweaking, which would obsolete the non technical worker roles. There’s no software engineer I know that can’t do the job of their non technical colleagues, but the inverse is certainly not true. “Devs” will not be called “devs” in this hypothetical future. They will be design and solutions specialists and their background and ability to understand, tweak, design, and prompt these AI to arrive at solutions will prove incredibly important. Its funny how people arrive at the conclusion that now suddenly technical expertise is useless and all the clueless MBAs with their “ideas” will be spraying prompts at the AI and somehow creating flawless scalable applications without any intervention at all. As a PM I’m desperately trying to acquire technical skill because I can see the writing on the wall in that not having it will probably render me near useless in this new future.
@@thechannelihadtomakeforthi9495 you're just confirming my point. The role of software engineer will be gone. A few software engineers will rise into the software architecture and designer roles for big projects and big clients. The rest will become part of a large hoard of low value AI prompters.
@@stinger4712 I don’t think it automatically means low value. I think it’s more likely we will enter an age where customers are asking for way more value for similar budgets. If AI greatly accelerates the speed at which products can be developed in the way you’re implying, then customers are not going to settle for similar value at similar cost. They are going to flock to companies offering them equal value for less money, and subsequently they are going to have more money to spend on more value. The only way I don’t see this being the case, is if customers are happy with the software they have and don’t want any improvements or new products/features, which I don’t see as being the case especially with the massive backlogs I’ve seen throughout my career (and unhappy customers always asking for outrageous things we just can’t do within budget). In this scenario, I think there’s increased demand for talented engineers who can do a little bit of everything and leverage AI to vastly speed up their deliveries. We might now expect a whole feature done in a few days rather than a few weeks, but that might also mean we do 5 features in a few weeks rather than 1. Of course, I could be wrong, and we could enter an age of tech recession where people move away from technology and software to focus on other pursuits, but I think we’re a long way off from this.
@@thechannelihadtomakeforthi9495it's not just about AI accelerating your work. It will take the below average to expert level. I'm trained as an electrical engineer and I'm good at it but I considered myself mediocre when it came to writing software for electrical systems. Then came chatGPT and overnight I began delivering software for the most complex systems. All I needed was the systems knowledge to quickly displace the high cost developers. That's what's going to happen across domains.
This will be like Office for Windows where Office originally came free with the Windows program. Now Office comes with in with a subscription charge. Simple AI’s free like Chat GPT but anything on a developer’s desktop will eventually become a subscription and the charges climb every few years.
I am a software developer and when using ai i can vouch that it makes you 2-3 times more productive. This equates to less programers needed. I have already noticed that now every job has over 200 applicants... Way over what it used to be like
I’m not worried at AI taking my job I’ll just learn how to program the AI. Besides currently I do customer service and I’ve been told for years AI is coming for my job, but AI may be able to guide the customer through the return process the AI dosn’t understand how people work, and that if I don’t highlight the collection link there is a 90% chance the next email from the customer is how do I get it collected. AI doesn’t understand that what people ask isn’t necessarily what they want to know the answer to.
I don't agree with this. I recently had Claude 3.5 make this very complicated piece of software for me that manages leads and bookings for my business, it is integrated with Google calendar and Google drive. All I was doing was just describing the functionality I wanted and giving the model any errors from the console. It also did extensive testing of the application. Soon software developers will not be needed or at least much much less.
From your brief description, that software has probably been written thousands of times already. Software for booking systems is so common and it's a very easy task for something that's been trained on all the code on the web to complete. Even just integrating a new feature into complex software, like RUclips for example, would need a lot more intervention from an actual developer.
@@technophobian2962 Have you seen that company that is generating open world games from text prompts? Did you see Google had a neural network recreate the Doom engine? I actually had Claude 3.5 code something like the Doom engine. I find it most impressive that it can tailor make specifically what you ask for.
I’ve been programming since my childhood 1990 and as of today I’m still employed as a programmer and still don’t see LLM replacing humans yet. Useful tool, and someday it may replace humans, but we are definitely not there yet.
In simple terms you must be knowing what you are doing. You just can't become a software engineer overnight because you have ai. Neither is software engineering being replaced. I have also been using ai and it's very useful
I appreciate your channel,I learn a lot from you.However ,plz don't do the click bait bs tile ,I don't think it will be good for both parties in the long term,cheers
Human energy is much cheaper than energy required to run such AI systems indefinitely for complete automation. So most jobs will evolve instead of completely getting replaced. Mother Earth has really ensured only natural things survive on it.
The fact that this CLICK BAIT title is a real concern - logically, development of AI is heading towards phasing out developers. The indicator being Click Bait. Disagreeing with this is like saying guys who click on videos with beautiful big breasted women do not like boobs. Which is false. Even if not a general purpose if replacing software development - monetizing in such a stigma Will align AI with out-dating software engineers. Like even if unnatural but because people. Technically if EVERYONE lost their jobs at the same time - would have zero impact on the economy. Food for thought. Like when demand for high jobs is high the dollar is strong, but if demand dropped to near zero overnight - how supply demands works would just affect the underlining value of the dollar and If, for example - If I develop AI in hope of making other people lose their job since I am unemployed would make me wealthier since there would be less wealthier people in contrast - So if everyone did this at same time Should, in theory have no influence on individuals personal worth.
I think anyone who really wants to learn these stuff can be a developer. Right now, even CS majors are having a hard time getting an entry level job or an internship. Why? School does not really teach them the practical stuff.
Sorry the future of AI is creating blueprints. You'll have blueprints for all these different types of software that the AI will follow and create variations on these blueprints to create new software.
Spot on. My CS degree is from 1991....the CRITICAL THINKING hasn't changed and being able to design is a skill. The syntax changes and has changed, but having a logical approach, structuring logic, and knowing when something is bad, is a long term marketable skill.
When AI can take one code base and move it to another without error. Then AI is interesting. For example take a react web app to vue with one requested prompt. Anything less isn’t anything to talk about.
15 years ago I was pretty much using the same tools and languages as I am using now :) that is c/c++ (yes, stl existed too back then 😁) and assembly too. And even visual studio 2005 was OK to use. And faster unlike those Javascript based IDEs we are using in 2024. But I am not saying that generative AI is not useful - it is but it is not going to replace a decent coder any time soon. BTW I code for living since 2002 and I wrote my first signal processing code in assembly in 1993.
If we see from starting and take maths as example then, in start math was greatly used and there was no machines to calculate simpler problems only mathematician can do those problem or someone who knows about math and after we upgraded and created our first calculator then we thought the job of the 'mathematician or the people who worked on shops and in business to calculate money expenditure' will be vanished because of calculator but as we know it doesn't happened because after calculator we were able to solve problems of that time when it was difficult but it is not now, So we now started focusing on more difficult problems and on more no of simpler problems to be solved at the same time with the 'help' of calculator which makes it easy to solve. And it is same in programming we are now going to develop after a decade we will be focusing on a very high tech program or software where openai will be a help for us as a calculator to a student to solve addition and subtraction
When ai came everyone thought humans will be not that much useful for work or job but it doesn't happened because after ai we got developed and our efficiency of job also increased so we then needed ai for help to do our job together with more efficiency. And humans always are evolving
Now even some human can calculate faster than and with efficiency and accuracy than the calculator as we see in shows like talent search show or anything.(America got talent). Maybe I don't know exactly name of those shows but I saw some clips and short videos of it
At the end of the day, you still need to know software engineering to debug and iterate on code written by AI. What AI will do, however, is turn some solo devs into sole proprietor AAA studios.
What do you think about the future of software engineers?
Same as network engineers, not dead at all as a career but also not what it was 15, 10 or even 5 years ago.
Over
As I think it will be more amazing and productive. It should be more
First of all o1 is not a model, it is mostly a wrapper around a model that helps to squeeze more from it.
I been in the field for almost 2 years as a software engineer, and everything you said was spot on. I love using chatGPT. I've used to get started and when I have ideas for complex ideas, I feed the ideas to ChatGPT and get instance feedback if the solution I think is possible. This helps me to be more productive.
Software engineers need to embrace change(actually everyone regardless if its your job or real life).
If software developers become obsolete then any other job will have the same faith. Simples as that. I don’t know why the fear only happens about coding… what about math and physics automation?
Exactly
Maths too...
Maths and physics have been oversaturated for years...
Except AI is exceptionally better at coding than humans@@TechWithTim
@@GameinTheSkin no it's not, you're clearly not a developer.
Sensible video.
Don't fear guys. Even if you're out of job someday, something will turn up to you because of your knowledge of technology.
Street beggars meet your description
@@davedsilva look like you dont even have a job or knowledge about coding. i can use mine hacking website and selling those data on black market. i can even make spyware and sell it to someone
@@davedsilva Why do you care brother?
Even if tech guys like us become beggars, you should strive to live a good life.
Why so much hate?
Once all of the wealth is concentrated in the hands of a few thousand tech bros, they will of course need services. Being so filthy rich, means that you basically don't have to do any task that you find to be boring. New positions will spring up that are unheard of today. For example, "Butt wiper". I know if I was worth a trillion dollars, I would want to pay someone to wipe my butt instead of doing it myself.
We coders aren’t obsolete, but our field has become incredibly saturated. Now, anyone with even a basic understanding can create sophisticated applications without any coding experience. The game has changed; creativity, not just technical skill, is now the key. With powerful tools available, anyone with a spark of imagination can bring their ideas to life without spending a fortune on developers. I started coding back in the mid-'80s, but it’s clear this territory is no longer exclusive to us
Yep, execution is becoming a lot easier, so creativity is going to become a lot more valuable. Using code to make tools that are actually useful. But it will be easy for anyone to copy… meaning lots of competition.
The cost of software is going to drop severely.
Yes, super saturated.
Fewer dev roles and a smaller percentage of those actually paying well.
But its good only for small softwares, try to work on larger software and it is barely useful.
yes. Then if you don't know what you are doing the software will implode and your only option will be write it from 0.
The people who don’t know how to code and only use chat gpt are not competition in the job market.
Yes Tim, you are right. Software Engineering is more than just coding.
enterprise-ai AI fixes this (Code complete projects in PHP or Python). Is it over for developers?
Considering the speed of the development, software engineering as in programming will change a lot. Just imagine how chatGPT will look like in 5-10 years. Or even 20 years, 50 years. Many developers will loose their jobs, because ChatGPT can generate code way faster. What you do in one hour, chatGPT does in 10 seconds.
Anything that is well documented out and is a straight-forward question ChatGPT and O1 can perform exceptionally. Anything which requires nuance and wholistic understanding both conceptually and factually, like communication of business problems and solving them, they perform really poorly on
you can't extrapolate. What if this is 90% of as good as it will ever get? We don't know how good it is capable of getting. We don't know that it will be economical compared to humans.
What we do know now is that they can NOT write code that solves complex problems faster than a software developer can. Suppose this algorithm is perfect and correct. What about the training data? It is trained off of open source code online, most of which is bad code. How can it correct itself if it doesn't know what is correct? The more code that AI writes, the more AIs that will be trained off of that code. And all of the errors will compound and explode. There might not even be enough data to train these things properly.
I think only copy and paste developers will loose their jobs because they don't understand what they are doing and can't even fix Simple bugs
yes bad code. i seriously doubt you understand how it works under the hood and its flaws
@@awesomedavid2012You're correct, Sir. I haven't heard anyone rave about how great and novel the software is that is being generated. It can do pretty good at regurgitating what it was trained on but not at prividing new innovation.
I don't know why but whenever I tries to learn coding RUclips start suggesting these types of videos😂
People don't understand just how complex an app ecosystem are right now. There are so many things you have to update/maintain/compliance be it an api or libary or even just Google or apple terms and conditions or security updates etc.. AI is not going to do that for you.
It already is. Developers are paid primary to code. And if AI can do that for you. What stops are person not well versed in tech doing your job for you. They can take a crash course on the diffrent tech stacks and chose from there. Programming has abstraction layers and with more them more people can code. This means supply goes and demand for experts goes down. meaning we get paid less. you dont want you job to on the same level as a fast food worker. Which it is going to be.
@@bravo90_ I've been programming in my niche and I just cannot see how your statement jives with my reality. There is simply too much complexity and unique problems that generative AI cannot solve, let alone even attempt to solve. Experience is the best teacher and computer system are nowhere near being able to simulate a brain
But the problem is not for senior programmers. Here juniors are the one who suffer. No compony will give you such big responsibility and technical jobs right after university.
It used to be, you get hired to code under supervision , and learn from your seniors, in 2 or 3 years you become better programmer.
But now if they start automating junior jobs, sw engeenering become a mid senior job.
It become like cybersecurity.
@@Euquila thats what they are littery doing the tech is there its just up to someone to connect all the dots
AI is our best chance of eliminating all this needless complexity and building the stack from scratch.
I have a friend who works at aconstruction buildings. He is a manual worker. He is probably the most happy and mentally stabile people I know. In his job he does details, not lifting heavy stones or something. He works with cables and paints walls, etc. He doesn't have to worry about all that.
Programming is more respected and people consider it much better than what my friend does.
But still I think soon I will give up on that IT industry. I spend 3 years in IT and it is too much for me. Always uncertainty, infinite learning and fight. I just wanna be happy honestly, even if I am not the richest person in the room.
"and i will see you in the next onnnnneeee" ah Tim, you truly have no idea how many times I hear you say that on a weekly and even daily basis haha and it always puts a smile on my face
Its always "..just another level of abstraction.." until its not😁😁😁
Programmers are going the same way as human cashiers at the checkout in the supermarket. The Lidl in my neighbourhood used to have 5 human checkout points. Now they have 8 self scan checkout points and 1 human cashier and guess what? There is a waiting line for the selfscan points and even when there is nobody at the human cashier point, customers will still prefer the selfscan machines.
Well, here in the UK the opposite is now happening. People don't really like those machines for large numbers of items. As such, several large supermarket chains are now pulling the machines out and employing more people for checkout roles. Indeed, Booths in the north of England has got rid of all their machines. Asda, as I recall, intend to halve the number of machines. So your argument doesn't hold water here in the UK, at least.
Dude, who even does the self chekout? Thas weird
Its obvious you are not a programmer.
@@cbnewham5633 I think that is fascinating, why would they do that. Do you know more of the reasoning behind their decision?
@@paraglide01 surveying their customers, listening to complaints and observing the huge lines of people waiting at the manned checkouts while the self service ones are pretty free. For a large shop with a cartfull of items, people get to chat to the checkout person and they only have to wait for the goods to be rolled to them for easy packing, not have to fumble around scanning and doing the packing which takes a lot longer. Lastly, people just feel they are unpaid employees when they have to do all their own scanning and packing.
Whether you agree or not, AGI (Artificial General Intelligence) is the beginning of the end for complex tasks being solely human. The next model, or perhaps the one after that, will conquer these challenges. But make no mistake-it's only a matter of time before it arrives.
for complex tasks being solely human.? for simple tasks too. why not
Not sure.
I've read some academics believe we may need a new paradigm beyond LLM to go significantly further with reasoning and what we call "intelligence".
They argue that the rate of improvement of AIs is slowing down, and that you cannot get incrementally to AGI on the path we are currently using. That we cannot scale indefinitely using the current foundations. Who knows?
AGI is not going to happen for many years. LLMs are not a viable route. o1-preview, as I've shown on my channel, cannot do basic maths. Without that you just have a fancy writing machine, not an intelligence that will perform any function.
Even Meta Chief AI scientist is skeptical about the AGI. It require a giant leap from current AI models/ algorithms to achieve AGI.
@@cbnewham5633 Its all scare mongering. Am I saying avoid this tech not at all but cmon.
As a junior dev i appreciate this message.
It is going to end software developers once they add agent-based architecture into the mix. Frog in boiling water type thing. They already have the tech they're just rolling it out slowly to avoid regulations.
yeah like they barely put any effort into marketing / hyping up o1' s release unlike 4o's. They just randomly decided to make it accessible and let us knew via twitter.
It will take time for companies to adopt the technology. In 10 years coding as it’s done today will be obsolete.
@@jimpresser3438lol on what planet is coding obsolete
@@hendrx On this one in 5 to 10 years. I just retired as a dev and I can see what’s coming from testing these models. Right now they are great tools to accelerate your work, but realistically do you actually think you will be writing software in 5 to 10 years at the rate this technology is improving? You will just tell them what you want written and it will generate the code. This new model was able to actually do graduate level physics problems out of Jackson. This is the standard Electromagnetic’s book used in Masters and PhD physics courses. I used the book for my masters in Physics and I was blown away seeing this thing to some of those problems which require partial differential equations to solve.
@@jimpresser3438
Your argument was used 5-10 years ago. AI can still only solve rather simple problems, or answers with straight answers, such as those from a university level text book.
It cannot create ideas or solutions, it can only apply the ideas from it'a dataset. If you just reitred, you'd definitely know that in a real world software problem, you cannot plug and place something from stackoverflow. You can only do that if the problem is simple.
I use AI everyday at work, yet the reality I encounter is not the one you are describing. It can only speed up the developing process of the simple parts of a system, and it is still performing rather poorly.
At this point AI is in the same boat as fusion energy it's "only 10 years away!"
I have no need to worry seeing how much struggles in real life tech stacks firat hand.
I’m still in high school, but I think the o1 model will help programmers by saving time instead of taking their jobs. Plus, AI will open up new jobs for people like designers and engineers.
This is true. My only concern is once you need less people to program because they’re more efficient it’s possible that demand goes down.
High school boy knows better than some grown up people.
The reason why companies hire programmers is to save time. AI helps save time with less people.
till you graduate , things are gonna change drastically kid
turns out high school students are smarter than most companies
I retired from my previous industry (public education: English teacher, school counselor, and admin) about two years ago. And that's when this wave of LLMs started taking off (not exponentially, but definitely increasing in pace). I've had several friends and recruiters tell me that my main advantage for job seeking will be my broad soft skills gained from my previous industry: dependability and reliability, communication skills (oral and written), creativity and imagination, teamwork (both vertical and horizontal), critical thinking across multiple spectra, adaptability, empathy, and eagerness to learn (and engage others in my learning).
Tim, that advice is outdated, especially for juniors. The people getting hired today are already in the industry. If you're just starting out, it's really tough. Employers won't hire anyone without experience, but you can't get experience without a job. They're outsourcing to cheaper countries, and a lot of entry-level tasks are being automated. The only places still hiring juniors are government programs or large corporations, usually through CS graduate programs.
To be honest, you're more likely to get hired as a plumber, electrician, or handyman than as a developer right now. The tech industry is still great in many ways, but the job market for developers is brutal unless you have a master’s degree or strong connections. If those opportunities ever existed, I would’ve found something by mid-last year, not spent 15 months job-hunting. If anyone disagrees, that's fine, but it's not just my opinion-this is the reality of the current job market.
You absolutely can. Make a project.
It seems to me you are ascribing the result of various trends to AI, and it s wrong from my pov for various reasons:
1
AI lets "bad developers" access the brain output of "good developers"
That means that the California tech bros are losing competitivity very, very rapidly, vs the rest of the world.
If you go to East Asia, you can get coders for a tenth of the US price, and thanks to LLMs the difference in output has vastly been reduced.
2
People who went to prestigious unis and were able to secure employment through the name, resting on their laurels because they had exclusive access to knowledge, are now facing competition from self taught devs from remote places.
3
Because LLMs make the knowledge available in various languages, non eng speakers can now access and learn more easily.
Local mgmt is thus more inclined to hire them since they are able to communicate properly.
Getting barely eng speaking Kor mgmt to instruct a CA based eng only dev teams was a nightmare.
Now they have the same capabilities in Korea, they will hire Koreans, as they are cheaper as well.
4
Increasingly "software devs" are seen as just IT, and companies require less dev skill, and more Network / Systems / Infra skills, since the automation trend is picking up, and that needs a broad understanding of things.
All in all, don t be an "average to good" dev in the US, you are too expensive and if you can t work for the giants the other companies will outsource the job.
And do not be a "I can only code" IT person, you can have less advanced knowledge in coding and develop an understanding of the other fields, that will help with getting a job tremendously since you are then versatile in entry positions.
@@cayjutler1262 He's already wasted 15 months - how long do you suggest he should slog away on a project?
Will his project make him stand out from the tens of thousands doing the same thing?
Will his project counter the shortcomings his CV obviously has to employers?
It might be better to spend the time retraining in a profession that's actually hiring juniors or developing a side hustle like affiliate marketing.
Developers are 10 a penny and we're never going back to the heyday of easy tech jobs for everyone.
@howtopassthat
This is just partly true, but not 100% true.
The market is competitive? yes this is correct. But it's not like fresher or junior got 0% chance like the way you are implying. They still have a chance, even the emerging of AI, AI now is just an assistant, nothing more. For 10 years in the future, I cannot say, but right now, no sight of replacement happening because tool is still a tool, and this tool still need human intervention in order to function properly.
So instead of looking at negative sides, we should moving forward and keep updating ourself to not become obsoleted, then the chance still many. But if you are lazy, "you already dead".
If you truly love software engineering (notice I didn’t say just coding), and never quit, and keep learning, and keep growing, you WILL stand out and you will succeed. Also notice I didn’t say it won’t be hard. In life things worth keeping and cherishing are usually hard anyway. For context, it took me 7 years to break into the field. I “quit” and got back at it many times during that process. Can’t tell how many times I felt dumb as a rock. Now I’m coming up on becoming a senior full stack engineer soon. I hope this is encouraging!
As a person who has been in this industry for 30+ years, I totally agree with you on this.
Eventually the title won’t be a question anymore, just acceptance.
Yes, accept the limits of ai.
Good video, thank you Tim I hope all is well!
100 % agree - so clearly stated . i’d like to see a screencast showing how you actually use ai in your work.
Thanks Tim, this is refreshing to know. The points you highlighted are very good. I Have always seen AI in programming as and advance code completion, nothing more.
Than you are not that smart as you think you are. Sorry.
I'm still a 2nd yr CS student and 2b honest this whole AI advancements used to bother me. This was until I started using chatgpt. Instead of just asking it to generate code for a task so that I can quickly finish and submit my weekly practicals, I started having "conversations" about whatever task I want to do and this helped me a lot in spotting its mistakes in the code it generates before I'd even intergrate that code in my project. In a way it gave me time to focus more on what could go wrong on the project/practical I'd be writing. After I finish writing my code, I'd normally paste some of my files on it so that it can check if I didn't miss anything that can maybe mess up my code, especially with handling exceptions coz I'm not very good at that.
Great video, noted👌
That's exactly the whole plan of openAI lol...
So the more better ai becomes the less people will be needed simple m
When i was a young engineer we were told to try and "Engineer out the technicians". AI "WILL" be able to Engineer out the Engineers! It will not stop with engineers, AI will be everywhere and SOON! This is not a bad thing in the long run. Now this does not engineer out innovation, in fact I think it will accelerate it! A few months ago there were blogs saying AI has peaked-lol. I remember hearing all these people saying look we are flattening out, but now here we are with another WOW with AI. We have not even scratched the surface of what the current customer level AI can do! I can not even imagine on what is going on in the heart of OPEN AI right now with GPT 5!!! We are in the dawn of a new technology revolution and AI will lead the way! AI will lead in things like cancer research, anything medical, energy maybe it will solve fusion, technology, longevity, space exploration, food and farming, and so many more i can not even think of right now!!!
AI code is going to give every person a developer side kick, that also knows how to implement numerical methods to solve problems.
Engineers will have AI write tools for them to solve engineering problems, but won’t replace engineers.
Scientists can have it help model their data, visualize, etc. it’s going to save so much time for STEM workers.
There will still be software engineers and developers to direct the AI, but far fewer humans.
One developer can now do the work today that it took 2 or 3 people before AI.
Obviously that means the job market is now saturated.
Perfect analysis. This is an evolution and a tool, and this changes our software development methodologies and know-how, but this is just the beginning. When AI will be fully automated with more sophisticated app templates, coder work will be limited to know how to ask a customised app responding to customer's enquiries. And after, there will be another evolution. Just think that the biggest human evolution is the communication between people. LLM/AI is the communication between humans and machines. And at last, there will be the communication between machines. That's the point where humans will be over.
Developers are safe now, until next year when autonomous AI agent will be coming. I dont think it is a 'natural evolution' of software development, it is more a revolution and a lot of programmers risk to loose their job in the next two years.
That’s delusional. Most AI companies won’t even be around in two years.
@@tradingviewindicatorguides895 I say that a lot of developers risk to loose their jobs, not every developers. Maybe not the techleads. And I didn't say that is was optimal at the moment, but in a near future, maybe with AI agents, it will certainly replace many developers.
@@tradingviewindicatorguides895gpt 3.5 solves leetcode problems , u have to ask properly that's itself tough
@@tradingviewindicatorguides895 for me it makes peaces of good code
In the future, only the people who have great ideas will be worthy. Their mind and AI collaboration.
I absolutely agree with Tim, AI is like our Friend, and it is not a replacing the jobs, in fact it's creating more jobs for software developer and prompt engineer
Hi Tim, great great video. Thanks for restoring my confidence. I am learning java mainly using Ai actually.
And it is true, things are changing in many other fields too, for ex I used to be a manual turner metal engineer in the past and not it is all computerised in general, incredible what this machines can do now.
Not that I fully agree but you kinda need to keep up with the times somehow if you want to work.
Thanks again
I fully agree with you. Coding is/ can be very tiring. So if you can save brainpower by using this tools for the low-mid level stuff, you can focus on an the more difficult things with your limited brainpower.
Full disclosure I am a Senior Software developer. Just watch Star Trek the Next Generation. They tell the computer what they need and it just does it. That's where we are headed. These LLM models are at best 1st Gen releases and they are already this good. In 2 or 3 more generations I expect they will be so good that software development will really be a thing of the past. They are already creating models that can generate rudimentary 3d games. Ten years from now at this rate I am pretty sure we will have perfect responses to our requests no matter what they are.
If it will achive this level, most industries will be replacable, people won't have jobs, leaving for guaranteed income from goverment
@@patryk5654 Most hands on jobs will be safe - electricians, plumbers, mechanics, salespersons, the list goes on.
@@DM-sc4zy robots + AI will solve that as well
@@TheBrothersTalk On a larger timescale though, robotics isn't where it needs to be now to replace fine motor tasks even if the "brain" (the LLM models) are smart enough to know what to do, the "body" (the robots) aren't precise enough to execute the actions yet.
AI can make 10000 3d games but if it is not somehow getting data from my neurons its never going to make a game alike mine.
You are absolutely right. I agree 💯%!
Wow i think you are absolutely right coz it means even all other jobs shall be replaced too if the software engineers are replaced there is no way other jobs can survive
Email was going to replace post men...
But post is still used for official documents and stuff. Post is very prevalent imo
It has majority for communication happens on the web. Post is to slow in this world. If you could get physical products online then. post would be dead
A poor example. Here in the UK it has killed regular letter post.
Like this guy is going to give any sort of honest advice, or feedback. I’ve learned most of these software RUclipsrs are just like like any content creator… say anything as long as it aligns with their bottom line(money/views).
Yes, for seniors this is great, for mid-juniors is really a problem, at least in the short term
AI will make it more competitive. Regular people armed with Chatgpt can now make amazing things
They can make a crud app or some other toy project? Congratz. Would be better to make it yourself so you can be better than AI.
Yep. Tetris, Snake, and I'm sure they'll crack PacMan at some point. I can tell you from experience last year, getting GPT to write a more complex game from scratch ended up with several days of herding it to do what I could have done in half a day, complex code that was not understandable, and repetitions of code because it kept forgetting it had already written utility functions. It worked - but it was enormously inefficient and the code base was really unmanageable.
So true. I have written code for the back-end of a transcription software with the help of Chatgpt-o1 even though I have no background in tech.
@@CricketFantasies5 would it not have been more fun and more insightful to have understood how to code it yourself, even if you used the AI to help you? That seems to me to be more advantageous than getting it to produce a black box which you don't understand and cannot debug.
@@cbnewham5633 well, I am in content creation field. So, I don't want to spend my time learning to code, rather I want to use it in whatever meaningful way I can.
At normal circumstances, I could have bought a subscription of an already present software built by a developer, but AI helped me develop my own software (only back-end) saving my money and time (to learn coding and the software works perfectly fine.
That's why, I say that AI will give the software developers a tough challenge as more and more people will do it themselves or the developers will face severe competition among themselves leading to a portion of them becoming jobless.
Totally agree with the take here. AI might change the way we code, but it’s not going to replace the need for skilled developers anytime soon! 💪
it will replace the need for most skilled developers
Ai is going to take our jobs. You asked what I'm going to do instead of run full speed at a brick wall? I started my own business using ai. I'm looking to expand my empire now. The problem with dev is the amount of work that goes into it just to be made obsolete in only a matter of time.
5 seconds in and straight to the point, Thank you
like all things, working smarter not harder.
Let's come back 1 year later and revisit this video.
Tbh I am scared.
This whole idea of software developers being replaced has 2 years now since Chat-GPT was released, and nothing happened yet...
@@JVWorks yet
🤣🤣🤣 you never touched coding
@@JVWorksIt sure did not happen but the models become much better. There will be software engineers, but fewer. This is the same moment as CAD introduction to mechanical engineers, now one ME can do the work of five.
True, true... I work in automotive industry as a industrial engineer but my background is automation programmer. All modifications on machines have to be outsourced because we do not have team for that. As I understand the topic very well, I'm able to implement those modifications nearly for free with help of AI as my "junior programmers". So there is power in using those tools, I save a lot of money because usually even small software additions costs thousands of euros plus external companies can't to it ASAP because they have their own projects. We can apply something like 80/20 here. 20% of a problem is actual coding, but I will have to still do the 80% of that job. On the other hand, those 20% of coding usually takes 80-90% of that time. With AI, this time can be decreased by 50-80%, what is great. at the end, everyone is happy, corp. that I work for valuates me even more and I'm able to do what I love at position where I shouldn't code at all.
"The data flywheel refers to the self-reinforcing cycle where the more data a system collects, the more value it can provide, leading to a compounding effect on insights and innovation."
Unfortunately for people just getting into this field, the flywheel has just started. There's no going back. Enjoy the ride as AI takes over
The way I think this will play out: a lot more people will start producing code with AI. For example, an analyst will use it to write some python to write a colab script to connect data. Millions will start writing code like this. This code will break and they won't be able to fix it. That's where you come in. You will be so busy fixing code, so learn how to specialise in AI code repair and you will always have a job.
Hello bro can you make a roadmap video from beginner to advanced (absolute essential topics and projects) for your videos to learn python . There are so many good videos, so I couldn't understand where to start. Thank you for your time for this comment and content.
Interestingly, the time spent writing code jas shifted to making sure that AI output actually matches my needs properly. Following it blindly is a sure way to break things
Its just a matter of time when we have new softwares that are built on AI. Then its a new paradigm, but sure people will be in the loop for sure but we dont know what will be entirely automated that used to be code by hand. Time will tell🙌
Just knowing how to code is like making a brick when you’re asked to build a house.
Been a developer for over 25 years, have been trying to get these new AI tools to accomplish tasks so I can get more done. It's always close but has yet to be 100% yet. The issue I keep seeing is that these companies keep trying to sensor things by adding invisible prompts and its screwing up the code return. The best i've got was running models locally but i've only been able to run smaller models due to the cost of hardware.
It´s great for new features, I imagine that those models are great for startups, however, most software development is 95% searching the right place and making sure that what you do doesn´t break things and 5% writing code.
100% mirrors my viewpoint as a fellow millennial.
The yak-yak software engineers will become obsolete, right !!
The future of application/system development will be impossible for the layman, only an engineer can think these things through. Lay people do not know what they want/need unless its given/offered to them.
I gotta be honest, I don't think AI will ever fully take over software development. However, I could see it doing to the developer field what stable robust servers did to network engineers. Like when I first started working as a network engineer, the formula was simple. One administrator per server or 100 users. However, by the time I left you might have a team of 3-4 managing 100 servers and thousands of employees. I can't even imagine what it's like today since it's been about 17 years.
So, I'd guess you'll still have developers, but you'll need much fewer per project and AI will do most of the simple lifting.
I guess this makes juniors like me obsolete. But it also allows faster dev time so that one can sell apps and the like i ugess.
Or, like with pretty much every other field where automation has happened, the amount or complexity of things being produced will increase, which means the same amount of workers will be required. What I'm more concerned about is that companies will find a way to use these changes as an excuse to exploit workers even more.
its not the end for software developers, although i expect the number of jobs in the industry will be reduced and the workflow will be a bit different.
and yes, like u are saying, coding will gradually become higher level and higher level although there will still be demand for those that actually understand the low level stuff.
I like to compare AI with everyday life tools. In the past we could clean with brooms. They still exist but now we also have vacuum cleaners, leaf blowers or outdoors vehicles that make the task easier or even vacuum robots that will do it themselves. Same for cutting grass, from scissors to manual mowers to all kinds of machines. Also remember when frameworks appeared? I think it's all part of progress and it's more tools that get added. We still need to be part of the process to make them work as we want.
Wise words. You speak like an old and wise (I'm 61)
Supply and demand. Productivity will go up radically, cost will go down and hence more demand for software. This may translate into more or fewer programmers (probably fewer) but "everyone" will not become a programmer. That said, I think Tim is correct with what he is saying, but I think at some point most programmers who work on repetitive type apps will have a very difficult time of it.
The thing is biologics is not keeping up with AI and machine. One should use ai to evolve biologics. I do not mean zombie, but robust super biologic to keep up.
The future of software development is problem solving and ability to colab with companies, u have to understand everything within
Everything is a panic. You’ll find articles panicking about not enough people having babies, and within the same article, warning that the cities are over populated and we’re using up all the natural resources. People need to choose, pick one or the other. And realistically, how much code does the world really need? At what point should we start focusing on things like clean water, mental health, not working 80% of our waking hours, etc?
It was an informative and fun video to watch. I respect your thoughts and agree with them
Change the way you think, it does not effect developers in negative way, but a oppotunity write more code and effective. So as a developer my self it make life easier not misserable. replacing a human with AI , we are not their yet.
Its not about being replaced, its about 10 devs effort for 2-3 devs can also do with AI (the 7-8 people is jobless tho)
I went for this field cause I love coding and it seems like that is going away in the next years. I don’t know how I could possibly feel encouraged right now..
Welp, I just started learning Python about a week ago and now I see this video?! Realistically what is the point of continuing with my journey when I currently know incredibly little and AI seems like it will be a step ahead whilst I learn. Back to the trades I guess....
you just started learning Python
you just started learning Python but chatgpt knows it. it can create simple python programs
I'm in the same situation man
Why are people still learning programming? I’m confident there’s little room left for freshers in this field. At this rate, within 1 to 2 years, only a handful of freshers will be needed to support senior developers. Experienced programmers will remain in demand as some human oversight will still be required. Yet, thousands are competing for the same roles, and senior developers will get them in the near future. I don’t understand why people are still rushing into this field. Freshers should look into new opportunities and have the courage to pursue something different. After all, humanoid bots are likely to take over most jobs soon anyway!
I'm keen to understand how my view might be wrong
The problem is that there is no safe job. You want to be medic or psychologist? If AI can replace developers, it will be also able to give medical examination and do mental support to patients. Lawyers? The same. Managers? It will be able to very boost productivity of managers, less managers needed. Manual workers? Sure, but there will be much more of them in the market (former white collars) and humanoids robots will slowly start to take over.
I need it to immigrate and I have already spent so much learning it. Giving it up would mean I've spent a soul crushing amount of energy on it that won't be replenished in a very long time if ever. I also know testing, so I keep it in might as a fall-back position.
Why I ever started coding? This is the only thing I find interesting in IT.
It would be interesting to see let’s say a 1 year difference between someone who just started to learn coding through AI vs. someone who has been a developer for 5-10 years but just started to use AI.
Choose a starting date and compare your project/startup after 1 year. How far did they get? Was it actually useful?
Interested to see where this all goes!
I can't help with a direct comparison, but I've been doing coding for 35 years. AI is a useful tool, but for anything that's complex you really have to know what you are doing - AI hallucinates and the code doesn't always address edge cases. Great for boiler plate stuff or getting started with APIs you haven't used before, but I wouldn't just ship code straight from an AI without proper reviews.
Hey Tim, how about we start coding again, using AI this time. Cant wait to see what you will do with these tools 🤪
This is just a cope. Developers are definitely going the way of the dodo. We're at the early stages of this AI thing and it's already this good. In a couple of year a handful of non developers will be able to huddle around a table or over zoom discuss a problem and design a structured solution. The AI will be able to build that solution and make it available in one afternoon.
Developers will be replaced by structured planners
Dev channels like this can't scare away their audience or else no one will watch their videos anymore. They have to keep giving junior devs hope that they're education and channel views still have meaning.
It’s far more likely that software engineers would rise to the role of design and AI prompting and tweaking, which would obsolete the non technical worker roles. There’s no software engineer I know that can’t do the job of their non technical colleagues, but the inverse is certainly not true.
“Devs” will not be called “devs” in this hypothetical future. They will be design and solutions specialists and their background and ability to understand, tweak, design, and prompt these AI to arrive at solutions will prove incredibly important.
Its funny how people arrive at the conclusion that now suddenly technical expertise is useless and all the clueless MBAs with their “ideas” will be spraying prompts at the AI and somehow creating flawless scalable applications without any intervention at all.
As a PM I’m desperately trying to acquire technical skill because I can see the writing on the wall in that not having it will probably render me near useless in this new future.
@@thechannelihadtomakeforthi9495 you're just confirming my point. The role of software engineer will be gone. A few software engineers will rise into the software architecture and designer roles for big projects and big clients. The rest will become part of a large hoard of low value AI prompters.
@@stinger4712 I don’t think it automatically means low value.
I think it’s more likely we will enter an age where customers are asking for way more value for similar budgets. If AI greatly accelerates the speed at which products can be developed in the way you’re implying, then customers are not going to settle for similar value at similar cost. They are going to flock to companies offering them equal value for less money, and subsequently they are going to have more money to spend on more value.
The only way I don’t see this being the case, is if customers are happy with the software they have and don’t want any improvements or new products/features, which I don’t see as being the case especially with the massive backlogs I’ve seen throughout my career (and unhappy customers always asking for outrageous things we just can’t do within budget).
In this scenario, I think there’s increased demand for talented engineers who can do a little bit of everything and leverage AI to vastly speed up their deliveries. We might now expect a whole feature done in a few days rather than a few weeks, but that might also mean we do 5 features in a few weeks rather than 1.
Of course, I could be wrong, and we could enter an age of tech recession where people move away from technology and software to focus on other pursuits, but I think we’re a long way off from this.
@@thechannelihadtomakeforthi9495it's not just about AI accelerating your work. It will take the below average to expert level. I'm trained as an electrical engineer and I'm good at it but I considered myself mediocre when it came to writing software for electrical systems. Then came chatGPT and overnight I began delivering software for the most complex systems. All I needed was the systems knowledge to quickly displace the high cost developers. That's what's going to happen across domains.
This will be like Office for Windows where Office originally came free with the Windows program.
Now Office comes with in with a subscription charge.
Simple AI’s free like Chat GPT but anything on a developer’s desktop will eventually become a subscription and the charges climb every few years.
it is not a problem, companies will pay as for office365
Dear Tim, you forgot to plug your programming course. 😁
To me this doesn't indeed replace developers but make fewer them needed to build entire system.
Not yet. But GPT5, coupled with agents, could definitely be a major nail in the coffin.
I am a software developer and when using ai i can vouch that it makes you 2-3 times more productive. This equates to less programers needed. I have already noticed that now every job has over 200 applicants... Way over what it used to be like
I’m not worried at AI taking my job I’ll just learn how to program the AI. Besides currently I do customer service and I’ve been told for years AI is coming for my job, but AI may be able to guide the customer through the return process the AI dosn’t understand how people work, and that if I don’t highlight the collection link there is a 90% chance the next email from the customer is how do I get it collected. AI doesn’t understand that what people ask isn’t necessarily what they want to know the answer to.
I don't agree with this. I recently had Claude 3.5 make this very complicated piece of software for me that manages leads and bookings for my business, it is integrated with Google calendar and Google drive. All I was doing was just describing the functionality I wanted and giving the model any errors from the console. It also did extensive testing of the application. Soon software developers will not be needed or at least much much less.
From your brief description, that software has probably been written thousands of times already. Software for booking systems is so common and it's a very easy task for something that's been trained on all the code on the web to complete. Even just integrating a new feature into complex software, like RUclips for example, would need a lot more intervention from an actual developer.
@@technophobian2962 for now maybe.
@@technophobian2962 Have you seen that company that is generating open world games from text prompts? Did you see Google had a neural network recreate the Doom engine? I actually had Claude 3.5 code something like the Doom engine. I find it most impressive that it can tailor make specifically what you ask for.
I’ve been programming since my childhood 1990 and as of today I’m still employed as a programmer and still don’t see LLM replacing humans yet. Useful tool, and someday it may replace humans, but we are definitely not there yet.
I agree,100%, but I probably differ from you and then I think that someday might be as soon as 2025
its over for begginer engineers, but to the top OG engineers, its the 6th Infinity Stone :-)
In simple terms you must be knowing what you are doing. You just can't become a software engineer overnight because you have ai. Neither is software engineering being replaced.
I have also been using ai and it's very useful
I appreciate your channel,I learn a lot from you.However ,plz don't do the click bait bs tile ,I don't think it will be good for both parties in the long term,cheers
Human energy is much cheaper than energy required to run such AI systems indefinitely for complete automation.
So most jobs will evolve instead of completely getting replaced.
Mother Earth has really ensured only natural things survive on it.
The fact that this CLICK BAIT title is a real concern - logically, development of AI is heading towards phasing out developers. The indicator being Click Bait. Disagreeing with this is like saying guys who click on videos with beautiful big breasted women do not like boobs. Which is false.
Even if not a general purpose if replacing software development - monetizing in such a stigma Will align AI with out-dating software engineers. Like even if unnatural but because people.
Technically if EVERYONE lost their jobs at the same time - would have zero impact on the economy. Food for thought. Like when demand for high jobs is high the dollar is strong, but if demand dropped to near zero overnight - how supply demands works would just affect the underlining value of the dollar and If, for example - If I develop AI in hope of making other people lose their job since I am unemployed would make me wealthier since there would be less wealthier people in contrast - So if everyone did this at same time Should, in theory have no influence on individuals personal worth.
I think anyone who really wants to learn these stuff can be a developer. Right now, even CS majors are having a hard time getting an entry level job or an internship. Why? School does not really teach them the practical stuff.
Sorry the future of AI is creating blueprints. You'll have blueprints for all these different types of software that the AI will follow and create variations on these blueprints to create new software.
Maybe AI will get a level behind our imagination, but you will always need humans who understands this technology!!!
Spot on. My CS degree is from 1991....the CRITICAL THINKING hasn't changed and being able to design is a skill. The syntax changes and has changed, but having a logical approach, structuring logic, and knowing when something is bad, is a long term marketable skill.
When AI can take one code base and move it to another without error. Then AI is interesting. For example take a react web app to vue with one requested prompt. Anything less isn’t anything to talk about.
15 years ago I was pretty much using the same tools and languages as I am using now :) that is c/c++ (yes, stl existed too back then 😁) and assembly too. And even visual studio 2005 was OK to use. And faster unlike those Javascript based IDEs we are using in 2024. But I am not saying that generative AI is not useful - it is but it is not going to replace a decent coder any time soon.
BTW I code for living since 2002 and I wrote my first signal processing code in assembly in 1993.
This is actually the beginning of software development...
If we see from starting and take maths as example then, in start math was greatly used and there was no machines to calculate simpler problems only mathematician can do those problem or someone who knows about math and after we upgraded and created our first calculator then we thought the job of the 'mathematician or the people who worked on shops and in business to calculate money expenditure' will be vanished because of calculator but as we know it doesn't happened because after calculator we were able to solve problems of that time when it was difficult but it is not now,
So we now started focusing on more difficult problems and on more no of simpler problems to be solved at the same time with the 'help' of calculator which makes it easy to solve.
And it is same in programming we are now going to develop after a decade we will be focusing on a very high tech program or software where openai will be a help for us as a calculator to a student to solve addition and subtraction
When ai came everyone thought humans will be not that much useful for work or job but it doesn't happened because after ai we got developed and our efficiency of job also increased so we then needed ai for help to do our job together with more efficiency.
And humans always are evolving
Now even some human can calculate faster than and with efficiency and accuracy than the calculator as we see in shows like talent search show or anything.(America got talent). Maybe I don't know exactly name of those shows but I saw some clips and short videos of it
If you read along please pin me and like if you are not the channel owner
What type of camera do you use?
funfact: c was actually considered a high level language back in the day
Excelente análise, parabéns!
When it comes to unit testing, AI is a life saver. It's saved me hours
Do you use GitHub copilot or just regular ChatGPT ?
At the end of the day, you still need to know software engineering to debug and iterate on code written by AI. What AI will do, however, is turn some solo devs into sole proprietor AAA studios.
thxx broo good content!!! ❤
Choosing this as my career seems like a waste of fucking time. I should’ve gone to trade and welded or done electrical work.
Learn to plumb!
Red Seal
Take 20, action!